ArticlePDF Available

Computer Vision-Based Wood Identification: A Review

November 2022
Forests 13(12):2041

DOI:10.3390/f13122041

License
CC BY 4.0

Authors:

José Luís Amorim da Silva

CITAR - Universidade Católica Portuguesa

Rui Bordalo

Universidade Católica Portuguesa

José Pissarra

GreenUPorto

Wood identification is an important tool in many areas, from biology to cultural heritage. In the fight against illegal logging, it has a more necessary and impactful application. Identifying a wood sample to genus or species level is difficult, expensive and time-consuming, even when using the most recent methods, resulting in a growing need for a readily accessible and field-applicable method for scientific wood identification. Providing fast results and ease of use, computer vision-based technology is an economically accessible option currently applied to meet the demand for automated wood identification. However, despite the promising characteristics and accurate results of this method, it remains a niche research area in wood sciences and is little known in other fields of application such as cultural heritage. To share the results and applicability of computer vision-based wood identification, this paper reviews the most frequently cited and relevant published research based on computer vision and machine learning techniques, aiming to facilitate and promote the use of this technology in research and encourage its application among end-users who need quick and reliable results.

General scheme of machine learning method for image classification (based on [63,68]).

…

General pipeline of deep learning models for image classification (based on [63,68]).

…

Summary of computer-assisted wood identification systems.

…

Wood image datasets available for computer vision-based wood identification research, adapted from [63].

…

Research and applications of deep learning technologies.

…

Figures - available via license: Creative Commons Attribution 4.0 International

Content may be subject to copyright.

Available via license: CC BY 4.0

Content may be subject to copyright.

Forests 2022, 13, 2041. https://doi.org/10.3390/f13122041 www.mdpi.com/journal/forests

Review

Computer Vision-Based Wood Identification: A Review

José Luís Silva 1,*, Rui Bordalo 1, José Pissarra 2 and Paloma de Palacios 3

1 Research Center for the Science and Technology of the Arts, School of Arts, Universidade Católica

Portuguesa, 4169-005 Porto, Portugal

2 Green UPorto and Department of Biology, Faculty of Sciences, University of Porto, 4169-005 Porto, Portugal

3 Department of Natural Systems and Resources, School of Forestry and Natural Environment Engineering,

Universidad Politécnica de Madrid, 28040 Madrid, Spain

* Correspondence: jlamorimdasilva@gmail.com

Abstract: Wood identification is an important tool in many areas, from biology to cultural heritage.

In the fight against illegal logging, it has a more necessary and impactful application. Identifying a

wood sample to genus or species level is difficult, expensive and time-consuming, even when using

the most recent methods, resulting in a growing need for a readily accessible and field-applicable

method for scientific wood identification. Providing fast results and ease of use, computer vision-

based technology is an economically accessible option currently applied to meet the demand for

automated wood identification. However, despite the promising characteristics and accurate results

of this method, it remains a niche research area in wood sciences and is little known in other fields

of application such as cultural heritage. To share the results and applicability of computer vision-

based wood identification, this paper reviews the most frequently cited and relevant published re-

search based on computer vision and machine learning techniques, aiming to facilitate and promote

the use of this technology in research and encourage its application among end-users who need

quick and reliable results.

Keywords: computer vision; machine learning; deep learning; convolutional neural networks;

image recognition; wood anatomy; wood identification; illegal logging

1. Introduction

Illegal logging is one of the most pressing environmental issues, particularly in trop-

ical countries with large forest areas and botanical groups that are highly valued in inter-

national markets. Illegal logging is currently the most profitable ecological crime world-

wide, accounting for 10 to 30% of the global timber trade [1,2].

Although Amazonian forests are traditionally seen as the hotspot of illegal logging,

areas such as Southeast Asia, Central Africa and Russia, home to roughly 60% of the

world’s forests, are unfortunately experiencing a surge in this crime [1,3]. The financial

impact of illegal logging is estimated at 52 to 157 billion dollars a year [1], but, more im-

portantly, the environmental damage, in many cases irreversible, can also have a global

ecological impact [4].

Several institutional and international legal measures have been put in place to pre-

vent overexploitation and irreversible loss of species and habitats [5–7]. Innovative pro-

grammes are emerging [8,9], solid research is under way [10–14], and research with sig-

nificant impact and news items are being shared worldwide [15].

The impact of wood identification extends beyond illegal trading and ecological is-

sues. Wood identification is paramount for the timber industry, civil and structural engi-

neering, criminology, archaeology, art history, ethnography, and conservation and resto-

ration, and many other disciplines.

Despite the multiple wood identification methods now available, the varied results,

costs, accessibility, deployment time and limiting factors hinder their applicability to real-

Citation: Silva, J.L.; Bordalo, R.;

Pissarra, J.; de Palacios, P. Computer

Vision-Based Wood Identification: A

Review. Forests 2022, 13, 2041.

https://doi.org/10.3390/f13122041

Academic Editor: Christian Brischke

and Cristina Nabais

Received: 21 October 2022

Accepted: 29 November 2022

Published: 30 November 2022

Publisher’s Note: MDPI stays neu-

tral with regard to jurisdictional

claims in published maps and institu-

tional affiliations.

Licensee MDPI, Basel, Switzerland.

This article is an open access article

distributed under the terms and

conditions of the Creative Commons

Attribution (CC BY) license

(https://creativecommons.org/license

s/by/4.0/).

Forests 2022, 13, 2041 2 of 26

world identification. This paper presents an overview of the changes that have occurred

in wood identification methods and a review of computer vision-based wood identifica-

tion, which is currently one of the fastest developing research areas in artificial intelligence

(AI) with very promising results and high identification accuracy. In this technique, visual

data are processed from any given image to extract the relevant features in order to make

a decision.

Analogic and Digital Systems

Historically, wood identification methods mainly comprised the study of chemical

and physical and anatomical features aspects of wood. Methods such as macroscopy,

which uses the physical characteristics of wood observable to the naked eye or with a 10×

hand lens, and microscopy, which resorts to light compound microscopes to the observa-

tion of multiple cell typologies that constitute the wood, were the first to be used. The

main limitation of these methods is that wood cannot always be identified at the species

level. As a result, there has been an emergence of multiple techniques such as near-infra-

red spectroscopy [16–18], DNA barcoding [19–21], mass spectrometry [22–24], and X-ray

tomography [25–27], with optical microscopy still used as a confirmation method for the

results of these techniques.

However, an important contribution was made with the advent of computer-based

technologies, which rapidly became a preferred option for constructing species databases

and hosting identification tools. Several wood identification databases and software based

on digital technology have been made available, including GUESS [28,29], CSIROID [30]

and the DELTA system [31], three of the most significant early programmes in achieving

the goal of wood identification. As a proof of concept, the importance of these systems

was fundamental for the development of what is today defined as computer-assisted

wood identification. The results they obtained made considerable progress compared to

earlier methods, especially with regard to the time required and identification accuracy.

The DELTA-Intkey for commercial timbers is the only one of these three systems still in

use.

2. Online Reference Databases for Wood Identification

This section briefly describes all the digital reference databases available online, to

the best of our knowledge. The common objective of online identification keys is to enable

and facilitate analysis of wood anatomical features and, ultimately, identification of the

wood. Table 1 summarises the computer-assisted wood identification systems mentioned

above.

Forests 2022, 13, 2041 3 of 26

Table 1. Summary of computer-assisted wood identification systems.

Year

Name

Taxa

Area

Identification

Number of

Images

Access

Comments

Reference

2000

Commercial timbers: descriptions, il-

lustrations, identification, and infor-

mation retrieval

404 hardwoods

Major forest re-

gions of the

world

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

[32]

2000

Anatomy of European and North

American Woods

325 hardwoods

101 softwoods

Europe

and

North America

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

Includes features adapted to identifi-

cation of carbonised woods from ar-

chaeological contexts

[33]

2003

Wood database of the Forestry and

Forest Products Research Institute

781

Japan

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

[34]

2004

InsideWood

7653 modern hard-

woods;

235 modern soft-

woods;

2173 fossil hard-

woods

Global

Microscopic descrip-

tions and illustrations

58,146 modern

hardwoods;

3807 fossil

hardwoods;

1482 modern

softwoods

Freely availa-

ble

Includes 61,578 searchable images

[35]

2004

Wood anatomy of central European

species

133 hardwoods

and softwoods

Europe

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

Includes macroscopic and micro-

scopic images and descriptions

[36]

2005

CITESwoodID

44 CITES woods;

31 look-a-like spe-

cies

Specific forest re-

gions of the

world

Macroscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

Includes abundant extra information

[37]

2005

Key to a Selection of Arid Australian

Hardwoods and Softwoods

58 hardwoods and

softwoods

Australia

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

Detailed information about each spe-

cies

[38]

2010

Brazilian Commercial Timbers

275 species

Brazil

Macroscopic features;

chemical and physical

tests

n.a.

Freely availa-

ble

[39]

2011

Pl@ntwood

110 hardwoods

Amazonia

Microscopic descrip-

tions and illustrations

n.a.

[40]

2013

Forest Species Database—Microscopic

112 hardwoods

and softwoods

Tropical forests

Microscopic descrip-

tions and illustrations

2240

Freely availa-

ble

Includes 2240 searchable microscopic

images

[41,42]

2014

Forest Species Database—Macro-

scopic

41 hardwoods and

softwoods

Brazil

Macroscopic descrip-

tions and illustrations

2942

Freely availa-

ble

Includes 2942 searchable macro-

scopic images

[43,44]

Forests 2022, 13, 2041 4 of 26

2016

MacroHOLZdata

150 hardwoods

and softwoods

Global

Macroscopic descrip-

tions and illustrations

n.a.

Free of charge

on request

Available in English, German and

Spanish

[45]

2016

Softwoods

53 softwoods

Major forest re-

gions of the

world

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

[46]

2018

Forest Species Classifier

112 hardwoods

and softwoods

Brazil

Macroscopic and mi-

croscopic illustrations

5182

Freely availa-

ble

[47]

41 hardwoods and

softwoods

Freely availa-

ble

2018

Charcoal

44 hardwoods

Brazil

Microscopic features

528

Available for

research only

[48]

2019

Charkey

507 hardwoods

and softwoods

French Guiana

Microscopic descrip-

tions and illustrations

n.a.

Freely availa-

ble

Highly detailed SEM images

[49]

n.a.

Softwood Retrieval System for Conif-

erous Wood

180 softwoods

China

≥1000

n.a.

Under development

[50]

2021

UTForest—UTFPR Classificador

44 hardwoods and

softwoods

Brazil

Macroscopic descrip-

tions

1318

Freely availa-

ble

[51]

Under de-

velopment

Mader app

n.a.

Microscopic features

26,000

n.a.

Database with 1000 images per spe-

cies

[52]

n.a.—not available.

Forests 2022, 13, 2041 5 of 26

2.1. Commercial Timbers: Descriptions, Illustrations, Identification and Information Retrieval

Among the several DELTA-INTKEY online identification keys that have been devel-

oped, including CITESwoodID and Softwoods, this interactive identification key devel-

oped in 2000 and updated in 2018–2019 is an integrated database of microscopic descrip-

tions and illustrations of 409 internationally traded hardwood taxa [32]. It covers major

forest regions of the world and is freely available online.

2.2. Anatomy of European and North American Woods

Created in 2000, this system includes 426 wood taxa [33] and provides an interactive

identification key for the common non-commercial wood species of Europe and North

America. It includes 325 hardwood species and 101 softwood species (native and intro-

duced) with 145 features and 15 sets. It has two extra sets that isolate the features applica-

ble to identification of modified and carbonised wood from palaeobotanical contexts. A

useful feature, which is missing from the commercial timbers key, is a set that isolates the

features extracted from the IAWA standards.

2.3. Wood Database of the Forestry and Forest Products Research Institute

This online database created in 2003 focuses on the identification of 781 Japanese tree

species, substantiating the descriptions on the IAWA list of hardwood features [53]. It

includes a multiple-entry key and an image database and is freely available online [34].

2.4. InsideWood

Developed in 2004, InsideWood is by far the largest and best-known online identifi-

cation key [35]. It is a multiple access key based on the IAWA hardwood list [53] and is

freely available online. It includes keys for hardwoods, softwoods and fossil hardwoods,

and more than 10,030 microscopic anatomic descriptions covering all regions of the world,

with more than 63,435 searchable images. It is a centralised database that integrates all the

anatomical data available for modern wood.

2.5. Wood Anatomy of Central European Species

This is a completely revised and updated version of Schweingruber’s work [54], cre-

ated in 2004 [36] and last updated in 2007. It is a web-based identification key with 133

species, accompanied by macroscopic and microscopic descriptions. It also provides in-

formation such as sample preparation, staining and other procedures, and is freely avail-

able online.

2.6. CITESwoodID

CITES (Convention on International Trade in Endangered Species of Wild Fauna and

Flora) [55] is an agreement that was drawn up in 1963 after a meeting of members of IUCN

(The World Conservation Union). One of many initiatives intended to contribute to the

goals of the agreement, CITESwoodID was developed in 2005 [37] and last updated in

2017. As the name indicates, the platform focuses on CITES species and is an interactive

identification key of macroscopic descriptions with an integrated database. It includes il-

lustrations of 44 CITES protected woods and 31 look-a-like trade species. It provides com-

prehensive, detailed information about each species, with advice on how to avoid misin-

terpretations, and numerous explanatory notes of the relevant features and procedures

for description and identification. It is freely available online.

2.7. Key to a Selection of Arid Australian Hardwoods and Softwoods

Stemming from doctoral research [38], this interactive key focuses on Australian

woods. It is hosted on the Lucid website and includes 58 wood-producing species of arid

Australia, particularly non-commercial species. It is mostly based on specimens from

Forests 2022, 13, 2041 6 of 26

northeast South Australia, southwest Queensland and far western New South Wales and

is freely available online.

2.8. Brazilian Commercial Timbers—Interactive Wood Identification Key

As the name suggests, this is an interactive identification system focusing on Brazil-

ian species. Made available in 2010 [39], it was developed in collaboration with the Forest

Products Laboratory (LPF) and the Brazilian Forest Service (SFB). It is hosted on the Lucid

website and includes 275 species, among them Brazilian CITES-listed timber species. All

the nomenclature was revised in 2020 according to the Brazilian Flora Species List. The

key works by analysing macroscopic features and chemical and physical tests on the

woods. It is freely available online.

2.9. Pl@ntwood

Pl@ntwood [40] was developed in 2011 and is described by the authors as an interac-

tive graphical identification tool based on the IDAO system, specifically designed to be

user friendly. It comprises 110 Amazonian tree species belonging to 34 angiosperm fami-

lies and includes microscopic morphological features.

2.10. The Forest Species Database—Microscopy (FSDM)

Created in 2013, this online database [41,42] comprises 2240 microscopic images of

112 species, 85 genera and 30 families of both hardwoods and softwoods. It is freely avail-

able online.

2.11. The Forest Species Database—Macroscopy (FSDM)

This online database for forest species identification [43,44] was made available in

2014. It includes 2942 macroscopic images of 41 Brazilian forest species and is freely avail-

able online.

2.12. MacroHOLZdata

MacroHOLZdata [45], created in 2002 and made available for the first time in 2016,

is another interactive identification key with an integrated database for macroscopic wood

descriptions. Completely redesigned in 2022, it is available in German, English and Span-

ish, and includes 150 common hardwood and softwood commercial timbers. The database

is free of charge.

2.13. Softwoods

Softwoods [46] is a database that was made available from 2016 onwards. This inter-

active identification key with an integrated database focuses only on gymnosperms, as

the title suggests. It provides microscopic descriptions and illustrations of 53 taxa in the

world’s main forest regions. It is free on request.

2.14. Forest Species Classifier

Made available in 2018, Forest Species Classifier is the result of a master’s degree [47].

It is a user-friendly online database focusing on Brazilian forest species. It uses macro-

scopic [43] and microscopic [41] databases and includes microscopic images of 112 species

and macroscopic images of 41 species, with a total of 5182 images. It is freely available

online.

2.15. UTForest—UTFPR Classificador

This new version of the Forest Species Classifier platform [51] has been available

since 2021. It allows macroscopic identification of 44 native species of Brazil and includes

1318 images.

Forests 2022, 13, 2041 7 of 26

2.16. Charcoal

Developed in 2018, this database comprises charcoal samples of 44 Brazilian hard-

wood forest species, using 528 images [48,56]. It is available for research purposes only.

2.17. CharKey

This 2019 electronic identification key is described by the authors [49] as the first

computer-aided identification key designed for charcoals from French Guiana. It uses

SEM photographs to illustrate the anatomical features of 507 species belonging to 274 gen-

era and 71 families. Most of the descriptions were taken from Détienne et al. [57], and

follow the IAWA list of microscopic features for hardwood identification [53]. The key

contains 289 “items”, and its main aim is to identify specimens to the genus level. It is

freely available online.

2.18. Softwood Retrieval System (SRS) for Coniferous Wood

The Softwood Retrieval System (SRS) for Coniferous Wood [50] is an online identifi-

cation key with descriptions and micrographs of 180 Chinese coniferous wood species

(155 species with descriptions and microphotographs and 25 species with only micropho-

tographs) from nine families and more than 1000 images showing anatomical details. The

system is searchable by an interactive multiple-entry key. The microphotographs were

collected from slices of 115 coniferous species provided by the Wood Collection of the

Chinese Academy of Forestry (Beijing, CAFw) and 40 coniferous species from the Herbar-

ium of Southwest Forestry University (Kunming, SWFUw). The descriptions use features

from the IAWA List of Microscopic Features for Softwood Identification [58]. The system

supports three retrieval methods for coniferous wood retrieval: species name, anatomical

characteristics, and microscopic anatomical images (in test).

2.19. Mader App

This mobile app is under development. Its goal is to contribute to the global wood

identification effort and the fight against illegal logging using AI [52]. The project com-

prises 26 species, with a vast image database of 1000 images per species, aiming to obtain

maximum intraspecific variability for each species. The images were taken using a porta-

ble microscope and the app aims to obtain real-time recognition of samples. Preliminary

data from the authors indicate that accuracy is 95%. The authors intend to make the app

available soon on the Play Store, and the database used will be freely available for neural

network training [52].

3. Computer Vision-Based Wood Identification

The digital systems described above are the foundation of the systems which, despite

their limitations, are currently used to identify wood, mostly based on computer vision

technology. They are applicable to several fields of research and industry, including neu-

robiology, autonomous vehicles, and facial recognition. Computer vision systems process

visual data from any given image or video to extract the required and relevant features to

make a decision [59].

This image recognition ability, also known as image classification, is one of the most

important research areas in AI and is most frequently based on supervised learning. In

this case, the network is required to create a model that learns from labelled images to

determine classification rules, then it classifies the input data based on these same rules

(generally used for image classification). In the case of unsupervised learning, it is the

model that obtains unknown information through unlabelled data (generally used for im-

age clustering) [60].

Machine learning can also decide what to do without human assistance from the data

recognised by computer vision (input data), using predesigned algorithms [61,62]. This

Forests 2022, 13, 2041 8 of 26

removes the need to teach the model the necessary features or procedures for wood iden-

tification [63].

Computer vision technology is very appealing to many researchers because of its

verifiable potential for field application [64] and proven ability to recognise and quantify

wood structure variations that are not easily discernible using strictly “human” analysis.

It is also an affordable resource [65] and, therefore, scalable. However, for the software to

correctly interpret the specific architecture structure of the samples analysed to such a

high level of precision, reference material must be constantly entered into the image da-

tabase so that it can recognise natural variations in wood structure [12].

Computer vision-based wood identification is the real-world application of combin-

ing two types of software with different approaches within AI [66,67].

Figure 1 shows a pipeline of this method.

Figure 1. General scheme of machine learning method for image classification (based on [63,68]).

3.1. Machine Learning

Machine learning operates primarily as software that recognises patterns from input

images that are processed to define a descriptive structure to which the unknown image

will be referenced [69]. This involves various stages, as follows.

3.2. Image Acquisition

The most frequently used types of image are macroscopic images (obtained without

magnification using a normal digital camera) [70–73], stereograms (stereoscopic images

obtained with hand lens magnification, ca. 10×) [66,67,74–76], micrographs (optical micro-

scopic images) [77–79], SEM images (up to 10,000×) [80], and X-ray computed tomography

(CT) images [26,81].

Light control and uniformity are significant issues in image processing [67,82,83].

They include techniques that are used to filter and normalise image brightness [84–86].

3.3. Image Datasets

Image dataset construction or availability is one of the most significant factors among

the multiple issues that can affect the performance of computer vision-based wood iden-

tification systems.

The more extensive the dataset is, the more naturally occurring biological variations

within a species will be accessed and learned by the model. However, because construct-

ing a dataset of wood samples is such a difficult and time-consuming task, most studies

use wood collections for references [41,70,75,82,87–90].

This limitation is countered to some extent by initiatives such as ImageNet [91]. Aim-

ing to advance computer vision and deep learning research, the ImageNet dataset was

made freely available to researchers worldwide. It contains 14.2 million images across

more than 20,000 classes. A similar process is under way with herbaria digitalisation [92–

94]. However, despite the efforts made [42,79,88,95,96], the lack of free access to world-

wide wood image datasets continues to be the main constraint for computer vision-based

wood identification [63].

Table 2 shows the main currently available datasets that have useful data for com-

puter vision-based wood identification research.

Forests 2022, 13, 2041 9 of 26

Table 2. Wood image datasets available for computer vision-based wood identification research,

adapted from [63].

Dataset

Description

Image

Type

Number of

Species

Number of Images

Accessibility

Reference

CAIRO

Commercial hardwood species of Malaysia

Stereo

3700

Inaccessible

[97]

FRIM

5200

[98]

LignoIndo

Commercial hardwood species of Indonesia

809

4854

[99]

ZAFU WS 24

Wood species at Zhejiang A&F University

480

[76]

RMCA

Commercial wood species of Central Africa

Micro

1221

Open

[79]

XDD

Major Fagaceae species of Japan

2449

[88]

Lauraceae species of East Asia

1658

[95]

WOOD-AUTH

Wood species of Greece

Macro

4272

[96]

UFPR

Wood species of Brazil

2942

[44]

UFPR

Micro

112

2240

[42]

3.4. Image Processing

Machine learning comprises two independent procedures: feature processing, also

known as extraction (extraction of relevant features from input images), and classification

(learning extracted features and querying image classification). There is, however, a pre-

vious step to image processing.

Pre-processing aims to convert the image into data that a specific algorithm can use

to extract the required features, thus reducing computational complexity and facilitating

subsequent processing [100]. The techniques used for this include greyscale conversion

and image cropping [72,75,79,89,101,102], filtering [84,86], image sharpening [75,103] and

denoising [80,104,105].

Another important pre-processing procedure is data splitting, where the dataset is

split into subsets, most commonly training, validation, and test sets. Data splitting is ulti-

mately used to create a training set, a validation set and a test set in order to later evaluate

the model performance. To understand the reasons for these sets, one should think that

machine learning systems mimic the human learning process based on examples. From

this training set, the system will learn to generalise in order to correctly classify the im-

ages. The validation set is used to avoid the system learning the images “from memory”

during this generalization process. Finally, the test set is used to check the reliability of

the learning process. The use of these distinct data samples is one of the earliest pre-pro-

cessing steps needed to evaluate any model’s performance.

More specifically, the system will interpret the images extracted from the designated

training set as nothing more than a combination of pixels. Each pixel will have a specific

intensity represented by a number, and in this way a matrix of numbers is formed. Image

processing is based on extracting elements such as points, blobs, angles, corners and

edges, and the patterns they form. Variability in the anatomy of each wood species is rep-

resented as patterns of distinct pixel intensities, arrangement, distribution, and aggrega-

tion. The variations detected by computer vision will be learned by machine learning. This

process is the fundamental operating system of computer vision for all applications, in-

cluding wood identification [63].

After the extracted features have been learned, a classification model is established

by a classifier and a test set is formed to evaluate the system’s learning. The images are

then input, allowing the classification model to complete the identification through feed-

back of the predicted classes of each image [63].

Computer vision detects and “sees” the input image using multiple feature extraction

algorithms, while machine learning selects the types of features to be extracted, in most

cases texture and local features.

Texture features work with the combination and arrangement of image elements

(pixel intensities and resulting patterns) [66,75,97,103,106]. The most frequently used tech-

niques are grey level co-occurrence matrix (GLCM), grey level aura matrix (GLAM), local

Forests 2022, 13, 2041 10 of 26

binary pattern (LBP), higher local order autocorrelation (HLAC), and Gabor filter-based

features (GFBF). Despite the individual capabilities of each technique, texture fusion of

different types of texture features has shown superior classification accuracy

[102,107,108].

Local features differ from texture features by not describing an image as a unit, but

by describing significant and important specific features (keypoints) such as edges, cor-

ners or points. The most frequently used algorithms are scale-invariant feature transform

(SIFT), speeded up robust features (SURF), oriented features-from-accelerated-segment-

test (FAST) and rotated binary-robust-independent-elementary-feature (BRIEF) (ORB),

and Accelerated-KAZE (AKAZE).

Beyond features typology, factors such as dimensionality reduction and feature se-

lection are also important, as a large number of features extracted from an image can sub-

stantially reduce the computational efficiency of classification models. To achieve this bal-

ance, methods such as R AutoEncoder [109], principal component analysis (PCA), linear

discriminant analysis (LDA), and genetic algorithms (GA) are used for dimension reduc-

tion of data sets [110].

Another important element is the classification models created to learn the extracted

features and establish classification rules. The most frequently used classifiers are k-near-

est neighbours (k-NN), support vector machines (SVM), and artificial neural networks

(ANN) [111,112]. These classification procedures can be executed either in on-site hard-

ware [113] or on a cloud-based interface [90].

Machines can be easily misled by factors such as the source of images, which can be

acquired in the field using mobile phones [90] or in a laboratory-controlled environment

[71,114], and variables including different thicknesses, orientation, staining, digital arte-

facts and other variations, which is why many thin sections from historical wood collec-

tions are useful only to the trained human eye [10].

4. Deep Learning

Deep learning is among the most notable and promising of the many branches of

machine learning research.

As a neural network that attempts to simulate the function, structure and behaviour

of the human brain (Figure 2), it has the capacity to process and “learn” large amounts of

data [115,116].

Figure 2. General pipeline of deep learning models for image classification (based on [63,68]).

Its multiple different architectures include ANN [117], deep neural networks (DNN)

[118], recurrent neural networks (RNN) [119], deep reinforcement learning (DRL) [120],

and convolutional neural networks (CNN) [121]. The fields to which these have been ap-

plied are so vast that they are very difficult to summarise, but they include computer vi-

sion [122], forensic research [123], climate science [124], machine translation [125], classic

literature [126] and bioinformatics [127], to name just a few.

Among these multiple architectures, it is mostly ANNs and CNNs that are applied

to wood characterisation and identification.

Table 3 summarises this research and the applications of deep learning technologies.

Forests 2022, 13, 2041 11 of 26

Table 3. Research and applications of deep learning technologies.

Reference

Database

Species Geo-

graphic Origin

Image Type

Section Type

Number of

Species

Anatomically

Similar Species

Number of Im-

ages

Image

Analysis

Program

Pre-Pro-

cessing Im-

ages

Features De-

scriptor

Classifiers

CNN

Model

Classification Accu-

racy

[128]

Samples from natural

forests

Canary Islands

(Spain)

Biometric data

Transverse

Tangential

Radial

Yes

n.a.

WinCell

PCA

n.a.

Feedforward mul-

tilayer perceptron

network

n.a.

92.0%

[80]

No specific source

Global

SEM

Transverse

101

n.a.

LDA

GLCM

Multiple classifica-

tion methods

Hardwoods vs. soft-

woods

(LOOCV) 89%,

(EVT) 93%.

7 species

(LOOCV) 81%,

(EVT) 80%.

[129]

Samples from natural

forests

Turkey

Micro

Transverse

Tangential

Radial

Yes

n.a.

SVM

with linear kernel

function

n.a.

95.2%

[78]

LWA-UFP

Brazil

Micro

n.a.

112

2240

n.a.

GLCM

LBP

SVM

LBP

n.a.

SVM—98.6%

LBP—86%

[130]

LWA-UFP

Brazil

Macro

Transverse

2050

n.a.

LBP

CLBP

Colour-based

features

SVM

3-Con-

vNeta

95.77%

Micro

112

2240

LBP

GLCM

LPQ

LPQ + GLCM

SVM

97.32%

[72]

LWA-UFP

Brazil

Macro

Transverse

2942

n.a.

Texture fusion

strategy

Two-level divide-

and-conquer

n.a.

97.77%

[131]

Samples from natural

forests

Iberian Penin-

sula

Biometric data

Transverse

Tangential

Radial

Yes

n.a.

WinCell

n.a.

Resilient back-

propagation

algorithm

Feedforward mul-

tilayer perceptron

network

n.a.

81.2%

[132]

Samples from the tim-

ber industry

Korea

Macro

Transverse

Yes

33.730

n.a.

LeNet3

99.3%

[79]

TXWD-RMCA

Democratic Re-

public of the

Congo

Micro

Transverse

1221

CellB (ver-

sion 3.2,

Olympus)

GSC

LPQ

LBP

LDA

KNN

n.a.

88% species level

89% genus level

90% family level

Forests 2022, 13, 2041 12 of 26

[87]

LWA-UFP

Brazil

Macro

Transverse

47.024

n.a.

IMAGE

DATA GEN-

ERATOR

class of

KERAS

n.a.

Resnet50

98.3%

[113]

SJRw;

MADw;

RBw

Central and

South America,

Africa

Stereo

Transverse

Yes

2303

n.a.

End-to-end

trained image clas-

sifiers

VGG16

87.4%–97.5%

[48]

LWA-UFP

Brazil

Micro

Transverse

528

n.a.

LBP

SVM

n.a.

93.9%

Inception_v3

n.a.

95.7%

[133]

LWA-UFP

Brazil

Macro

Transverse

2942

n.a.

Scale dataset 100%

Micro

112

2240

Macroscopic dataset

98.73%

Microscopic dataset

99.11%

[134]

Samples from the tim-

ber industry

Korea

Macro

Transverse

Tangential

Yes

33,815

n.a.

Multiple

models

ensemble

98%

[135]

FFPRI

Japan

NIR-HSI

Tangential

Radial & between

the two planes

n.a.

PCA

90.5%

[136]

SJRw;

MADw;

RBw

Central and

South America

Micro

Transverse

Tangential

Radial

Yes

n.a.

SVM

n.a.

91.4%

[137]

TXWD-RMCA;

HNM;

MADw;

CM, Inc.;

OSU;

Central America

and Central Af-

rica

DART-

TOFMS

Slivers for metabo-

lome profiling

n.a.

Binning

threshold

n.a.

82.2%

[74]

Samples from natural

forests

Amazonia At-

lantic region

Macro

Transverse

2000

n.a.

Adapthisteq

GLCM

SVM

n.a.

97.7%

[67]

TXWD-RMCA;

HNM;

MADw;

CM, Inc.;

OSU;

Central America

and Central Af-

rica

Stereo

Transverse

n.a.

ResNet34

Species level

81.9%

Genus level

96.1%

[82]

DSB-FWRC

North America

Stereo

Transverse

1869

n.a.

Incep-

tionV4_Re

sNetV2

92.6%

[138]

2012 ImageNet

Brazil

Stereo

Transverse

281

n.a.

RiLPQ

kNN

DenseNet

98.8%

Forests 2022, 13, 2041 13 of 26

Tangential

Radial

[70]

CVLO-CELOS

Suriname

Macro

Transverse

1.2 million

n.a.

Threshold

n.a.

Inception-

98%

[77]

Neotropical

regions

Micro

Transverse

112

2240

n.a.

LBP

SVM

ResNet10

95.6%

[139]

Samples from Lumber

yard

North America

Macro

Tangencial

3158

n.a.

SGD optimizer

Adam optimizer

ResNet-50

Dense-

Net-121

Mo-

bileNet-

98.2%

[140]

GACD

Global

XRF

Tangential

Radial & in be-

tween the two

planes

n.a.

1D CNN

model

99%

[73]

Collected

from trunks of leafing

trees

Europe

Macro

Transverse

n.a.

Train set

n.a.

Residual

convolu-

tional en-

coder net-

work

Wood patch classifi-

cation—93%

Wood core classifi-

cation—98.7%

[141]

TXWD-RMCA

Democratic Re-

public of the

Congo

Micro

Transverse

Tangential

Radial

n.a.

GSC

LPQ

MVRF

n.a.

95%

n.a. not available. Institutions: CM, Inc.—Carlton McLendon, Inc.; CVLO-CELOS—Centrum voor Landbouwkundig Onderzoek/Centre for Agricultural Research

in Suriname; DSB-FWRC—Department of Sustainable Bioproducts/Forest and Wildlife Research Center; FFPRI—Forestry and Forest Products Research Institute

(Japan); GACD—Garman Art Conservation Department, SUNY—Buffalo State; HNM—IB—Herbario Nacional de México; Instituto de Biología; LWA-UFP—

Laboratory of Wood Anatomy at the Federal University of Parana; MADw—USDA Forest Products Laboratory Wood Collection of Madison, Wisconsin; OSU—

Oregon State University; PV—Private vendor; RBw—Botanic Garden of Rio de Janeiro, Brazil; SJRw—Samuel J. Record Collection; TXWD-RMCA—Tervuren

Xylarium Wood Database—Royal Museum for Central Africa. Models & Techniques: DART-TOFMS—Direct Analysis in Real Time, Time of Flight Mass Spec-

trometry; EVT—External validation test; GLCM—Grey level co-occurrence matrix; GSC—Grey scale conversion; kNN—k-nearest neighbours; LDA—linear dis-

criminant analysis; LOOCV—Leave-one-out-cross-validation; MVRF—Multi-view random forest; PCA—Principal component analysis; RF—Random Forests;

RiLPQ—Rotation Invariant Local Phase Quantisation; SGD optimizer—Stochastic gradient descent; SVM—Support vector machine; TL—Transfer learning; XRF—

X-ray fluorescence spectrometry.

Forests 2022, 13, 2041 14 of 26

4.1. Artificial Neural Networks (ANN)

Artificial neural networks are not only one of the main investigation methods, but

also constitute the foundation of deep learning [63]. These mathematical structures in-

spired by biological neural networks are a form of supervised or unsupervised learning

that show high ability to learn from examples given to them and extrapolate the infor-

mation when applied to future non-identified samples. This ability to reproduce, model

and “learn” nonlinear processes has given ANNs widespread applications in multiple

disciplines [79,117].

In the field of wood differentiation and identification, examples of research applying

ANNs include:

- Esteban et al. [128] used a feedforward multilayer perceptron (MLP) network, which

uses a similar structure to ANN to distinguish between Juniperus cedrus and J. phoe-

nicea var. canariensis, obtaining a 92% probability of correctly differentiating the spe-

cies;

- Mallik et al. [80] applied SEM to wood cross sections with 1500× magnification to

obtain species-level identification through the shape, number, area and distribution

of earlywood tracheids, processed by image segmentation, object recognition and

statistical methods. Their results showed that when distinguishing between hard-

woods and softwoods, a 0.89 accuracy was obtained using leave-one-out cross-vali-

dation and 0.93 using an external validation test (EVT), and when differentiating

seven wood species, they obtained a 0.81 accuracy using one-leave-out cross-valida-

tion and 0.80 using an EVT;

- The same microscopic features analysis was applied by Martins et al. [78], who used

microscopic transverse sections applying local phase quantisation (LPQ), local binary

patterns (LBP) and grey-level co-occurrence matrix (GLOM) to identify Brazilian spe-

cies. The process was applied to 112 species, 85 genera and 30 families, obtaining a

recognition rate of 98.6% for differentiation of hardwoods and softwoods and 86%

for discrimination of the 112 species;

- Turhan [129] used the SVM as a machine learning algorithm to differentiate Salix alba,

S. caprea and S. eleagnos, obtaining a 95.2% success rate;

- Filho et al. [72] used a two-level divide-and-conquer classification strategy to differ-

entiate 41 species of Brazilian flora, obtaining the highest accuracy level, of 97.77%;

- Esteban et al. [131] used a multilayer perceptron (MP) to differentiate Pinus sylvestris

L. and P. nigra Arn subsp. salzmannii (Dunal) Franco, obtaining 81.2% accuracy in the

testing set;

- Silva et al. [79] used microscopic images of cross sections of 77 commercial wood

species from the Democratic Republic of the Congo for surface texture analysis, re-

porting 88% successful identifications at species level, 89% at genus level and 90% at

family level.

- He et al. [136] applied machine learning classifiers SVM, Naive Bayes (NB), Decision

Tree C5.0 and ANN) to discriminate between Swietenia macrophylla King, S. mahagoni

(L.) Jacq and S. humilis Zucc. The best results were obtained with SVM, with an over-

all accuracy of 91.4%;

- Deklerck et al. [137] used machine learning not for image-based data processing, but

for metabolome profile obtained through direct analysis in real-time (DART™) ioni-

sation combined with time-of-flight mass spectrometry (TOFMS) to study the heart-

wood of 175 samples of 10 species of the Meliaceae family. Combining these tech-

niques resulted in accuracy levels of 82.2%;

- de Andrade et al. [74] generated 2000 macroscopic images of 21 species using a

smartphone and samples manually polished with a knife to replicate field conditions.

A grey level co-occurrence matrix for the development of classifiers based on SVM

was used, resulting in accuracies of 97.7%;

Forests 2022, 13, 2041 15 of 26

- Silva et al. [141] used 77 Congolese wood species as a reference base for applying a

multi-view random forest (MVRF) model for species-level identification. To ensure

information was not missed, the authors used images of the three anatomical planes.

The results showed that the concatenation of features from the transverse and tan-

gential planes clearly outperforms transverse-only analysis, while adding the radial

plane minimally improves the results obtained. The use of the MVRF model outper-

formed concatenation of LPQ features. The results showed that the supplementary

information added using three planes analysis and the model type considerably im-

prove the final results. Moreover, when evaluating the performance of the systems

developed, using the k-fold cross-validation scheme could have led to overestima-

tion of the results, so the authors applied a leave-k-tree-out approach during cross-

validation. The results showed that implementing this approach dramatically de-

creased accuracy compared with traditional cross-validation schemes.

4.2. Convolutional Neural Networks (CNN)

Convolutional neural networks are one of the most significant applications of ANNs.

In the AI context, a CNN is a class of feedforward ANN that has been successfully applied

to digital image processing analysis.

A CNN processes images more effectively by applying filtering techniques to ANNs

[116]. This is a powerful and accurate way of solving classification problems, and CNNs

are mainly credited for their role in image analysis, recognition, and classification. The

architecture of a CNN typically has multiple layers between input and output: three con-

volutional layers, a pooling layer and a fully connected layer. These layers process differ-

ent tasks during the image’s course. As the images progress through the distinct layers,

features such as edges, colours and shapes are extracted and interpreted. These features

are then learned and classified by the deep neural network, resulting ultimately in the

network’s ability to identify a specific object [63,116,142]. Other advantages are the capac-

ity of automatically recognise important features without human supervision.

CNNs have difficulty dealing with variance in the data presented, as tilted or rotated

images. This results in a limitation to encode an object’s orientation and position or pro-

cess spatially invariant data.

Research examples applied to wood identification include:

- Hafemann et al. [130] applied the CNN model 3-ConvNeta to identify macro images

of 41 species and micro images of 112 species. The results obtained 95.77% accuracy

for macroscopic images and 97.32% accuracy for microscopic images;

- Kwon et al. [132] applied six LeNet and MiniVGGNet CNN models to identify five

Korean softwood species (Cryptomeria japonica, Chamaecyparis obtuse, Pinus koraiensis,

P. densiflora, Larix kaempferi), using an iPhone 7 camera to obtain macroscopic images

of rough sawn surfaces from cross sections. Of all the CNN models tested, LeNet3

achieved the highest results and stability, with two extra layers added to the original

LeNet architecture. The identification accuracy obtained was 99.3%. The authors re-

ported that the software weight of the CNN created is small enough for installation

on a mobile device such as a smartphone;

- Maintaining the objective of ensuring field applicability, Kwon et al. [134] acknowl-

edged the real-world limitations of not including longitudinal wood surfaces. Using

mobile device cameras to obtain macroscopic images, they applied a combination of

models, obtaining the best results with LeNet2, LeNet3 and MiniVGGNet4. Their re-

sults showed an overall accuracy of 98% and an improvement on their earlier study,

particularly in the case of P. koraiensis and P. densiflora;

- Figueroa-Mata et al. [87] applied deep convolutional networks for identification of

41 Brazilian forest species from xylotheque samples at species level, achieving an ac-

curacy of 98.3%;

Forests 2022, 13, 2041 16 of 26

- Ravindran et al. [113] used CNNs to identify 10 neotropical species in the Meliaceae

family (Cabralea canjerana, Carapa guianensis, Guarea glabra, G. grandifolia, Khaya ivoren-

sis, K. senegalensis, and the CITES-listed Swietenia macrophylla, S. mahagoni, Cedrela fis-

silis, and C. odorata), using only the transverse surface. The results showed an accu-

racy of 87.4 to 97.5%;

- To develop an automatic classification system for charcoal, Maruyama et al. [48] ap-

plied two LBP configurations of as texture descriptors. As state-of-the-art machine

learning classifiers, SVM and random forests (RF) have shown the best results. Incep-

tion_v3 CNN was applied for representation learning evaluation. The database com-

prised 44 charcoal samples from Brazilian native species from natural forests. The

authors reported that both handcrafted features and RL achieved results of around

95% recognition rate;

- Oliveira et al. [133] used databases developed by Filho et al. [72] and Martins et al.

[78] to access cross sections of 2942 wood macroscopic images of 41 species and 2240

microscopic images of 112 species, applying CNNs to create three models. Based on

the results, the authors reported 100% recognition accuracy for the scale model,

98.73% for the macroscopic model, and 99.11% for the microscopic model;

- Kanayama et al. [135] applied a deep CNN approach to near-infrared hyperspectral

imaging (NIR-HSI) using a principal component (PC) algorithm to identify 120 sam-

ples of 38 hardwood species. The results obtained showed 90.5% accuracy;

- A CNN was also used by Ravindran and Wiedenhoeft [67] to compare the macro-

scopic field identification programme XyloTron, using an ImageNet pre-trained Res-

Net34 CNN, with mass spectrometry to differentiate 10 Meliaceae species used by

Deklerck et al. [137]. The results showed identification accuracy of 81.9% at the spe-

cies level and 96.1% at the genus level compared to 74.9% and 91.4%, respectively, in

the work by Deklerck et al. [137];

- Lopes et al. [82] applied the InceptionV4_ResNetV2 CNN to analyse macroscopic im-

ages of the end-grain of 10 xylarium North American hardwood species, producing

1869 images using a smartphone fitted with a 14× macro lens. Their results showed

an accuracy of 92.6%;

- de Geus et al. [138] applied the DenseNet CNN to recognise 281 species, using the

largest dataset of microscopic transverse, radial and tangential images available at

the time. Rotation invariant LPQ (RiLPQ) showed the best results of the feature de-

scriptors used. The authors reported an identification accuracy of 98.8%;

- Olschofsky and Köhl [70] applied Inception-v3, an image classification model using

a CNN for feature recognition and classification, pre-trained with 1.2 million images.

The CITES-protected species Cedrella odorata was chosen and compared with 13 other

tropical tree species for recognition. The results with the pre-trained CNNs had 98%

accuracy, but when other tree species not used for training were added, the classifi-

cation accuracy fell to 87%;

- The ResNet101 CNN, associated with an SVM as classifier, was applied by Lens et al.

[77] to species-level identification of 112 mainly neotropical tree species, using only

transverse sections but focusing on microscopic rather than macroscopic analysis.

The results showed successful identification in 95.6% of cases;

- Wu et al. [139] applied deep convolutional neural networks (CNNs) for the identifi-

cation of 11 rough saw hardwood North American species based on tangential plane

images only. CNNs ResNet-50, DenseNet-121, as well as MobileNet-V2 were tested,

resulting in an overall accuracy of 98.2%.

- Shugar et al. [140] combined X-ray fluorescence spectrometry (XRF) and a CNN to

identify 48 wood specimens of both hardwoods and softwoods, mostly from heart-

wood and using either tangential or radial sections. They reported 99% identification

accuracy from the 66 datasets;

- In the study by Fabijańska et al. [73], a CNN with residual connections was tested to

identify 312 wood core scanned images of 14 European softwood and hardwood tree

Forests 2022, 13, 2041 17 of 26

species, developing a wood patch classification and a wood core classification. The

results showed that the proposed model correctly recognised patch images in 93% of

cases and wood core images in 98.7%. Comparison of the results also showed that

this model outperformed the state-of-the-art convolutional neural network-based

model.

4.3. Generative Adversarial Networks (GANs)

Within deep learning, GANs [143] are described as neural networks that can learn to

generate realistic samples from the data on which they were trained.

They use a neural network as a generator that takes a random distribution of data as

input and learns to map that information to output the desired distribution of data. A

second neural network, known as a discriminator (a binary classifier), will use the input

and output images to determine the probability of the image originating as a training im-

age (real) or on the generator (fake), thus assessing the most likely class to which the out-

put image belongs [144].

Generative adversarial networks can produce highly realistic images using CNNs in

an unsupervised manner [145]. Their application extends to multiple fields of scientific

research, but they remain poorly explored in wood sciences [146–148].

- Addressing the possibility of eliminating economic and processing burdens in ac-

quiring images of worldwide wood species for machine-learning training purposes,

Lopes et al. [145] accessed 119 hardwood species references on the publicly available

Xylarium Digital Database [88]. Applying a style-based GAN, they successfully gen-

erated highly realistic and anatomically meaningful synthetic microscopic cross-sec-

tional images of hardwood species which they reported as virtually indistinguishable

from real cross-sectional images.

- To evaluate the resemblance, quality and pattern evaluation between the synthetic

and real cross sections, a structural similarity index measure (SSIM) and Fréchet in-

ception distance (FID) were applied and a visual Turing test (VTT) was performed

by wood anatomists to confirm the usefulness and realism of the GAN-generated

images. The results showed that the artificially generated images were indistinguish-

able from real microscopic cross-sectional images.

- The authors [145] reported that it is even feasible to generate synthetic hybrids based

on microscopic cross-sectional images from two parental species. This would have

considerable implications on wood science and technology, especially for estimating

the wood permeability, strength, density, or hydraulic potential, for example, of a

species that has not even been planted.

5. Field Applicable Wood Identification Systems

One of the most interesting features of computer vision-based wood identification

systems is their field application capability. Despite the consensus that it will be a long

time before this technology becomes readily available not only to researchers and law en-

forcement bodies, but also the general public, it is evident that this goal is reachable. Pro-

grammes already developed or under development to respond to field application needs

include:

5.1. MyWood-ID

Described as an automated wood identification mobile app [90,149], MyWood-ID

uses a smartphone with a retrofitted macro lens and machine vision for macroscopic wood

identification. The system uses a database of 20 species of timber native to Malaysia and

provides a simple and effective way to acquire macroscopic wood digital images. The im-

ages are then uploaded to a cloud server via an internet connection for immediate identi-

fication results. It is intended to be cost-effective, easily accessible and intuitive, and to

provide fast results.

Forests 2022, 13, 2041 18 of 26

These characteristics are evident when compared with other field deployable systems

[65,75]. As differentiating features of their wood identification system, the authors cite its

portability, lower initial cost, faster field deployment time, intuitive use, and continual

online database update. However, it requires a constant internet connection for results

and is operating-system-dependent (running only on iPhone 6 and 7). The main limitation

of this system is the lack of consistent light control for wood image acquisition, although

the authors indicate that this can be mitigated using the learning capability of a deep

learning algorithm. The results achieved by this system have an accuracy of 96% to 98%.

It is a paid app.

5.2. MyWood-Premium

This is an update of the previous app, developed by FRIM (Wood Anatomy Lab of

Forest Research Institute Malaysia) and UTAR (Universiti Tunku Abdul Rahman) [150].

The updated version comprises a database of 100 wood species native to Malaysia. It is

available only on iPhone, iPod touch and Mac, and requires iOS 8.0 or later. It recom-

mends the Ollo-clip™ Macro Lens with 21× magnification for optimum performance. It is

a free app.

5.3. Xylorix

Another recent approach to rapid field wood identification is Xylorix [151,152], a

platform that combines a suite of apps, tools and services. Xylorix Inspector is a wood

identification mobile app that uses macroscopic features for automated identification. It

is based on trained AI models to automatically identify the wood genus or species. How-

ever, a Xylorix WIDK-24X01 illuminated macro lens must be attached to the mobile phone

camera for correct performance. It is available on either Apple iOS or Android operating

systems and is supported by most mobile phones. Of the 24 species in the system database,

11 are free and the other 13 are paid.

5.4. XyloTron

XyloTron is a paid, open-source, image-based macroscopic field identification pro-

gramme designed for wood and charcoal identification [65,153]. It features adjustable and

controlled visible light, UV illumination capacity, and all the necessary software to control

the device, capture images, and deploy the trained classification models.

It works by capturing high-quality images of wood or charcoal samples with visible

or UV light. The identification accuracy for wood is described as 97.7%, increasing with

the use of UV light to 99.1% (e.g., identification confusion between Albizia sp., fluorescent,

and Inga sp., not fluorescent) and 98.7% for charcoal.

One limitation is that it is not a simple or easily deployable on-site system to use,

because it requires a permanent connection to a laptop computer.

Ravindran and Wiedenhoeft [67] compared the performance of XyloTron and MS for

species- and genus-level identification of 10 species of Meliaceae. The results showed a

similar species-level accuracy of the XyloTron and MS models, but higher genus-level ac-

curacy with XyloTron [67].

5.5. XyloPhone

To overcome visual aberrations (field distortion and spherical aberration), uncon-

trolled light sources, high prices, and a lack of real field applicability, Wiedenhoeft [83]

proposed the XyloPhone. Described as an open-source, 3D-printed imaging attachment

adaptable to virtually any smartphone for macroscopic image capture, it is a small, closed

plastic box that provides a fixed focal distance, exclusion of ambient light, and a choice of

visible or UV illumination. It is powered by a rechargeable external battery and a com-

mercially available lens, making it affordable and, according to the author, providing

comparable image quality to XyloTron.

Forests 2022, 13, 2041 19 of 26

To document features such as evenness of illumination, distortion, maximum reso-

lution, and spherical aberration, the author compared the Xylophone + iPhone (XPi), the

XyloPhone + Samsung (XPs), the Ollo Clip 14× + iPhone (OCi), and the Xylorix + iPhone,

with two distinct configurations. He reported that XyloPhone’s optical performance, es-

pecially when used with more recent smartphones, is clearly superior to the lenses/light-

ing arrays of other systems [83].

5.6. WIDER

WIDER is a battery-charged portable system that uses spectroscopy measurement

and machine-learning-based identification software. It comprises a database of 15 species

and the authors [154] reported accuracy results of 95%. It is part of a larger project that

was completed in 2021 and brought into use by USAID PEER Cycle 8 (Development of

Wood Identification System and Timber Tracking Database to Support Legal Trade). The

same project developed the ECVT 4D Dynamic [154] technology for monitoring tree phys-

iological processes.

5.7. IMAIapp

IMAIapp [155] is a wood identification mobile app that uses a lens attached to a

smartphone. Its purpose is to use convolutional neural networks (CNN) capable of carry-

ing out timber identification through machine learning of macroscopic elements observed

in photo enlargements. The difficulty of the problem lies in the number of classes that the

method is required to recognise automatically, from a total of 400 wood species and a high

number of macroscopic images. EfficientNet architecture is enhanced by a novel approach

for pre-processing that combines computer vision and data augmentation techniques ap-

plied to the original dataset. The use of classification models based on deep learning is the

leading technique with the best performance at present, and the innovative approach to

increase the quality of the training data makes the model integrated in IMAIapp robust to

rotation, illumination and zoom invariants. This means the app can be used in the field.

Using TensorFlow Lite libraries for Apple and Google platforms, the application works

standalone, is 100% executable from the mobile device and does not require a connection

to the Internet. IMAIapp is therefore a design using an edge-computing method that is

intended to avoid computing constraints on the mobile device on which it is installed. The

project is under development and the app will be free for Android and iOS.

6. Discussion

In the last 100 years, what we now call the traditional wood identification method

based on anatomical descriptions has followed well defined, standardised features to suc-

cessfully distinguish and identify the multiple families, genera and species of angiosperm

and gymnosperm trees. However, despite the many positive aspects of this method, it is

now evident that it is reaching its limit.

The main limitations are the identification uncertainty at species level, the time-con-

suming methodology, the lack of anatomists with the necessary training for the task, and

the associated costs of these professionals, mainly when on-site identifications are re-

quired. These limitations have a notable impact on the type of monitoring that can be

carried out, e.g., in the fight against illegal logging. Faster, more accurate and economi-

cally scalable methods are urgently needed.

As a valuable response to this need, the results obtained so far by computer vision-

based identification (CVBI) of wood and, in particular, deep learning approaches, clearly

demonstrate that this method has enormous potential for wood identification and quan-

titative wood anatomy.

Despite the many obstacles remaining, this method is steadily adapting to overcome

limitations such as inter- and intra-anatomical variability, high anatomical resemblance,

Forests 2022, 13, 2041 20 of 26

non-homogeneous illumination, staining or deformed samples, and limited image data-

bases, among many other issues.

Of all the resources discussed, deep learning appears to be the most significant and

promising solution in AI developments, and CNN models applied to wood sciences are

one of the leading and most rapidly evolving systems. CNNs have exhibited a notably

more efficient capacity and accuracy for quantitative wood anatomy and feature recogni-

tion, alongside computer cost reduction.

Field-deployable identification systems appear to be the most important and impact-

ful option in computer vision-based wood identification. This resource, based on ad-

vances in communication technologies, will enable more prolific, increasingly accurate

and faster screening by authorities without human prejudice, particularly with regard to

illegal timber and charcoal trading.

Computer vision-based identification technology could become one of the most ef-

fective and unavoidable weapons in the fight against the illegal timber and charcoal trade,

as it enables individuals who are untrained in traditional identification to obtain highly

accurate and legally binding identifications on the spot.

The multiple future contributions of the technologies underpinning CVBI for wood

sciences are difficult to fully envision at present. However, it is of utmost importance to

overcome or at least mitigate the limitations that are severely hampering the development

and implementation of these systems.

Two of the most pressing issues are the limited number of digital databases, which

are specific to geographically restricted areas/species or inaccessible to the global research

community, and the lack of extensive field testing and verification hindering the accuracy

quantification of the systems.

The most urgent actions required are the construction of a freely accessible global

digital wood image database, the availability of this tool in a cloud-based system for ac-

cess everywhere, by everyone, and priority inclusion of CITES-listed species and their

look-a-likes.

Author Contributions: Conceptualisation, writing and original draft preparation J.L.S.; review and

editing R.B., J.P., P.d.P. All authors have read and agreed to the published version of the manuscript.

Funding: I would like to acknowledge the Portuguese Fundação para a Ciência e Tecnologia (FCT)

for the PhD grant (UI/BD/151009/2021) to José Luis Amorim Silva and for funding the Strategic

Projects UID/EAT/0622/2016 of CITAR—Centro de Investigação da Universidade Católica Escola

das Artes, Porto, Portugal.

Data Availability Statement: Data are contained within the article and are also available from the

corresponding author.

Acknowledgments: I would like to acknowledge Yong Haur Tay for his kind clarifications about

Xylorix, Pedro Luis Paula Filho for his kind clarifications about The Forest Species Database—Mac-

roscopic, Forest Species Classifier and UTForest—UTFPR Classificador, Hisashi Abe for his kind

clarifications about the Wood database of the Forestry and Forest Products Research Institute, and

Dra Cassiana Ferreira for her kind clarifications about the Mader app.

Conflicts of Interest: The authors declare no conflict of interest.

References

1. May, C. Transnational Crime and the Developing World; Global Financial Integrity: Washington, DC, USA, 2017.

2. Nellemann, C. Green Carbon, Black Trade: A Rapid Response Assessment on Illegal Logging, Tax Fraud and Laundering in the

World’s Tropical Forests. 2012. Available online: https://wedocs.unep.org/20.500.11822/8030 (accessed on 27 April 2022).

3. EU; EC. Forests. Available online: https://ec.europa.eu/environment/forests/illegal_logging.htm (accessed on 27 April 2022).

4. Bösch, M. Institutional quality, economic development and illegal logging: A quantitative cross-national analysis. Eur. J. For.

Res. 2021, 140, 1049–1064. https://doi.org/10.1007/s10342-021-01382-z.

5. UN. Convention on International Trade in Endangered Species of Wild Fauna and Flora. Available online:

https://cites.org/sites/default/files/eng/disc/CITES-Convention-EN.pdf (accessed on 27 April 2022).

Forests 2022, 13, 2041 21 of 26

6. E.U. European Union Timber Regulation. Available online: https://eur-lex.europa.eu/legal-con-

tent/EN/TXT/?uri=CELEX%3A32010R0995 (accessed on 27 April 2022).

7. European Union; Austrian Development Cooperation; The World Bank; IUCN.; WWF ENPI EAST FLEG—European Neigh-

borhood and Paternaship Instrument East Countries Forest Law Enforcement and Governance Program. Available online:

https://www.enpi-fleg.org/ (accessed on 6 June 2022).

8. FAOUN; UNDP. UNEP UN-REDD Programme. Available online: https://www.un-redd.org/ (accessed on 27 April 2022).

9. UNDER; UNEP. FAOUN Preventing, Halting and Reversing The Degradation Of Ecosystems Worldwide. Available online:

https://www.decadeonrestoration.org/ (accessed on 27 April 2022).

10. Schmitz, N.; Beeckman, H.; Blanc-Jolivet, C.; Boeschoten, L.E.; Braga, J.J.W.B.; Cabezas, J.A.; Chaix, G.; Crameri, S.; Degen, B.;

Deklerck, V.; et al. Overview of Current Practices in Data Analysis for Wood Identification. A Guide for the Different Timber Tracking

Methods; GTTN-European Forest Institute: Joensuu, Finland, 2020.

11. Schmitz, N.; Beeckman, H.; Cabezas, J.A.; Cervera, M.T.; Espinoza, E.; Fernandez-Golfin, J.; Gasson, P.; Hermanson, J.; Jaime

Arteaga, M.; Koch, G.; et al. The Timber Tracking Tool Infogram. Overview of Wood Identification Methods’ Capacity; Global Timber

Tracking Network, GTTN Secretariat, European Forest Institute and Thünen Institute: Joensuu, Finland, 2019.

12. Dormontt, E.E.; Boner, M.; Braun, B.; Breulmann, G.; Degen, B.; Espinoza, E.; Gardner, S.; Guillery, P.; Hermanson, J.C.; Koch,

G.; et al. Forensic timber identification: it’s time to integrate disciplines to combat illegal logging. Biol. Conserv. 2015, 191, 790–

798. https://doi.org/10.1016/j.biocon.2015.06.038.

13. United Nations. International Consortium on Combating Wildlife Crime (2016). Best Practice Guide for Forensic Timber Iden-

tification; United Nations Office on Drugs and Crime: New York, NY, USA, 2016.

14. ITTO. Biennial Review and Assessment of the World Timber Situation. Available online: https://www.itto.int/direct/topics/top-

ics_pdf_download/topics_id=6783&no=1 (accessed on 27 April 2022).

15. Interpol Illegal Logging in Latin America and Caribbean Inflicting Irreversible Damage-INTERPOL. Available online:

https://www.interpol.int/News-and-Events/News/2022/Illegal-logging-in-Latin-America-and-Caribbean-inflicting-irreversi-

ble-damage-INTERPOL (accessed on 27 April 2022).

16. Abe, H.; Watanabe, K.; Ishikawa, A.; Noshiro, S.; Fujii, T.; Iwasa, M.; Kaneko, H.; Wada, H. Simple separation of torreya nucifera

and chamaecyparis obtusa wood using portable visible and near-infrared spectrophotometry: Differences in light-conducting

properties. J. Wood Sci. 2016, 62, 210–212. https://doi.org/10.1007/s10086-016-1541-z.

17. Pace, J.H.C.; Latorraca, J.-V.D.F.; Hein, P.R.G.; de Carvalho, A.M.; Castro, J.P.; da Silva, C.-E.S. Wood species identification from

Atlantic forest by near infrared spectroscopy. For. Syst. 2019, 28, e015. https://doi.org/10.5424/fs/2019283-14558.

18. Snel, F.A.; Braga, J.W.B.; da Silva, D.; Wiedenhoeft, A.C.; Costa, A.; Soares, R.; Coradin, V.T.R.; Pastore, T.C.M. Potential field-

deployable NIRS identification of seven dalbergia species listed by CITES. Wood Sci. Technol. 2018, 52, 1411–1427.

https://doi.org/10.1007/s00226-018-1027-9.

19. Akhmetzyanov, L.; Copini, P.; Sass-Klaassen, U.; Schroeder, H.; de Groot, G.A.; Laros, I.; Daly, A. DNA of centuries-old timber

can reveal its origin. Sci. Rep. 2020, 10, 20316. https://doi.org/10.1038/s41598-020-77387-2.

20. Jiao, L.; Liu, X.; Jiang, X.; Yin, Y. Extraction and amplification of DNA from aged and archaeological populus euphratica wood

for species identification. Holzforschung 2015, 69, 925–931. https://doi.org/10.1515/hf-2014-0224.

21. Wagner, S.; Lagane, F.; Seguin-Orlando, A.; Schubert, M.; Leroy, T.; Guichoux, E.; Chancerel, E.; Bech-Hebelstrup, I.; Bernard,

V.; Billard, C.; et al. High-throughput DNA sequencing of ancient wood. Mol. Ecol. 2018, 27, 1138–1154.

https://doi.org/10.1111/mec.14514.

22. Carmona, R.J.; Wiemann, M.C.; Baas, P.; Barros, C.; Chavarria, G.D.; McClure, P.J.; Espinoza, E.O. Forensic identification of

CITES appendix I cupressaceae using anatomy and mass spectrometry. IAWA J. 2020, 41, 720–739.

https://doi.org/10.1163/22941932-bja10002.

23. Espinoza, E.O.; Wiemann, M.C.; Barajas-Morales, J.; Chavarria, G.D.; McClure, P.J. Forensic analysis of cites-protected dalbergia

timber from the americas. IAWA J. 2015, 36, 311–325. https://doi.org/10.1163/22941932-20150102.

24. Zhang, M.; Zhao, G.J.; Liu, B.; He, T.; Guo, J.; Jiang, X.; Yin, Y. Wood discrimination analyses of pterocarpus tinctorius and

endangered pterocarpus santalinus using DART-FTICR-MS coupled with multivariate statistics. IAWA J. 2019, 40, 58–74.

https://doi.org/10.1163/22941932-40190224.

25. Ge, Z.; Chen, L.; Luo, R.; Wang, Y.; Zhou, Y. The detection of structure in wood by X-ray CT imaging technique. BioResources

2018, 13, 3674–3685. https://doi.org/10.15376/biores.13.2.3674-3685.

26. Kobayashi, K.; Hwang, S.-W.; Okochi, T.; Lee, W.-H.; Sugiyama, J. Non-destructive method for wood identification using con-

ventional X-ray computed tomography data. J. Cult. Herit. 2019, 38, 88–93. https://doi.org/10.1016/j.culher.2019.02.001.

27. Tazuru, S.; Sugiyama, J. Wood identification of japanese shinto deity statues in matsunoo-taisha shrine in kyoto by synchrotron

X-ray microtomography and conventional microscopy methods. J. Wood Sci. 2019, 65, 60. https://doi.org/10.1186/s10086-019-

1840-2.

28. Wheeler, E.A.; LaPasha, C.A. A microcomputer based system for computer-aided wood identification. IAWA Bull. 1987, 8, 347–

354.

29. LaPasha, C.A. General unknown entry and search system. A program package for computer-assisted identification. Suppl. North

Carolina Agric. Resour. Serv. 1986, 474, 18.

30. Ilic, J. Computer aided wood identification using csiroid. IAWA J. 1993, 14, 333–340.

31. Dallwitz, M.J. A general system for coding taxonomic descriptions. Taxon 1980, 29, 41–46. https://doi.org/10.2307/1219595.

Forests 2022, 13, 2041 22 of 26

32. Richter, H.G.; Dallwitz, M.J. Commercial Timbers: Descriptions, Illustrations, Identification, and Information Retrieval. Avail-

able online: https://www.delta-intkey.com/wood/en/index.htm (accessed on 10 October 2021).

33. Heiss, A.G. Anatomy of European and North American Woods—An Interactive Identification Key. Available online:

http://www.holzanatomie.at/ (accessed on 9 February 2022).

34. Forestry & Forest Products Research Institute. Wood Database of the Forestry & Forest Products Research Institute. Available

online: https://db.ffpri.go.jp/WoodDB/index-E.html (accessed on 2 May 2022).

35. Wheeler, E.A. InsideWood. Available online: https://insidewood.lib.ncsu.edu/search;jsessionid=hYhHqrsAf-

kTKM8JGVm0e3WjZLOdRCfo3_1Y5k6Zq?0 (accessed on 9 January 2022).

36. Schoch, W.; Heller-Kellenberger, I.; Schweingruber, F.; Kienast, F.; Schmatz, D. Wood Anatomy of Central European Species.

Available online: http://www.woodanatomy.ch/authors.html (accessed on 3 May 2022).

37. Richter, H.G.; Gembruch, K.; Koch, G. CITESwoodID: Descriptions, Illustrations, Identification, and Information Retrieval.

Available online: https://www.delta-intkey.com/citeswood/index.htm (accessed on 9 February 2022).

38. Barker, J.A.; Flinders, B.A.H. Key to a Selection of Arid Australian Hardwoods & Softwoods. Available online: https://keys.lu-

cidcentral.org/keys/v3/arid/default_wip.htm (accessed on 4 January 2022).

39. Coradin, V.T.R.; Camargos, J.A.A.; Pastore, T.C.M.; Christo, A.G. Brazilian commercial timbers: Interactive identification key

based on general and macroscopic features madeiras comerciais do brasil: Chave interativa de identificação baseada em

caracteres gerais e macroscópicos. Available online: https://keys.lucidcentral.org/keys/v4/madeiras_comerciais_do_brasil/

index_en.html (accessed on 2 May 2022).

40. Sarmiento, C.; Détienne, P.; Heinz, C.; Molino, J.-F.; Grard, P.; Bonnet, P. Pl@ntwood: A computer-assisted identification tool

for 110 species of amazon trees based on wood anatomical features. IAWA J. 2011, 32, 221–232. https://doi.org/10.1163/22941932-

90000053.

41. Martins, J.; Oliveira, L.S.; Nisgoski, S.; Sabourin, R. The forest species database—Microscopy. Available online:

https://web.inf.ufpr.br/vri/databases/forest-species-database-microscopic/ (accessed on 3 May 2022).

42. UFPR Forest Species Database—Microscopic. Available online: https://web.inf.ufpr.br/vri/databases/forest-species-database-

microscopic/ (accessed on 21 April 2022).

43. Filho, P.L.P.; Oliveira, L.S.; Nisgoski, S.; Britto, A.S. The Forest Species Database—Macroscopic. Available online:

https://web.inf.ufpr.br/vri/databases/forest-species-database-macroscopic/ (accessed on 3 May 2022).

44. UFPR Forest Species Database—Macroscopic. Available online: https://web.inf.ufpr.br/vri/databases/forest-species-database-

macroscopic/ (accessed on 23 April 2021).

45. Richter, H.G.; Oelker, M.; Koch, G. MacroHOLZdata—Computer Aided Macroscopic Wood Identification and Information on

Properties and Utilization of Trade Timbers. CD-ROM. Available online: http://macroholzdata.appstor.io/ (accessed on 23 April

2021).

46. Richter, H.G.; Dallwitz, M.J. Commercial Timbers: Descriptions, Illustrations, Identification, and Information Retrieval. Avail-

able online: https://www.delta-intkey.com/wood/index.htm (accessed on 3 May 2022).

47. de Oliveira, W. Forest Species Classifier. Available online: http://reconhecimentoflorestal.md.utfpr.edu.br./#/pt/classificador

(accessed on 3 May 2022).

48. Maruyama, T.M.; Oliveira, L.S.; Britto, A.S.; Nisgoski, S. Automatic classification of native wood charcoal. Ecol. Inform. 2018, 46,

1–7. https://doi.org/10.1016/j.ecoinf.2018.05.008.

49. Bodin, S.C.; Scheel-Ybert, R.; Beauchêne, J.; Molino, J.-F.; Bremond, L. CharKey: An electronic identification key for wood char-

coals of French Guiana. IAWA J. 2019, 40, 75-S20. https://doi.org/10.1163/22941932-40190227.

50. EyeWood, S.F.U. Softwood Retrieval System for Coniferous Wood. Available online: http://woodlab.swfu.edu.cn/#/(accessed

on 23 April 2021).

51. Filho, P.L.d.P. UTForest—UTFPR Classificador. Available online: https://clb.lamia.sh.utfpr.edu.br/classification (accessed on 8

June 2022).

52. Ferreira, C.A.; Inga, J.G.; Vidal, O.D.; Goytendia, W.E.; Moya, S.M.; Centeno, T.B.; Vélez, A.; Gamarra, D.; Tomazello-Filho, M.

Identification of tree species from the peruvian tropical amazon “selva central” forests according to wood anatomy. BioResources

2021, 16, 7161–7179.

53. Wheeler, E.A.; Baas, P.; Gasson, P.E. IAWA List of microscopic features for hardwood identification: With an appendix on non-

anatomical information. IAWA Bull. 1989, 10, 219–332. https://doi.org/10.1163/22941932-90000496.

54. Schweingruber, F.H. Microscopic Wood Anatomy : Structural Variability of Stems and Twigs in Recent and Subfossil Woods

from Central Europe; Swiss Federal Institute for Forest: Birmensdorf, Swiss, 1990; ISBN 3905620022 9783905620023.

55. CITES. Available online: https://cites.org/eng (accessed on 2 May 2022).

56. Menon, L.T.; Laurensi, I.A.; Penna, M.C.; Oliveira, L.E.S.; Britto, A.S. Data augmentation and transfer learning applied to char-

coal image classification. In Proceedings of the 2019 International Conference on Systems, Signals and Image Processing (IWS-

SIP), Osijek, Croatia, 5–7 June 2019; pp. 69–74.

57. Détienne, P.; Jacquet, P.; Mariaux, A. Manuel d’identification des bois tropicaux. tome 3 : Guyane Française. In Manuel D’iden-

tification des Bois Tropicaux; CIRAD: Montpellier, France, 1982; Volume 3, p. 315, ISBN 139782876145962.

58. Richter, H.G.; Grosser, D.; Heinz, I.; Gasson, P.E. IAWA list of microscopic features for softwood identification. IAWA J. 2004,

25, 1–70.

Forests 2022, 13, 2041 23 of 26

59. Voulodimos, A.; Doulamis, N.; Doulamis, A.; Protopapadakis, E. Deep learning for computer vision: A brief review. Comput.

Intell. Neurosci. 2018, 2018, 7068349. https://doi.org/10.1155/2018/7068349.

60. Li, N.; Shepperd, M.; Guo, Y. A Systematic review of unsupervised learning techniques for software defect prediction. Inf. Softw.

Technol. 2020, 122, 106287. https://doi.org/10.1016/j.infsof.2020.106287.

61. Jordan, M.I.; Mitchell, T.M. Machine learning: Trends, perspectives, and prospects. Science 2015, 349, 255–260.

https://doi.org/10.1126/science.aaa8415.

62. Khan, A.I.; Al-Habsi, S. Machine learning in computer vision. Procedia Comput. Sci. 2020, 167, 1444–1451.

https://doi.org/10.1016/j.procs.2020.03.355.

63. Hwang, S.-W.; Sugiyama, J. Computer vision-based wood identification and its expansion and contribution potentials in wood

science: A review. Plant Methods 2021, 17, 47. https://doi.org/10.1186/s13007-021-00746-1.

64. Ravindran, P.; Ebanyenle, E.; Ebeheakey, A.A.; Abban, K.B.; Lambog, O.; Soares, R.; Costa, A.; Wiedenhoeft, A.C. Image based

identifcation of ghanaian timbers using the xylotron: Opportunities, risks and challenges. In Proceedings of the 33rd Conference

on Neural Information Processing Systems (NeurIPS 2019), Vancouver, BC, Canada, 9–14 December 2019.

65. Ravindran, P.; Thompson, B.J.; Soares, R.K.; Wiedenhoeft, A.C. The xylotron: Flexible, open-source, image-based macroscopic

field identification of wood products. Front. Plant Sci. 2020, 11, 1015. https://doi.org/10.3389/fpls.2020.01015.

66. Tou, J.Y.; Tou, P.; Lau, P.Y.; Tay, Y.H. Computer vision-based wood recognition system. In Proceedings of International Work-

shop on Advanced Image Technology. 2007. Available online:

https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=Computer+vision-based+wood+recognition+system.+In+Pro-

ceedings+of+International+Workshop+on+Advanced+Image+Technology%2C+2007&btnG= (accessed on 27 April 2022).

67. Ravindran, P.; Wiedenhoeft, A.C. Comparison of two forensic wood identification technologies for ten meliaceae woods: Com-

puter vision versus mass spectrometry. Wood Sci. Technol. 2020, 54, 1139–1150. https://doi.org/10.1007/s00226-020-01178-1.

68. Du, M.; Liu, N.; Hu, X. Techniques for interpretable machine learning. Commun. ACM 2019, 63, 68–77.

https://doi.org/10.1145/3359786.

69. Martins, A.L.R.; Marcal, A.R.S.; Pissarra, J. Modified DBSCAN algorithm for microscopic image analysis of wood. In Iberian

Conference on Pattern Recognition and Image Analysis; Springer: Cham, Switzerland, 2019; pp. 257–269.

70. Olschofsky, K.; Köhl, M. Rapid field identification of cites timber species by deep learning. Trees For. People 2020, 2, 100016.

https://doi.org/10.1016/j.tfp.2020.100016.

71. Barmpoutis, P.; Dimitropoulos, K.; Barboutis, I.; Grammalidis, N.; Lefakis, P. Wood species recognition through multidimen-

sional texture analysis. Comput. Electron. Agric. 2018, 144, 241–248. https://doi.org/10.1016/j.compag.2017.12.011.

72. Filho, P.L.P.; Oliveira, L.S.; Nisgoski, S.; Britto, A.S. Forest species recognition using macroscopic images. Mach. Vis. Appl. 2014,

25, 1019–1031. https://doi.org/10.1007/s00138-014-0592-7.

73. Fabijańska, A.; Danek, M.; Barniak, J. Wood species automatic identification from wood core images with a residual convolu-

tional neural network. Comput. Electron. Agric. 2021, 181, 105941. https://doi.org/10.1016/j.compag.2020.105941.

74. de Andrade, B.G.; Basso, V.M.; de Figueiredo Latorraca, J.V. Machine vision for field-level wood identification. IAWA J. 2020,

41, 681–698. https://doi.org/10.1163/22941932-bja10001.

75. Khalid, M.; Lew, E.; Lee, Y.; Yusof, R.; Nadaraj, M. Design of an intelligent wood species recognition system. Int. J. Simul. Syst.

Sci. Technol. 2008, 9, 9–19.

76. Wang, H.; Zhang, G.; Qi, H. Wood recognition using image texture features. PLoS ONE 2013, 8, e76101.

https://doi.org/10.1371/journal.pone.0076101.

77. Lens, F.; Liang, C.; Guo, Y.; Tang, X.; Jahanbanifard, M.; da Silva, F.S.C.; Ceccantini, G.; Verbeek, F.J. Computer-assisted timber

identification based on features extracted from microscopic wood sections. IAWA J. 2020, 41, 660–680.

https://doi.org/10.1163/22941932-bja10029.

78. Martins, J.; Oliveira, L.S.; Nisgoski, S.; Sabourin, R. A database for automatic classification of forest species. Mach. Vis. Appl.

2013, 24, 567–578. https://doi.org/10.1007/s00138-012-0417-5.

79. da Silva, N.R.; De Ridder, M.; Baetens, J.M.; Van den Bulcke, J.; Rousseau, M.; Bruno, O.M.; Beeckman, H.; Van Acker, J.; De

Baets, B. Automated classification of wood transverse cross-section micro-imagery from 77 commercial central-African timber

species. Ann. For. Sci. 2017, 74, 30. https://doi.org/10.1007/s13595-017-0619-0.

80. Mallik, A.; Tarrío-Saavedra, J.; Francisco-Fernández, M.; Naya, S. Classification of wood micrographs by image segmentation.

Chemom. Intell. Lab. Syst. 2011, 107, 351–362. https://doi.org/10.1016/j.chemolab.2011.05.005.

81. Kobayashi, K.; Akada, M.; Torigoe, T.; Imazu, S.; Sugiyama, J. Automated recognition of wood used in traditional japanese

sculptures by texture analysis of their low-resolution computed tomography data. J. Wood Sci. 2015, 61, 630–640.

https://doi.org/10.1007/s10086-015-1507-6.

82. Lopes, D.J.V.; Burgreen, G.W.; Entsminger, E.D. North american hardwoods identification using machine-learning. Forests 2020,

11, 298. https://doi.org/10.3390/f11030298.

83. Wiedenhoeft, A.C. The XyloPhone: Toward democratizing access to high-quality macroscopic imaging for wood and other

substrates. IAWA J. 2020, 41, 699–719. https://doi.org/10.1163/22941932-bja10043.

84. Yu, H.; Cao, J.; Luo, W.; Liu, Y. Image retrieval of wood species by color, texture, and spatial information. In Proceedings of the

2009 International Conference on Information and Automation, Zhuhai/Macau, China, 22–24 June 2009; pp. 1116–1119.

85. Yusof, R.; Khalid, M.; Khairuddin, A.S.M. Application of kernel-genetic algorithm as nonlinear feature selection in tropical

wood species recognition system. Comput. Electron. Agric. 2013, 93, 68–77. https://doi.org/10.1016/j.compag.2013.01.007.

Forests 2022, 13, 2041 24 of 26

86. Zamri, M.I.P.; Cordova, F.; Khairuddin, A.S.M.; Mokhtar, N.; Yusof, R. Tree species classification based on image analysis using

improved-basic gray level aura matrix. Comput. Electron. Agric. 2016, 124, 227–233. https://doi.org/10.1016/j.compag.2016.04.004.

87. Figueroa-Mata, G.; Mata-Montero, E.; Valverde-Otarola, J.C.; Arias-Aguilar, D. Using deep convolutional networks for species

identification of xylotheque samples. In Proceedings of the 2018 IEEE International Work Conference on Bioinspired Intelli-

gence (IWOBI), San Carlos, Costa Rica, 18–20 July 2018; pp. 1–9.

88. Kobayashi, K.; Kegasa, T.; Hwang, S.S.-W.; Sugiyama, J. Anatomical features of fagaceae wood statistically extracted by com-

puter vision approaches: Some relationships with evolution. PLoS ONE 2019, 14, e0220762. https://doi.org/10.1371/jour-

nal.pone.0220762.

89. Souza, D.V.; Santos, J.X.; Vieira, H.C.; Naide, T.L.; Nisgoski, S.; Oliveira, L.E.S. An automatic recognition system of brazilian

flora species based on textural features of macroscopic images of wood. Wood Sci. Technol. 2020, 54, 1065–1090.

https://doi.org/10.1007/s00226-020-01196-z.

90. Tang, X.J.; Tay, Y.H.; Siam, N.A.; Lim, S.C. MyWood-ID. Rapid and robust automated macroscopic wood identification system

using smartphone with macro-lens. In Proceedings of the 2018 International Conference on Computational Intelligence and Intelligent

Systems—CIIS 2018; ACM Press: New York, NY, USA, 2018; pp. 37–43.

91. Stanford, V.L.; University, S.; University, P. ImageNet. Available online: https://www.image-net.org/ (accessed on 21 April

2021).

92. Seregin, A.P. Moscow digital herbarium: A consortium since 2019. Taxon 2020, 69, 417–419. https://doi.org/10.1002/tax.12228.

93. New Your Botanical Garden, N. Index Herbariorum. Available online: http://sweetgum.nybg.org/science/ih/ (accessed on 21

April 2022).

94. Soltis, P.S. Digitization of herbaria enables novel research. Am. J. Bot. 2017, 104, 1281–1284.

95. Sugiyama, J.; Hwang, S.W.; Kobayashi, K.; Zhai, S.; Kanai, I.; Kanai, K. Database of cross sectional optical micrograph from

KYOw lauraceae wood. Available online: https://repository.kulib.kyoto-u.ac.jp/dspace/handle/2433/245888 (accessed on 21

April 2022).

96. Wood-Auth. Barmpoutis, P. WOOD-AUTH Dataset A (Version 0.1). Available online: https://doi.org/10.2018/wood.auth (ac-

cessed on 21 April 2022).

97. Nasirzadeh, M.; Khazael, A.A.; bin Khalid, M. Woods recognition system based on local binary pattern. In Proceedings of the

Second International Conference on Computational Intelligence, Communication Systems and Networks, CICSyN 2010, Liver-

pool, UK, 28–30 July 2010; pp. 308–313.

98. Khalid, M.; Yusof, R.; Khairuddin, A.S.M. Improved tropical wood species recognition system based on multi-feature extractor

and classifier. Int. J. Electr. Comput. Eng. 2011, 5, 495–501.

99. Damayanti, R.; Prakasa, E.; Krisdianto; Dewi, L.M.; Wardoyo, R.; Sugiarto, B.; Pardede, H.F.; Riyanto, Y.; Astutiputri, V.; Pan-

jaitan, G.R.; et al. LignoIndo: Image database of indonesian commercial timber. IOP Conf. Ser. Earth Environ. Sci. 2019, 374, 12057.

https://doi.org/10.1088/1755-1315/374/1/012057.

100. Kour, A.; Yv, V.; Maheshwari, V.; Prashar, D. A review on image processing. Int. J. Electron. Commun. Comput. Eng. 2012, 4,

2278–4209.

101. Martins, J.; Oliveira, L.S.; Britto, A.S.; Sabourin, R. Forest species recognition based on dynamic classifier selection and dissim-

ilarity feature vector representation. Mach. Vis. Appl. 2015, 26, 279–293. https://doi.org/10.1007/s00138-015-0659-0.

102. Yusof, R.; Khalid, M.; Khairuddin, A.S.M. Fuzzy logic-based pre-classifier for tropical wood species recognition system. Mach.

Vis. Appl. 2013, 24, 1589–1604. https://doi.org/10.1007/s00138-013-0526-9.

103. Yusof, R.; Rosli, N.R.; Khalid, M. Tropical wood species recognition based on gabor filter. In Proceedings of the 2009 2nd Inter-

national Congress on Image and Signal Processing, Tianjin, China, 17–19 October 2009; pp. 1–5.

104. Brunel, G.; Borianne, P.; Subsol, G.; Jaeger, M.; Caraglio, Y. Automatic identification and characterization of radial files in light

microscopy images of wood. Ann. Bot. 2014, 114, 829–840. https://doi.org/10.1093/aob/mcu119.

105. Kobayashi, K.; Hwang, S.-W.; Lee, W.-H.; Sugiyama, J. Texture analysis of stereograms of diffuse-porous hardwood: Identifi-

cation of wood species used in tripitaka koreana. J. Wood Sci. 2017, 63, 322–330. https://doi.org/10.1007/s10086-017-1625-4.

106. Tou, J.Y.; Tay, Y.H.; Lau, P.Y. Rotational invariant wood species recognition through wood species verification. In Proceedings

of the 2009 First Asian Conference on Intelligent Information and Database Systems, Dong hoi, Vietnam, 1–3 April 2009; pp.

115–120.

107. Cavalin, P.R.; Kapp, M.N.; Martins, J.; Oliveira, L.E.S. A multiple feature vector framework for forest species recognition. In

Proceedings of the Proceedings of the 28th Annual ACM Symposium on Applied Computing—SAC ’13; ACM Press: New York, NY,

USA, 2013; p. 16.

108. Yusof, R.; Khairuddin, U.; Rosli, N.R.; Ghafar, H.A.; Azmi, N.M.A.N.; Ahmad, A.; Khairuddin, A.S.M. A Study of feature ex-

traction and classifier methods for tropical wood recognition system. In Proceedings of the TENCON 2018—2018 IEEE Region

10 Conference, Jeju, Republic of Korea, 28–31 October 2018; pp. 2034–2039.

109. Lewis, N.D. Deep Learning Made Easy with R: A Gentle Introduction for Data Science; Platform, C.I.P., Ed.; CreateSpace Independent

Publishing Platform: Scotts Valley, CA, USA, 2016; ISBN 1519514212.

110. Zhuo, L.; Cheng, B.; Zhang, J. A Comparative study of dimensionality reduction methods for large-scale image retrieval. Neu-

rocomputing 2014, 141, 202–210. https://doi.org/10.1016/j.neucom.2014.03.014.

111. Fukushima, K. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by

shift in position. Biol. Cybern. 1980, 36, 193–202. https://doi.org/10.1007/BF00344251.

Forests 2022, 13, 2041 25 of 26

112. Lu, D.; Weng, Q. A Survey of image classification methods and techniques for improving classification performance. Int. J.

Remote Sens. 2007, 28, 823–870. https://doi.org/10.1080/01431160600746456.

113. Ravindran, P.; Costa, A.; Soares, R.; Wiedenhoeft, A.C. Classification of CITES-listed and other neotropical meliaceae wood

images using convolutional neural networks. Plant Methods 2018, 14, 25. https://doi.org/10.1186/s13007-018-0292-9.

114. Andrade, B.G.D.; Vital, B.R.; Carneiro, A.D.C.O.; Basso, V.M.; Pinto, F.D.A.D.C. Potential of texture analysis for charcoal classi-

fication. Floresta e Ambient. 2019, 26, e20171241. https://doi.org/10.1590/2179-8087.124117.

115. LeCun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature 2015, 521, 436–444. https://doi.org/10.1038/nature14539.

116. Goodfellow, I.; Bengio, Y.; Courville, A. Deep Learning; MIT Press: Cambridge, UK, 2016; ISBN 9780262035613.

117. Abiodun, O.I.; Kiru, M.U.; Jantan, A.; Omolara, A.E.; Dada, K.V.; Umar, A.M.; Linus, O.U.; Arshad, H.; Kazaure, A.A.; Gana, U.

Comprehensive review of artificial neural network applications to pattern recognition. IEEE Access 2019, 7, 158820–158846.

https://doi.org/10.1109/ACCESS.2019.2945545.

118. Schmidhuber, J. Deep learning in neural networks: An overview. Neural Networks 2015, 61, 85–117.

https://doi.org/10.1016/j.neunet.2014.09.003.

119. Yu, Y.; Si, X.; Hu, C.; Zhang, J. A review of recurrent neural networks: LSTM cells and network architectures. Neural Comput.

2019, 31, 1235–1270. https://doi.org/10.1162/neco_a_01199.

120. Panzer, M.; Bender, B. Deep reinforcement learning in production systems: A systematic literature review. Int. J. Prod. Res. 2022,

60, 4316–4341. https://doi.org/10.1080/00207543.2021.1973138.

121. Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; et al. Recent advances in convo-

lutional neural networks. Pattern Recognit. 2018, 77, 354–377. https://doi.org/10.1016/j.patcog.2017.10.013.

122. Huang, T.S. Computer vision: Evolution and promise. Comput. Sci. 1996. https://doi.org/10.5170/CERN-1996-008.21.

123. Carriquiry, A.; Hofmann, H.; Tai, X.H.; VanderPlas, S. Machine learning in forensic applications. Significance 2019, 16, 29–35.

https://doi.org/10.1111/j.1740-9713.2019.01252.x.

124. Camps-Valls, G.; Tuia, D.; Zhu, X.X.; Reichstein, M. Deep Learning for the Earth Sciences; Camps-Valls, G., Tuia, D., Zhu, X.X.,

Reichstein, M., Eds.; Wiley: Hoboken, NJ, USA, 2021; ISBN 9781119646143.

125. Singh, S.P.; Kumar, A.; Darbari, H.; Singh, L.; Rastogi, A.; Jain, S. Machine translation using deep learning: An overview. In

Proceedings of the 2017 International Conference on Computer, Communications and Electronics (Comptelix), Jaipur, India, 1–

2 July 2017; pp. 162–167.

126. Clanuwat, T.; Bober-Irizar, M.; Kitamoto, A.; Lamb, A.; Yamamoto, K.; Ha, D. Deep learning for classical japanese literature. In

Proceedings of the Workshop on Machine Learning for Creativity and Design, Vancouver, BC, Canada, 9 December 2018.

127. Li, Y.; Huang, C.; Ding, L.; Li, Z.; Pan, Y.; Gao, X. Deep learning in bioinformatics: Introduction, application, and perspective in

the big data era. Methods 2019, 166, 4–21. https://doi.org/10.1016/j.ymeth.2019.04.008.

128. Esteban, L.G.; Fernández, F.G.; de Palacios, P.; Romero, R.M.; Cano, N.N. Artificial neural networks in wood identification: The

case of two juniperus species from the canary islands. IAWA J. 2009, 30, 87–94. https://doi.org/10.1163/22941932-90000206.

129. Turhan, K.; Serdar, B. Support vector machines in wood identification: The case of three salix species from Turkey. Turkish J.

Agric. For. 2013, 37, 249–256. https://doi.org/10.3906/tar-1205-47.

130. Hafemann, L.G.; Oliveira, L.S.; Cavalin, P. Forest species recognition using deep convolutional neural networks. In Proceedings

of the 2014 22nd International Conference on Pattern Recognition, Stockholm, Sweden, 24–28 August 2014; pp. 1103–1107.

131. Esteban, L.G.; de Palacios, P.; Conde, M.; Fernández, F.G.; García-Iruela, A.; González-Alonso, M. Application of artificial neural

networks as a predictive method to differentiate the wood of Pinus sylvestris, L. and pinus nigra arn subsp. Salzmannii (dunal)

franco. Wood Sci. Technol. 2017, 51, 1249–1258. https://doi.org/10.1007/s00226-017-0932-7.

132. Kwon, O.; Lee, H.G.; Lee, M.-R.; Jang, S.; Yang, S.-Y.; Park, S.-Y.; Choi, I.-G.; Yeo, H. Automatic wood species identification of

korean softwood based on convolutional neural networks. J. Korean Wood Sci. Technol. 2017, 45, 797–808.

https://doi.org/10.5658/WOOD.2017.45.6.797.

133. de Oliveira, W.; Filho, P.L.d.P.; Martins, J.G. Software for forest species recognition based on digital images of wood. FLORESTA

2018, 49, 543–552. https://doi.org/10.5380/rf.v49 i3.60075.

134. Kwon, O.; Lee, H.G.; Yang, S.-Y.; Kim, H.; Park, S.-Y.; Choi, I.-G.; Yeo, H. Performance enhancement of automatic wood classi-

fication of korean softwood by ensembles of convolutional neural networks. J. Korean Wood Sci. Technol. 2019, 47, 265–276.

https://doi.org/10.5658/WOOD.2019.47.3.265.

135. Kanayama, H.; Ma, T.; Tsuchikawa, S.; Inagaki, T. Cognitive spectroscopy for wood species identification: Near infrared hyper-

spectral imaging combined with convolutional neural networks. Analyst 2019, 144, 6438–6446.

https://doi.org/10.1039/C9AN01180C.

136. He, T.; Marco, J.; Soares, R.; Yin, Y.; Wiedenhoeft, A. Machine learning models with quantitative wood anatomy data can dis-

criminate between swietenia macrophylla and swietenia mahagoni. Forests 2019, 11, 36. https://doi.org/10.3390/f11010036.

137. Deklerck, V.; Mortier, T.; Goeders, N.; Cody, R.B.; Waegeman, W.; Espinoza, E.; Van Acker, J.; Van den Bulcke, J.; Beeckman,

H. A protocol for automated timber species identification using metabolome profiling. Wood Sci. Technol. 2019, 53, 953–965.

https://doi.org/10.1007/s00226-019-01111-1.

138. de Geus, A.R.; da Silva, S.F.; Gontijo, A.B.; Silva, F.O.; Batista, M.A.; Souza, J.R. An analysis of timber sections and deep learning

for wood species classification. Multimed. Tools Appl. 2020, 79, 34513–34529. https://doi.org/10.1007/s11042-020-09212-x.

139. Wu, F.; Gazo, R.; Haviarova, E.; Benes, B. Wood identification based on longitudinal section images by using deep learning.

Wood Sci. Technol. 2021, 55, 553–563. https://doi.org/10.1007/s00226-021-01261-1.

Forests 2022, 13, 2041 26 of 26

140. Shugar, A.N.; Drake, B.L.; Kelley, G. rapid identification of wood species using XRF and neural network machine learning. Sci.

Rep. 2021, 11, 17533. https://doi.org/10.1038/s41598-021-96850-2.

141. Da Silva, N.R.; Deklerck, V.; Baetens, J.; Van den Bulcke, J.; De Ridder, M.; Rousseau, M.; Bruno, O.M.; Beeckman, H.; Acker, J.

Van; De Baets, B.; et al. Improved wood species identification based on multi-view imagery of the three anatomical planes.

Prepr. from Res. Sq. 2022, 18, 79. https://doi.org/10.21203/rs.3.rs-1167349/v1.

142. Chai, J.; Zeng, H.; Li, A.; Ngai, E.W.T. Deep learning in computer vision: A critical review of emerging techniques and applica-

tion scenarios. Mach. Learn. with Appl. 2021, 6, 100134. https://doi.org/10.1016/j.mlwa.2021.100134.

143. Goodfellow, I.; Pouget-Abadie, J.; Mirza, M.; Xu, B.; Warde-Farley, D.; Ozair, S.; Courville, A.; Bengio, Y. Generative adversarial

nets. In Proceedings of the Advances in Neural Information Processing Systems, Montreal, QC, Canada, 8–13 December 2014.

144. Yi, X.; Walia, E.; Babyn, P. Generative adversarial network in medical imaging: A review. Med. Image Anal. 2019, 58, 101552.

https://doi.org/10.1016/j.media.2019.101552.

145. Lopes, D.J.V.; Monti, G.F.; Burgreen, G.W.; Moulin, J.C.; dos S. Bobadilha, G.; Entsminger, E.D.; Oliveira, R.F. Creating high-

resolution microscopic cross-section images of hardwood species using generative adversarial networks. Front. Plant Sci. 2021,

12, 760139.https://doi.org/10.3389/fpls.2021.760139.

146. Hu, K.; Wang, B.; Shen, Y.; Guan, J.; Cai, Y. Defect identification method for poplar veneer based on progressive growing gen-

erated adversarial network and MASK R-CNN model. BioResources 2020, 15, 3040–3052.

147. Habite, T.; Abdeljaber, O.; Olsson, A. Automatic detection of annual rings and pith location along norway spruce timber boards

using conditional adversarial networks. Wood Sci. Technol. 2021, 55, 461–488. https://doi.org/10.1007/s00226-021-01266-w.

148. Shu, D.; Park, S.W.; Kwon, J. 3D Point cloud generative adversarial network based on tree structured graph convolutions. In

Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Republic of Korea, 27 October–

2 November 2019; pp. 3858–3867.

149. UTAR; FRIM. MyWood-ID. Available online: https://mywoodid.frim.gov.my/ (accessed on 4 May 2022).

150. UTAR; FRIM. MyWood-Premium. Available online: https://mywoodid.frim.gov.my/ (accessed on 20 July 2022).

151. Tay, Y.H. XYLORIX: An AI-as-a-service platform for wood identification. In Proceedings of the IAWA-IUFRO International

Symposium for Updating Wood Identification, Beijing, China, 20–22 May 2019.

152. Agritix Xylorix. Available online: https://www.xylorix.com/products/ (accessed on 4 May 2022).

153. XyloTron.org XyloTron. Available online: https://xylotron.org/ (accessed on 4 May 2022).

154. The National Academies of Sciences, Engineering, and Medicine. Available online: https://sites.nationalacade-

mies.org/PGA/PEER/PEERscience/PGA_195537 (accessed on 4 May 2022).

155. Universidad Politécnica de Madrid, Universidad de Granada, Asociación Española del Comercio e Industria de la Madera.

(AEIM) IMAI App.

The application of chromatography and complementary techniques for wood identification — State of the art and future directions

Article

Jun 2024

Wood identification, whether it involves distinguishing between known species or identifying unfamiliar wood samples, is a scientific field increasingly valuable across various disciplines ranging from biology to criminology, structural engineering, and art conservation. It carries a growing economic, commercial, social, and ecological significance. Numerous scientific methods have been employed for wood identification. Visual analysis techniques, such as macroscopy, optical microscopy, scanning electron microscopy, X-ray tomography, and computer-assisted wood identification, play a pivotal role. Analytical approaches like mass spectrometry, near-infrared spectroscopy, thermogravimetry, and DNA barcoding have also been utilised. However, chromatography stands out due to its exceptional precision in providing quantitative and qualitative results, prompting its development and widespread use. This paper offers a critical review of the role of chromatography in wood identification. Advantages are highlighted, including the capability to identify specific components unique to each wood species, thus achieving remarkable differentiation at the species level. Additionally, potential drawbacks are discussed, such as the time-intensive sample preparation procedures, mainly when dealing with materials like wood-polymer composites (WPC), and the absence of an open-source database. This comprehensive analysis aims to provide a broader perspective on the possibilities and limitations of this technique.

Automatic Wood Species Classification and Pith Detection in Log CT Images

Article

Full-text available

Dec 2024

This article focuses on the need for digitalization in the forestry and timber sector using information from CT scans of logs. The National Forest Centre (Slovak Republic) operates a unique 3D CT scanner for wooden logs at the Stráž Biotechnology Park. This real-time scanner generates a 3D model of a log, displaying the wood's internal features/defects. To optimize log-cutting plans effectively, it is necessary to automatically detect and classify these features and defects in real time, leveraging computer vision principles. Artificial intelligence, specifically neural networks, addresses this need by enabling solutions for tasks of this nature. Building a highly efficient neural network for detecting wood features and defects requires creating a database of log scans and training the network on these data. This is a time-intensive process, as it involves manually marking internal features and defects on hundreds of CT scans of various wood types. A functional neural network for detecting internal wood defects represents a significant advancement in sector digitalization, paving the way for further automation and robotization in wood processing. For the forestry sector to remain competitive, efficiently process raw materials, and improve product quality, the effective application of CT scanning technology is essential. This technological innovation aligns the sector more closely with leaders in other fields, such as the automotive, engineering, and metalworking industries.

Quantifying solid volume of stacked eucalypt timber using detection-segmentation and diameter distribution models

Article

Full-text available

Nov 2024

Accurate timber quantification is essential in forestry and the timber industry, impacting harvest planning, processing, pricing, and overall supply chain management. Traditional methods for estimating the volume of stacked timber, often reliant on manual measurements, are time-consuming and prone to error. This research aims to develop an accurate procedure for estimating the volume of stacked eucalypt timber in yards. The proposed procedure combines automatic log detection and diameter distribution models. Automatic log detection was achieved using advanced computer vision techniques, specifically the You Only Look Once version 9 (YOLOv9) model, which automates the identification and counting of individual logs within a stack. We used stem diameter distribution models to estimate total stack volume based on log counts and probability densities. This approach ensures high accuracy and efficiency, significantly reducing the time and effort required for volume estimation. The dataset used for this study includes diameter measurements from a pre-harvest inventory of eucalypt trees aged 6 and 7 years, alongside videos of stacked timber. The YOLOv9 model was trained to detect logs from these videos, achieving high precision in object detection and segmentation tasks. Performance metrics such as Box Precision, Box Recall, and mean Average Precision (mAP) were used to evaluate the model's effectiveness. The results indicate that the model generalizes well to new data, with high accuracy in both validation and test sets. Among the distribution models evaluated, the generalized extreme value (GEV) distribution provided the best fit for the stem diameter data, allowing for accurate volume predictions. This procedure, which integrates automatic log detection with diameter distribution models, offers a scalable solution applicable to large and complex timber stacks. Finally, a repository was established to allow users to test the proposed method. Future works will focus on refining the model's accuracy and expanding its applicability across different species, forest production and log conditions.

WoodYOLO: A Novel Object Detector for Wood Species Detection in Microscopic Images

Preprint

Full-text available

Nov 2024

Wood species identification plays a crucial role in various industries, from ensuring the legality of timber products to advancing ecological conservation efforts. This paper introduces WoodYOLO, a novel object detection algorithm specifically designed for microscopic wood fiber analysis. Our approach adapts the YOLO architecture to address the challenges posed by large, high-resolution microscopy images and the need for high recall in localization of the cell type of interest (vessel elements). Our results show that WoodYOLO significantly outperforms state-of-the-art models, achieving performance gains of 12.9% and 6.5% in F2 score over YOLOv10 and YOLOv7, respectively. This improvement in automated wood cell type localization capabilities contributes to enhancing regulatory compliance, supporting sustainable forestry practices, and promoting biodiversity conservation efforts globally.

WoodYOLO: A Novel Object Detector for Wood Species Detection in Microscopic Images

Article

Full-text available

Oct 2024

Automating Wood Species Detection and Classification in Microscopic Images of Fibrous Materials with Deep Learning

Article

Full-text available

May 2024
MICROSC MICROANAL

We have developed a methodology for the systematic generation of a large image dataset of macerated wood references, which we used to generate image data for nine hardwood genera. This is the basis for a substantial approach to automate, for the first time, the identification of hardwood species in microscopic images of fibrous materials by deep learning. Our methodology includes a flexible pipeline for easy annotation of vessel elements. We compare the performance of different neural network architectures and hyperparameters. Our proposed method performs similarly well to human experts. In the future, this will improve controls on global wood fiber product flows to protect forests.

Wood-species identification based on terahertz spectral data augmentation and pseudo-label guided deep clustering

Article

Full-text available

Dec 2023

Image classification based on deep learning with automatic relevance determination and structured Bayesian pruning

Article

Full-text available

Aug 2024

Classification of Woody Biomass Images Using Neural Networks and Attention Mechanisms

Article

Aug 2024

Wood biomass is a crucial resource in the field of bioenergy. It is commonly used as a feedstock in bioenergy generation and biochemical production. However, to fully make use of the characteristics of distinct woody biomass, before delivering it to subsequent processing steps, it is necessary to classify and distinguish the different classes of wood biomass. Currently, companies tend to hire operators to carry out the classification task with the help of a digital inventory of biomass classes. However, operators require a great deal of training to correctly group the wood biomass into broad groups with relatively similar quality. Moreover, the judgment bias of operators and the small differences among materials, such as the visual similarity of shavings and sawdust, can lead to potential classification inaccuracy. Considering these challenges, it is inevitable to explore an alternative solution. Thanks to the low cost of the collection of digital images, this work proposes the use of deep neural networks to automatically and precisely classify images. However, this is a challenging task. Specifically, the discrimination in features among certain classes is small, especially with the involvement of some unexpected backgrounds. To address the challenge, compared with the traditional practice that a mere feature vector is extracted for an image, a feature extraction module is applied to produce sets of diverse feature vectors from a single image to explore more visual information. Self-attention mechanisms are intensively utilized in the stage of extraction. The effectiveness of the proposed method is experimentally shown in our collected data set. After a voting strategy is applied, the average accuracy of our method is 96.2%.

Wood Type Identification via Neural Networks and Spectral Analysis: An Advanced Algorithmic Solution

Conference Paper

Dec 2023

Improved wood species identification based on multi-view imagery of the three anatomical planes

Article

Full-text available

Jun 2022
PLANT METHODS

Background The identification of tropical African wood species based on microscopic imagery is a challenging problem due to the heterogeneous nature of the composition of wood combined with the vast number of candidate species. Image classification methods that rely on machine learning can facilitate this identification, provided that sufficient training material is available. Despite the fact that the three main anatomical sections contain information that is relevant for species identification, current methods only rely on transverse sections. Additionally, commonly used procedures for evaluating the performance of these methods neglect the fact that multiple images often originate from the same tree, leading to an overly optimistic estimate of the performance. Results We introduce a new image dataset containing microscopic images of the three main anatomical sections of 77 Congolese wood species. A dedicated multi-view image classification method is developed and obtains an accuracy (computed using the naive but common approach) of 95%, outperforming the single-view methods by a large margin. An in-depth analysis shows that naive accuracy estimates can lead to a dramatic over-prediction, of up to 60%, of the accuracy. Conclusions Additional images from non-transverse sections can boost the performance of machine-learning-based wood species identification methods. Additionally, care should be taken when evaluating the performance of machine-learning-based wood species identification methods to avoid an overestimation of the performance.

Identification of tree species from the Peruvian tropical amazon “Selva Central” forests according to wood anatomy

Article

Full-text available

Sep 2021
BIORESOURCES

The “Selva Central” of Peru is characterized by its forest species richness that produces quality wood for countless uses. Therefore, it is necessary to identify the wood and its macroscopic anatomy, which is an important tool for the botanical identification of tree species. For this purpose, 13 sawmills located in 3 provinces were selected that exploit several tree species of “Selva Central”. Sampling of representative woods was carried out, identified by common names and, in the laboratory, they were polished, examined, and grouped by the similarity of the macroscopic anatomical structure, leading to the tree species identification. Twenty tree species were identified, belonging to 17 genera, with emphasis on the Lauraceae and Fabaceae families. However, Moraceae, Meliaceae, Lecythidaceae, Euphorbiaceae, Bignoniaceae, Myristicaceae, Combretaceae, and Burseraceae families were also identified. The anatomical structures of all the identified tree species were described, transversal and longitudinal tangential cross section images were collected, and a species identification key was constructed. The implications and importance of tree species identification via wood anatomy were discussed, in terms of controlling forest exploitation, traceability of the production chain, and the future development of an artificial intelligence tree-species identification method.

Creating High-Resolution Microscopic Cross-Section Images of Hardwood Species Using Generative Adversarial Networks

Article

Full-text available

Oct 2021

Microscopic wood identification plays a critical role in many economically important areas in wood science. Historically, producing and curating relevant and representative microscopic cross-section images of wood species is limited to highly experienced and trained anatomists. This manuscript demonstrates the feasibility of generating synthetic microscopic cross-sections of hardwood species. We leveraged a publicly available dataset of 119 hardwood species to train a style-based generative adversarial network (GAN). The proposed GAN generated anatomically accurate cross-section images with remarkable fidelity to actual data. Quantitative metrics corroborated the capacity of the generative model in capturing complex wood structure by resulting in a Fréchet inception distance score of 17.38. Image diversity was calculated using the Structural Similarity Index Measure (SSIM). The SSIM results confirmed that the GAN approach can successfully synthesize diverse images. To confirm the usefulness and realism of the GAN generated images, eight professional wood anatomists in two experience levels participated in a visual Turing test and correctly identified fake and actual images at rates of 48.3 and 43.7%, respectively, with no statistical difference when compared to random guess. The generative model can synthesize realistic, diverse, and meaningful high-resolution microscope cross-section images that are virtually indistinguishable from real images. Furthermore, the framework presented may be suitable for improving current deep learning models, helping understand potential breeding between species, and may be used as an educational tool.

Deep reinforcement learning in production systems: a systematic literature review

Article

Full-text available

Sep 2021

Shortening product development cycles and fully customisable products pose major challenges for production systems. These not only have to cope with an increased product diversity but also enable high throughputs and provide a high adaptability and robustness to process variations and unforeseen incidents. To overcome these challenges, deep Reinforcement Learning (RL) has been increasingly applied for the optimisation of production systems. Unlike other machine learning methods, deep RL operates on recently collected sensor-data in direct interaction with its environment and enables real-time responses to system changes. Although deep RL is already being deployed in production systems, a systematic review of the results has not yet been established. The main contribution of this paper is to provide researchers and practitioners an overview of applications and to motivate further implementations and research of deep RL supported production systems. Findings reveal that deep RL is applied in a variety of production domains, contributing to data-driven and flexible processes. In most applications, conventional methods were outperformed and implementation efforts or dependence on human experience were reduced. Nevertheless, future research must focus more on transferring the findings to real-world systems to analyse safety aspects and demonstrate reliability under prevailing conditions.

Rapid identification of wood species using XRF and neural network machine learning

Article

Full-text available

Sep 2021

An innovative approach for the rapid identification of wood species is presented. By combining X-ray fluorescence spectrometry with convolutional neural network machine learning, 48 different wood specimens were clearly differentiated and identified with a 99% accuracy. Wood species identification is imperative to assess illegally logged and transported lumber. Alternative options for identification can be time consuming and require some level of sampling. This non-invasive technique offers a viable, cost-effective alternative to rapidly and accurately identify timber in efforts to support environmental protection laws and regulations.

Deep learning in computer vision: A critical review of emerging techniques and application scenarios

Article

Full-text available

Aug 2021

Deep learning has been overwhelmingly successful in computer vision (CV), natural language processing, and video/speech recognition. In this paper, our focus is on CV. We provide a critical review of recent achievements in terms of techniques and applications. We identify eight emerging techniques, investigate their origins and updates, and finally emphasize their applications in four key scenarios, including recognition, visual tracking, semantic segmentation, and image restoration. We recognize three development stages in the past decade and emphasize research trends for future works. The summarizations, knowledge accumulations, and creations could benefit researchers in the academia and participators in the CV industries.

Institutional quality, economic development and illegal logging: a quantitative cross-national analysis

Article

Full-text available

Oct 2021
EUR J FOREST RES

Matthias Bösch

Illegal logging is a global concern, associated with severe negative environmental, social and economic impacts, such as deforestation, degradation of biodiversity and loss of government revenues. Despite recent international efforts to combat illegal logging activities, the problem remains widespread. While the academic literature on the subject is extensive, little systematic research has been devoted to analysing the causes of illegal logging. Here, this knowledge gap is addressed with a cross-national assessment of factors hypothesized to impact illegal logging. The logistic regression analysis conducted in this study corroborates some widely held beliefs, but also provides some new insights on the factors that are important for whether illegal logging is likely to be a problem. It is shown that, besides physical-geographic characteristics, a number of factors relating to the level and speed of a country’s economic-institutional development are associated with illegal logging. These include gross domestic product per capita, economic growth, voice and accountability, rule of law and control of corruption. The findings also have implications for existing policies to tackle illegal logging activities.

Computer vision-based wood identification and its expansion and contribution potentials in wood science: A review

Article

Full-text available

Apr 2021
PLANT METHODS

The remarkable developments in computer vision and machine learning have changed the methodologies of many scientific disciplines. They have also created a new research field in wood science called computer vision-based wood identification, which is making steady progress towards the goal of building automated wood identification systems to meet the needs of the wood industry and market. Nevertheless, computer vision-based wood identification is still only a small area in wood science and is still unfamiliar to many wood anatomists. To familiarize wood scientists with the artificial intelligence-assisted wood anatomy and engineering methods, we have reviewed the published mainstream studies that used or developed machine learning procedures. This review could help researchers understand computer vision and machine learning techniques for wood identification and choose appropriate techniques or strategies for their study objectives in wood science.

Automatic detection of annual rings and pith location along Norway spruce timber boards using conditional adversarial networks

Article

Full-text available

Mar 2021
WOOD SCI TECHNOL

In the woodworking industry, detection of annual rings and location of pith in relation to timber board cross sections, and how these properties vary in the longitudinal direction of boards, is relevant for many purposes such as assessment of shape stability and prediction of mechanical properties of timber. The current work aims at developing a fast, accurate and operationally simple deep learning-based algorithm for automatic detection of surface growth rings and pith location along knot-free clear wood sections of Norway spruce boards. First, individual surface growth rings that are visible along the four longitudinal sides of the scanned boards are detected using trained conditional generative adversarial networks (cGANs). Then, pith locations are determined, on the basis of the detected growth rings, by using a trained multilayer perceptron (MLP) artificial neural network. The proposed algorithm was solely based on raw images of board surfaces obtained from optical scanning and applied to a total of 104 Norway spruce boards with nominal dimensions of 45×145×4500mm3

45\times 145\times 4500\,\hbox {mm}^{3}

. The results show that optical scanners and the proposed automatic method allow for accurate and fast detection of individual surface growth rings and pith location along boards. For boards with the pith located within the cross section, median errors of 1.4 mm and 2.9 mm, in the x- and y-direction, respectively, were obtained. For a sample of boards with the pith located outside the board cross section in most positions along the board, the median discrepancy between automatically estimated and manually determined pith locations was 3.9 mm and 5.4 mm in the x- and y-direction, respectively.