Figure 2 - uploaded by Pavol Prokop
Content may be subject to copyright.
a. Graphical visualization of characters 1-9. Axis x: 1-42 Thrips fuscipennis; 43-93 T. sambuci (length in µm).
Source publication
Being faced with practical problems in pest identification, we present a methodical paper based on artificial neural networks to discriminate morphologically very similar species, Thrips sambuci Heeger, 1854 and Thrips fuscipennis Haliday, 1836 (Thysanoptera: Thripinae), as an applied case for more general use. The artificially intelligent system m...
Similar publications
Here we present a simple, cost-effective, and sustainable mass rearing system for western flower thrips, Frankliniella occidentalis, with details of molecular identification. Our rearing methods are improved from other systems used previously because we have organized maintenance responsibilities that occur during a Monday through Friday work wk, a...
Citations
... An automated tool for identification of insects based on taxonomical keys with professional-level accuracy was developed up to family level identification of flies and beetles (Valan et al., 2019). Two species of thrips identified by using TRAJAN neural network simulator and statistical analysis based on morphological characters and morphometric data (Fedor et al., 2014). Neural network has been used for insect classification as harmful and non-harmful in cotton ecosystem (Gassoumi et al., 2000). ...
In the era of 21st century, agriculture is facing many challenges now-a-days to feed the world population. The population growth is increasing day by day and it expected to cross 10 billion by 2050. Agriculture farming plays significant role in growth of Indian economy. India stands second in farm production all over the world. After the green revolution, India face production loss with an estimate of US$ 36 billion. The agriculture production decreases mainly because of insect pests, diseases and weeds in important agricultural crops. Hence, there is a need of transition in farming system to adopt advanced and innovative technologies for more and sustainable production. In recent years Artificial intelligence gained popularity in agriculture and provides solutions in several areas like big data analysis, pest and disease forewarning models, mobile applications in IPM, Information and ICT based crop-advisory system, insect detection, pest and disease identification, etc. In the proposed paper, AI based applications discussed in detail to provide insights into innovative technologies and pave the way for knowledge dissemination and adoption of AI based technologies for more effective crop production and protection.
... Insect identification systems proposed in Solis-Sánchez et al. (2011), Xia et al. (2015, Espinoza et al. (2016), Ebrahimi et al. (2017), Lu et al. (2019), Rustia et al. (2020), Li and Yang (2020) and Li et al. (2021) considered these tiny insects as a target category, but have the ability to distinguish them only from other insect orders. Some studies (Fedor et al., 2008(Fedor et al., , 2014 analyse thrip morphometric details (the qualitative and quantitative traits of an insect's body) from microscope images for species level classification. These studies though, required manual measurement or calculation of morphometric traits from the microscope images which were then fed as features to an Feed Forward Neural Network (FFNN). ...
Accurate identification of insect pests is essential in crop management as they are one of the primary causes of yield losses. However, differences between insect species demand different pest control strategies. Hence, research on new technology for fine-grained classification of insect pests is potentially important. Morphologically similar microscopic pest species classification has received little attention in the literature, and is targeted by this study as a means to address the need for agricultural pest management. We propose a novel computational method for deep learning-based, fine-grained classification of microscopic insects using the Vision Transform (ViT) architecture. This architecture employs an attention mechanism motivated by domain knowledge. The proposed approach consists of two main modules, a Data Preprocessing Module to segment relevant insect features and split the insect into body segments to inform identification, and a Domain Knowledge-Driven Stacked Model based on ViT to generate the prediction from each body segment and to fuse predictions for each segment into an accurate species-level classification. We tested the approach using an image dataset of two economically devastating thrip species – Western Flower thrips (Frankliniella occidentalis) and Plague thrips (Thrips imaginis). These insects are small (∼1mm), exhibit minute inter-species differences, and require different pest control strategies. We compared our model with the original ViT model, RestNet101, and RestNet50. Experimental results achieve an F1-score of 0.978, a 3.27% improvement over the baselines. This is important in the horticultural context given the yield losses that these pest insects are known to cause if their populations remain incorrectly quantified.
... The morphometric analysis method used in this study for discriminating between lineages of T. tabaci requires a prior species identification method, especially when T. tabaci specimens are collected from the field, which may include other thrips species. Other Thrips spp., such as the T. hawaiiensis species group (Palmer and Wetton, 1987), Thrips parvispinus Karny (Johari et al., 2014), and two morphologically similar thrips species, Thrips fuscipennis Haliday and Thrips sambuci Hegger (Fedor et al., 2014), exhibit morphometric variations as well. Therefore, the discrimination of T. tabaci lineages based only on morphometric analysis may lead to misidentification in potentially mixed field-collected samples. ...
The onion thrips, Thrips tabaci Lindeman, is a cryptic species complex of three distinct lineages: L1, L2, and T, which exhibit considerable variation in their biological and ecological traits. The most accurate method for their identification is based on molecular techniques. This study aimed to investigate the morphometric variation of T. tabaci cryptic species complex and to distinguish characters that may be useful in discriminating the lineages. For this purpose, morphometric measurements were performed on the eggs, newly hatched first instar larvae, and newly emerged adults. Our results revealed significant differences in egg size between the three lineages. Moreover, the PCA analysis conducted on morpho-metric measurements of the first instar larvae and adults showed that females of the T lineage are very well separated from the females of the L lineages in the adult stage but not in the first instar larval stage. The distinction between the females is partially congruent with the results of genetic studies. Moreover, our results indicate that adult sexual dimorphism with regard to size in L1 and T lineages is not mediated by the size of eggs and first instar larvae, and this may be due to the different growth rates of males and females.
... Insect identification systems proposed in Solis-Sánchez et al. (2011), Xia et al. (2015, Espinoza et al. (2016), Ebrahimi et al. (2017), Lu et al. (2019), Rustia et al. (2020), Li and Yang (2020) and Li et al. (2021) considered these tiny insects as a target category, but have the ability to distinguish them only from other insect orders. Some studies (Fedor et al., 2008(Fedor et al., , 2014 analyse thrip morphometric details (the qualitative and quantitative traits of an insect's body) from microscope images for species level classification. These studies though, required manual measurement or calculation of morphometric traits from the microscope images which were then fed as features to an Feed Forward Neural Network (FFNN). ...
Accurate identification of insect pests is essential in crop management as they are one of the primary causes of yield losses. However, differences between insect species demand different pest control strategies. Hence, research on new technology for the fine-grained classification of insect pests is potentially important. Morphologically similar microscopic pest species classification has received little attention in the literature, and is targeted by this study as a means to address the need for agricultural pest management. We propose a novel computational method for deep learning-based, fine-grained classification of microscopic insects using the Vision Transform (ViT) architecture. This architecture employs an attention mechanism motivated by domain knowledge. The proposed approach consists of two main modules, a Data Preprocessing Module to segment relevant insect features and split the insect into body segments to inform identification, and a Domain Knowledge-Driven Stacked Model based on ViT to generate the prediction from each body segment and to fuse predictions for each segment into an accurate species-level classification. We tested the approach using an image dataset of two economically devastating thrip species - Western Flower thrips (Frankliniella occidentalis) and Plague thrips (Thrips imaginis). These insects are small (~1mm), exhibit minute inter-species differences, and require different pest control strategies. We compared our model with the original ViT model, RestNet101, and RestNet50. Experimental results achieve an F1-score of 0.952, a 4.50% improvement over the baselines. This is important in the horticultural context given the yield losses that these pest insects are known to cause if their populations remain incorrectly quantified.
The preprint is available at https://ssrn.com/abstract=4137865
... Similarly, an ANN has been developed that can identify 101 European thrips species with 95% reliability [216]. A three-layer ANN using seventeen morphological and 15 quantitative morphometric variables can discriminate two similar thrips species, T. sambuci Heeger and T. fuscipennis Haliday, with 100% accuracy [217]. ...
... The implementation of ANN shows potential to discriminate thrips species with high accuracy [216,217]. SVM-based image processing can be used for high-throughput diagnosis of thrips species [223]. The success of these techniques relies on active collaboration across the field of morpho-taxonomy and robotics. ...
Simple Summary: Thrips are important agricultural and forest pests. They cause damage by sucking plant sap and transmitting several plant viruses. Correct identification is the key for epidemiological studies and formulating appropriate management strategies. The application of molecular and electronic detection platforms has improved the morphological character-based diagnosis of thrips species. This article reviews research on molecular and automated identification of thrips species and discusses future research strategies for rapid and high throughput thrips diagnosis. Abstract: Thrips are insect pests of economically important agricultural, horticultural, and forest crops. They cause damage by sucking plant sap and by transmitting several tospoviruses, ilar-viruses, carmoviruses, sobemoviruses, and machlomoviruses. Accurate and timely identification is the key to successful management of thrips species. However, their small size, cryptic nature, presence of color and reproductive morphs, and intraspecies genetic variability make the identification of thrips species challenging. The use of molecular and electronic detection platforms has made thrips identification rapid, precise, sensitive, high throughput, and independent of developmental stages. Multi-locus phylogeny based on mitochondrial, nuclear, and other markers has resolved ambiguities in morphologically indistinguishable thrips species. Microsatellite, RFLP, RAPD, AFLP, and CAPS markers have helped to explain population structure, gene flow, and intraspecies heterogeneity. Recent techniques such as LAMP and RPA have been employed for sensitive and on-site identification of thrips. Artificial neural networks and high throughput diagnostics facilitate automated identification. This review also discusses the potential of pyrosequencing, microarrays, high throughput sequencing, and electronic sensors in delimiting thrips species.
... Morphology (body size and shape) has always been considered as an important factor which affects the patterns of inter-and intraspecific competition in ecological communities (Fedor et al., 2009(Fedor et al., , 2014Hernandéz et al., 2011;Romero et al., 2014). Virtually all organisms, as well as biological processes, exhibit some degree of plasticity (e.g. ...
... We According to our previous experience with morphometrics and phenotypic plasticity (Fedor et al., 2008(Fedor et al., , 2009(Fedor et al., , 2014, we selected 4 morphometric variables, which are well-visible and easy to measure: width of the head (HW), length of the head (HL), anterior pronotum width (AP) and posterior pronotum width (PP) analysed separately for males and females ( Fig. 2) under the digital image analysis system (microscope Leica DM1000 and image analyser software Leica LAS EZ, Version 2.0.0, © 2010). ...
... This is particularly the case for the Thysanoptera, as their presence on a crop does not necessarily imply a pest problem, as only one percent of the species of thrips are economically important pests (Parker, 1995;Monteiro, 2001). Nevertheless, their correct identifi cation depends on preparing slides, which can be used for identifi cation by specialists based on a study of their morphological characters under a microscope or by the use of cybertaxonomic tools (Fedor et al., 2014). The traditional process of preparing slides of thrips is well described (Mound &Kibby, 1998) andwas Eur. ...
Thrips are important agricultural pests and accurate identifi cation is important for their effective management. In order to determine species, however, they need to be mounted on slides and the traditional process is time-consuming. The aim of this paper is to describe a simple and fast method to prepare temporary slides for the routine identifi cation of thrips, which is not dependent on their colour and hardness. Four species of thrips of different colours were used in the preparations: Frankliniella occidentalis (yellow with brown tergal markings), Frankliniella schultzei (entirely brown), Haplothrips gowdeyii (dark brown to carmine)and Caliothrips phaseoli (brown to black). Slides of each species were prepared using three different methods: traditional (3 days), simplifi ed (6 h) and fast method (10 min). The thrips on the resulting slides were observed under a microscope and important structures used in their identifi cation were compared. The quality of the slides prepared using the traditional method was superior to those prepared using the other two methods if only the transparency and general position of the insects on the slides were considered. The transparency of the slides prepared using the simplifi ed method was also good, but only for the pale coloured species (yellow and grey-brown). The fast method, on the other hand, was very effi cient for routine identification since it resulted in slides of suffi cient quality for identifying species regardless of their colour. It is important, however, to stress that the
fast method is only suitable for preparing temporary slides for routine identifi cation and is not a substitute for the traditional method of preparing permanent slides.
Intraspecific trait variability, produced by genetic variation and phenotypic plasticity within species, allows the optimization of individual’s fitness in different conditions, ultimately enhancing survival and reproduction. We investigated variability in morphological traits of invasive thrips species Hercinothrips femoralis (O. M. Reuter, 1891) during simulated introduction and establishment in a novel environment. Six generations of this species were reared in laboratory for eight months. The initial phase of introduction was simulated by the transfer of thrips generations to environments with different environmental conditions varying in temperature and humidity. The statistical evaluation of seven measured morphological attributes (e.g., body length, wing length) was performed to analyse the morphological variability. Species phenotypic “explosion” in several morphological characters (especially total body length) was observed during the initial phase of introduction in generations brought from the primary site into novel environments with different conditions. Probable phenotypic specialization was observed during the generations following introduction under the same ecological conditions. Furthermore, the most variable morphological features were specified. This study goes beyond the taxonomic level, because its results and main idea can be applied to any invasive species.
This study proposes a deep-learning-based system for detecting and classifying Scirtothrips dorsalis Hood, a highly invasive insect pest that causes significant economic losses to fruit crops worldwide. The system uses yellow sticky traps and a deep learning model to detect the presence of thrips in real time, allowing farmers to take prompt action to prevent the spread of the pest. To achieve this, several deep learning models are evaluated, including YOLOv5, Faster R-CNN, SSD MobileNetV2, and EfficientDet-D0. EfficientDet-D0 was integrated into the proposed smartphone application for mobility and usage in the absence of Internet coverage because of its smaller model size, fast inference time, and reasonable performance on the relevant dataset. This model was tested on two datasets, in which thrips and non-thrips insects were captured under different lighting conditions. The system installation took up 13.5 MB of the device’s internal memory and achieved an inference time of 76 ms with an accuracy of 93.3%. Additionally, this study investigated the impact of lighting conditions on the performance of the model, which led to the development of a transmittance lighting setup to improve the accuracy of the detection system. The proposed system is a cost-effective and efficient alternative to traditional detection methods and provides significant benefits to fruit farmers and the related ecosystem.
Almost sixty years after the first published plea for more systematic research on thrips in Slovakia, the checklist undisputedly requires an appropriate revision with a special emphasis on the economic consequences of climate change and biological commodity trade globalisation synergic effects, followed by the dynamic and significant changes in the native biodiversity due to alien species introduction. The updated checklist contains 189 species ecorded from the area of Slovakia, from three families: Aeolothripidae Uzel, 1895 (15 species), Thripidae Stephens, 1829 (113 species) and Phlaeothripidae Uzel, 1895 (61 species), including 7 beneficiary and 35 economic pest elements, such as one A2 EPPO quarantine pest (Frankliniella occidentalis) and five potential transmitters of tospoviruses (F. occidentalis, F. intonsa, F. fusca, Thrips tabaci, Dictyothrips betae). Several species (e.g., Hercinothrips femoralis, Microcephalothrips abdominalis, F. occidentalis, T. flavus, T. tabaci, Limothrips cerealium, L. denticornis, etc.) may possess a heavy introduction and invasion potential with well-developed mechanisms for successful dispersion.
