No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
RoboWeedSupport - Presentation of a cloud based system bridging the gap between in-field weed inspections and decision support systems
Abstract
In order to exploit potentials of 20–40% reduction of herbicide use, as documented by use of Decision Support Systems (DSS), where requirements for manual field inspection constitute a major obstacle, large numbers of digital pictures of weed infestations have been collected and analysed manually by crop advisors. Results were transferred to: 1) DSS, which determined needs for control and connected, optimized options for control returned options for control and 2) convolutional, neural networks, which in this way were trained to enable automatic analysis of future pictures, which support both field- and site-specific integrated weed management.
... After detecting weeds' locations, their species can be estimated and weed distribution maps created (Dyrmann, 2017) that can be converted to optimal spray plans (Rydahl et al., 2017). In the agricultural domain, various methods for weed detection has been proposed. ...
Information about the presence of weeds in fields is important to decide on a weed control strategy. This is especially crucial in precision weed management, where the position of each plant is essential for conducting mechanical weed control or patch spraying. For detecting weeds, this study proposes a fully convolutional neural network, which detects weeds in images and classifies each one as either a monocot or dicot. The network has been trained on over 13 000 weed annotations in high-resolution RGB images from Danish wheat and rye fields. Due to occlusion in cereal fields, weeds can be partially hidden behind or touching the crops or other weeds, which the network handles. The network can detect weeds with an average precision (AP) of 0.76. The weed detection network has been evaluated on an Nvidia Titan X, on which it is able to process a 5 MPx image in 0.02 s, making the method suitable for real-time field operation.
... However, DSS do not fit well into farmers usual practices by requiring manual field inspections and identification of weeds constituting a major obstacle [19]. For the past three years, the Danish nationally funded project, RoboWeedSupport, has sought to bridge the gap between the potential Crop Protection Online (CPO) or IPMwise based herbicide savings and the required field inspections [24]. Initially smartphone cameras and later unmanned aerial drones (UAS) were used to collect images from the field for semi-automated weed discrimination and classification [25]. ...
In recent years, precision agriculture and precision weed control have been developed aiming at optimising yield and cost while minimising environmental impact. Such solutions include robots for precise hoeing or spraying. The commercial success of robots and other precision weed control techniques has, however, been limited, partly due to a combination of a high acquisition price and low capacity compared to conventional spray booms, limiting the usage of precision weeding to high-value crops. Nonetheless, conventional spray booms are rarely used optimally. A study by Jørgensen et al. (2007) has shown that selecting the right herbicides can lead to savings by more than 40 percent in cereal fields without decreasing the crop yield when using conventional sprayers. Therefore, in order to utilise conventional spray booms optimally, a preliminary analysis of the field is necessary. This analysis should determine which weeds are present in the field and the density of those weeds so that herbicides targeting those weeds may be selected.
Researchers have sought to detect and classify weeds and crops in images, but studies are limited regarding the number of plant species that can be discriminated and the flexibility of the camera setup.
In the present PhD thesis, requirements for the camera set-up are loosened, allowing the use of consumer grade cameras or even cell phones for weed species localisation and identification in images from conventionally grown fields. In total 4 537 images have been collected over three growth seasons from Danish fields. In these images 31397 plants are annotated with names, from which the 17 most frequent species are selected for automated classifiiiication. The automated classification consists of two steps: Initially, weeds are located in images after which, the weeds are classified.
Three types of weed localisation approaches are tested: Two approaches that perform a pixel-wise segmentation of plants, and one approach, that detects regions in images containing weeds. Common for all three approaches is that they aim at overcoming some of the challenges when working with images from fields: Namely changes in lighting, soil types, and plant stress due to lack of nutrition. The first of the suggested approaches segments plant material from the soil by using fuzzy C -means clustering combined with a threshold value for each pixel, which depends on the neighbourhood pixels, which helps to detect non-green stem regions. The second approach uses a fully convolutional neural network for segmenting pixels in three categories: Soil, weeds, and crops. The Neural Network is trained solely on modelled images but can segment weeds from maize with an intersection-over-union of between 0.69 and 0.93 for weeds and maize. Rather than segmenting images, the third approach produces region proposals that indicate weed locations in images. This method also uses a fully convolutional neural network, that enables it to detect weed instances in wheat fields despite occluding leaves.
The three methods for weed segmentation and localisation solve four problems in the field of camera based weed detection: handling of changing environments, handling of non-green plant stems, segmentation of weeds and crops that are overlapping, and instance detection in cereal fields with occluding leaves.
Following the detection of the weeds, the weed species are to be determined. For solving this problem, a convolutional neural network is used, which classifies the weeds with an overall accuracy of 87 percent for 17 species despite a severe degree of leaf occlusion.
Because of the ability to handle weed detection and classification in natural environments, these methods can potentially reduce the investment of farmers, and thus lead to a higher adoption rate than existing precision weed control techniques, resulting in huge potential savings regarding herbicide consumption.
... This has resulted in a dataset consisting of 1.427 images, in which 18.541 weeds have been annotated manually. The annotation of these weeds is made using a web portal, as described in Rydahl et al. (2016). Here, the user can mark each weed instance in the image and denote the species of each of the weeds. ...
This paper presents a method for automating weed detection in colour images despite heavy leaf occlusion. A fully convolutional neural network is used to detect the weeds. The network is trained and validated on a total of more than 17,000 annotations of weeds in images from winter wheat fields, which have been collected using a camera mounted on an all-terrain vehicle. Hereby, the network is able to automatically detect single weed instances in cereal fields despite heavy leaf occlusion.
Deep learning approaches have been found to be suitable for the agricultural field with successful applications to vegetable infection through plant disease. In this chapter, the authors discuss some widely used deep learning architecture and their practical applications. Nowadays, in many typical applications of machine vision, there is a tendency to replace classical techniques with deep learning algorithms. The benefits are valuable; on one hand, it avoids the need of specialized handcrafted features extractors, and on the other hand, results are not damaged. Moreover, they typically get improved.
Effective weed control, using either mechanical or chemical means, relies on knowledge of the crop and weed plant occurrences in the field. This knowledge can be obtained automatically by analyzing images collected in the field. Many existing methods for plant detection in images make the assumption that plant foliage does not overlap. This assumption is often violated, reducing the performance of existing methods. This study overcomes this issue by training a convolutional neural network to create a pixel-wise classification of crops, weeds and soil in RGB images from fields, in order to know the exact position of the plants. This training is based on simulated top-down images of weeds and maize in fields. The results show an pixel accuracy over 94% and a 100% detection rate of both maize and weeds, when tested on real images, while a high intersection over union is kept. The system can handle 2.4 images per second for images with a resolution of 1MPix, when using an Nvidia Titan X GPU.
This paper presents a method for automating weed detection in colour images despite heavy leaf occlusion. A fully convolutional neural network is used to detect the weeds. The network is trained and validated on a total of more than 17,000 annotations of weeds in images from winter wheat fields, which have been collected using a camera mounted on an all-terrain vehicle. Hereby, the network is able to automatically detect single weed instances in cereal fields despite heavy leaf occlusion.
Information on which weed species are present within agricultural fields is important for site specific weed management. This paper presents a method that is capable of recognising plant species in colour images by using a convolutional neural network. The network is built from scratch trained and tested on a total of 10,413 images containing 22 weed and crop species at early growth stages. These images originate from six different data sets, which have variations with respect to lighting, resolution, and soil type. This includes images taken under controlled conditions with regard to camera stabilisation and illumination, and images shot with hand-held mobile phones in fields with changing lighting conditions and different soil types. For these 22 species, the network is able to achieve a classification accuracy of 86.2%.
Crop Protection Online (CPO) is a decision support system, which integrates decision algorithms
quantifying the requirement for weed control and a herbicide dose model. CPO was designed to be
used by advisors and farmers to optimize the choice of herbicide and dose. The recommendations
from CPO for herbicide application in spring barley in Denmark were validated through field
experiments targeting three levels of weed control requirement. Satisfactory weed control levels at
harvest were achieved by a medium control level requirement generating substantial herbicide
reductions (~ 60% measured as the Treatment Frequency Index (TFI)) compared to a high level of
required weed control. The observations indicated that the current level of weed control required is
robust for a range of weed scenarios. Weed plant numbers 3 wk after spraying indicated that the
growth of the weed species were inhibited by the applied doses, but not necessarily killed, and that an
adequate level of control was reached later in the season through crop competition.
The highly complex knowledge of scientific disciplines makes nuanced analysis and modelling possible. However, the information produced often does not reach farmers because it is presented in a way that does not correspond to the way their work is carried out in practice. The decision support system Crop Protection Online is widely used by advisors and as a learning tool for students. Although the system has been validated in many field trials over the years and has shown reliable results, the number of end-users among farmers has been relatively low during the last 10 years (approximately 1000 farmers). A sociological investigation of farmers’ decision-making styles in the area of crop protection has shown that arable farmers can be divided into three major groups: (a) system-orientated farmers, (b) experience-based farmers and (c) advisory-orientated farmers. The information required by these three groups to make their decisions varies and therefore different ways of using decision support systems need to be provided. Decision support systems need to be developed in close dialogue and collaboration with user groups.
RoboWeedSupport - Sub millimeter weed image acquisition in cereal crops with speeds up till 50 km/h
- M S Laursen
- R N Jorgensen
- M Dyrmann
- R N Poulsen
Laursen MS, Jorgensen RN, Dyrmann M, Poulsen RN (2017). RoboWeedSupport -Sub
millimeter weed image acquisition in cereal crops with speeds up till 50 km/h. In this volume.
Adapting the Decision Support System CPOWeeds to optimize weed control in northern Spanish conditions. PhD dissertation, Departament de Hortofruticulture
- J M Montull
Montull JM (2016). Adapting the Decision Support System CPOWeeds to optimize weed
control in northern Spanish conditions. PhD dissertation, Departament de Hortofruticulture,
Botanica in Jardineria, Universitat de Lleida, Spain.
Norsk utgave av det danske beslutningsstøttesystemet Plantevaern Online for ugrassprøyting i korn
- K S Tørresen
- J Netland
- P Rydahl
Tørresen KS, Netland J, Rydahl P (2004). Norsk utgave av det danske beslutningsstøttesystemet Plantevaern Online for ugrassprøyting i korn. Grønn kunnskap 8 (2), 100-109.
Review of new technologies critical to effective implementation of Decision Support Systems (DSS's) and Farm Management Systems
- T Been
- A Berti
- N Evans
- D Gouache
- G Volkmar
- J E Jensen
- J Kapsa
- N Levay
- N Munier-Jolain
- S Nibouche
- M Raynal
- P Rydahl
Been T, Berti A, Evans N, Gouache D, Volkmar G, Jensen JE, Kapsa J, Levay N, Munier-Jolain N, Nibouche S, Raynal M, Rydahl P (2009). Review of new technologies critical to
effective implementation of Decision Support Systems (DSS's) and Farm Management
Systems (FMS's) Aarhus University, Denmark, 6th March 2009. http://www.endure-
Norsk utgave av det danske beslutningsstøtte-systemet Planteværn Online for ugrassprøyting i korn
- Tørresen
Review of new technologies critical to effective implementation of Decision Support Systems (DSS’s) and Farm Management Systems (FMS’s) Aarhus University
- T Been
- A Berti
- N Evans
- D Gouache
- G Volkmar
- J E Jensen
- J Kapsa
- N Levay
- N Munier-Jolain
- S Nibouche
- M Raynal
- P Rydahl