Neural Network Models for Weed Detection
Keywords:
convolutional neural networks, weed detection, precision agriculture, deep learning, semantic segmentation, object detection, transfer learning, multispectral imaging, site-specific weed management, image classification, unmanned aerial vehicles, agricultural roboticsAbstract
This paper summarizes the neural network models developed for weed detection in agricultural production systems published up to 2021, covering the classification, object detection, and semantic segmentation schemes. Differences in performance are evaluated and discussed in terms of accuracy, precision, recall and f1 score for convolutional networks with eight layers trained on early ImageNet-scale classification work, residual networks over 50 layers, and inception-style modules in various combinations on a range of single-crop row imagery to multi-species rangeland benchmarks with over 17,000 labeled images. Detection-based frameworks, such as the single-shot and two-stage, are evaluated by means of a mean average precision and a trade-off between inference speed (seconds per frame) and inference accuracy (mean average precision), with the fourth generation of the You Only Look Once architecture showing around 91.5 percent mean average precision on real-time inference speeds of almost 23 frames per second. Under field conditions, the weed-class F1-scores of segmentation-based approaches using multispectral red-green-blue near-infrared fusion were close to 0.80. The single most useful optimization strategy investigated was identified as transfer learning, which led to an accuracy increase of around 5.8 percentage points over networks trained from a random initialization of weights, whereas synthetic oversampling of minority weed classes resulted in an accuracy increase of around 3.2 percentage points. Comparative analysis shows that deep learning fares better than all traditional machine learning classifiers that rely on hand-crafted shape and texture features when evaluated under similar assessment conditions, achieving between 8 and 13% increase in accuracy. Neural network models are an effective and growing mature platform for site-specific weed management with a potential for further development in lightweighting the architecture, fusion of multispectral sensors, and generalization across environments.
Downloads
References
A. Ahmad, D. Saraswat, V. Aggarwal, A. Etienne, and B. Hancock, "Performance of deep learning models for classifying and detecting common weeds in corn and soybean production systems," Computers and Electronics in Agriculture, vol. 184, p. 106081, 2021.
M. H. Asad and A. Bais, "Weed detection in canola fields using maximum likelihood classification and deep convolutional neural network," Information Processing in Agriculture, vol. 7, no. 4, pp. 535–545, 2020.
M. D. Bah, A. Hafiane, and R. Canals, "Deep learning with unsupervised data labeling for weed detection in line crops in UAV images," Remote Sensing, vol. 10, no. 11, p. 1690, 2018.
A. Bakhshipour and A. Jafari, "Evaluation of support vector machine and artificial neural networks in weed detection using shape features," Computers and Electronics in Agriculture, vol. 145, pp. 153–160, 2018.
A. Bakhshipour, A. Jafari, S. M. Nassiri, and D. Zare, "Weed segmentation using texture features extracted from wavelet sub-images," Biosystems Engineering, vol. 157, pp. 1–12, 2017.
Y. Beeharry and V. Bassoo, "Performance of ANN and AlexNet for weed detection using UAV-based images," in Proc. 2020 3rd Int. Conf. Emerging Trends Elect., Electron. Commun. Eng. (ELECOM), 2020, pp. 163–167.
A. dos Santos Ferreira, D. M. Freitas, G. G. da Silva, H. Pistori, and M. T. Folhes, "Weed detection in soybean crops using ConvNets," Computers and Electronics in Agriculture, vol. 143, pp. 314–324, 2017.
A. dos Santos Ferreira, D. M. Freitas, G. G. da Silva, H. Pistori, and M. T. Folhes, "Unsupervised deep learning and semi-automatic data labeling in weed discrimination," Computers and Electronics in Agriculture, vol. 165, p. 104963, 2019.
B. Espejo-Garcia, N. Mylonas, L. Athanasakos, S. Fountas, and I. Vasilakoglou, "Towards weeds identification assistance through transfer learning," Computers and Electronics in Agriculture, vol. 171, p. 105306, 2020.
A. Etienne, A. Ahmad, V. Aggarwal, and D. Saraswat, "Deep learning-based object detection system for identifying weeds using UAS imagery," Remote Sensing, vol. 13, no. 24, p. 5182, 2021.
K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proc. 2016 IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 770–778.
A. Kamilaris and F. X. Prenafeta-Boldú, "Deep learning in agriculture: A survey," Computers and Electronics in Agriculture, vol. 147, pp. 70–90, 2018.
A. Krizhevsky, I. Sutskever, and G. E. Hinton, "ImageNet classification with deep convolutional neural networks," Communications of the ACM, vol. 60, no. 6, pp. 84–90, 2017.
Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," Nature, vol. 521, no. 7553, pp. 436–444, 2015.
A. Milioto, P. Lottes, and C. Stachniss, "Real-time blob-wise sugar beets vs weeds classification for monitoring fields using convolutional neural networks," ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., vol. IV-2/W3, pp. 41–48, 2017.
A. Olsen et al., "DeepWeeds: A multiclass weed species image dataset for deep learning," Scientific Reports, vol. 9, no. 1, p. 2058, 2019.
K. Osorio, A. Puerto, C. Pedraza, D. Jamaica, and L. Rodríguez, "A deep learning approach for weed detection in lettuce crops using multispectral images," AgriEngineering, vol. 2, no. 3, pp. 471–488, 2020.
G. G. Peteinatos, P. Reichel, J. Karouta, D. Andújar, and R. Gerhards, "Weed identification in maize, sunflower, and potatoes with the aid of convolutional neural networks," Remote Sensing, vol. 12, no. 24, p. 4185, 2020.
I. Rakhmatulin, A. Kamilaris, and C. Andreasen, "Deep neural networks to detect weeds from crops in agricultural environments in real-time: A review," Remote Sensing, vol. 13, no. 21, p. 4486, 2021.
I. Sa et al., "weedNet: Dense semantic weed classification using multispectral images and MAV for smart farming," IEEE Robotics and Automation Letters, vol. 3, no. 1, pp. 588–595, 2018.
I. Sa et al., "WeedMap: A large-scale semantic weed mapping framework using aerial multispectral imaging and deep neural network for precision farming," Remote Sensing, vol. 10, no. 9, p. 1423, 2018.
A. Wang, W. Zhang, and X. Wei, "A review on weed detection using ground-based machine vision and image processing techniques," Computers and Electronics in Agriculture, vol. 158, pp. 226–240, 2019.
J. Yu, A. W. Schumann, Z. Cao, S. M. Sharpe, and N. S. Boyd, "Weed detection in perennial ryegrass with deep learning convolutional neural network," Frontiers in Plant Science, vol. 10, Art. 1422, 2019.
J. Yu, A. W. Schumann, S. M. Sharpe, X. Li, and N. S. Boyd, "Detection of grassy weeds in bermudagrass with deep convolutional neural networks," Weed Science, vol. 68, no. 5, pp. 545–552, 2020.
J. Yu, S. M. Sharpe, A. W. Schumann, and N. S. Boyd, "Deep learning for image-based weed detection in turfgrass," European Journal of Agronomy, vol. 104, pp. 78–84, 2019.
J. Yu, S. M. Sharpe, A. W. Schumann, and N. S. Boyd, "Detection of broadleaf weeds growing in turfgrass with convolutional neural networks," Pest Management Science, vol. 75, no. 8, pp. 2211–2218, 2019.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
