Application of ResNet for identification of weeds in grain crops

 With the development of artificial intelligence and machine learning technologies, convolutional neural networks (CNN) have become one of the key tools in the field of computer vision. They are actively used to solve problems related to the identification and classification of objects. In agriculture, where the accuracy and speed of data processing are of critical importance, CNNs have found wide application, in particular, for the automatic recognition of weeds in crops. The purpose of this study is to develop image classifiers based on ResNet-18, ResNet-34 and ResNet-50 convolutional models  to identify weeds and classify their prevalence in grain crop areas. The results of phytosanitary monitoring of wheat (Triticum aestivum L.) and barley (Hordeum sativum L.) crops and photographs of accounting sites are used as initial information. During the monitoring, 19 weed species were identified that were present in the areas with varying degrees of intensity. To quantify the weediness, each species was assigned an identifier: 0 – if the species exceeded the economic threshold of harmfulness (ETH), 1 - if it did not exceed it. The main task of the classifier was to recognize weeds in photographs and determine one of two gradations of weediness of the area. The proposed approach demonstrated high efficiency. The accuracy of classifications on the test set of images was 95% on average. An error matrix was constructed to assess the reliability of the method, which also confirmed the high accuracy of predictions. The obtained results indicate the high efficiency of using ResNet architectures for automatic weed recognition tasks. Practical application of the developed classifiers allows for prompt phytosanitary diagnostics of crops, which contributes to the timely and accurate selection of herbicides for weed control. 

1. FAO. The State of Food and Agriculture 2020. Overcoming water challenges in agriculture, Rome, 2020, p. 210. DOI: 10.4060/cb1447en.

2. Zimdahl R.L., Basinger N.T. Fundamentals of weed science. Elsevier, 2024, p. 560.

3. Kamilaris A., Prenafeta-Boldú F.X. Deep learning in agriculture: A survey. Computers and electronics in agriculture, 2018, vol. 147, pp. 70–90. DOI: 10.1016/j.compag.2018.02.016.

4. Rejeb A., Abdollahi A., Rejeb K., Treiblmaier H. Drones in agriculture: A review and bibliometric analysis. Computers and electronics in agriculture, 2022, vol. 198, p. 107017. DOI: 10.1016/j.compag.2022.107017.

5. Hu W.J., Fan J., Du Y.X., Li B.S., Xiong N., Bekkering E. MDFC–ResNet: An agricultural IoT system to accurately recognize crop diseases. IEEE Access, 2020, vol. 8, pp. 115287–115298. DOI: 10.1109/ACCESS.2020.3001237.

6. Yağ İ., Altan A. Artificial intelligence-based robust hybrid algorithm design and implementation for real-time detection of plant diseases in agricultural environments. Biology, 2022, vol. 11 (12), p. 1732. DOI: 10.3390/biology11121732.

7. Kouadio L., El Jarroudi M., Belabess Z., Laasli  S.-E., Roni M.Z.K., Amine I.D.I., Mokhtari N., Mokrini F., Junk J., Lahlali R. A review on UAVbased applications for plant disease detection and monitoring. Remote Sensing, 2023, vol. 15 (17), p. 4273. DOI: 10.3390/rs15174273.

8. Peteinatos G.G., Reichel P., Karouta J., Andú- jar D., Gerhards R. Weed identification in maize, sunflower, and potatoes with the aid of convolutional neural networks. Remote Sensing, 2020, vol. 12 (24), p. 4185. DOI: 10.3390/rs12244185.

9. Li D., Li Y., Zhang Z. Analysis of convolutional neural networks-based approaches in fruit disease detection for smart agriculture applications. Journal of Optics, 2024, vol. 53, pp. 4256– 4265. DOI: 10.1007/s12596-023-01592-1.

10. Weihs B.J., Tang Z., Tian Z., Heuschele D.J., Siddique A., Terrill T.H., Zhang Z., York L.M., Xu Z. Phenotyping alfalfa (Medicago sativa L.) root structure architecture via integrating confident machine learning with ResNet-18. Plant Phenomics, 2024, vol. 6, p. 0251. DOI: 10.34133/plantphenomics.0251.

11. de Camargo T., Schirrmann M., Landwehr N., Dammer K.H., Pflanz M. Optimized deep learning model as a basis for fast UAV mapping of weed species in winter wheat crops. Remote Sensing, 2021, vol. 13 (9), p. 1704. DOI: 10.3390/rs13091704.

12. Riksen V.S., Shpak V.A. Using deep learning methods for detection and classification of weeds in Fagopyrum esculentum crops. Sibirskii vestnik sel'skokhozyaistvennoi nauki = Siberian Herald of Agricultural Science, 2024, vol. 54, no. 12 (313), pp. 5–14. (In Russian). DOI: 10.26898/0370-8799-2024-12-1.

13. Howard D., Gugger S. Deep learning with fastai and PyTorch: minimum formulas, minimum code, maximum efficiency. St. Petersburg, Peter, 2023. (In Russian).