N.V. Gapon1, A.V. Puzerenko2, M.M. Zhdanova3, D.V. Rudoy4, R.R. Ibadov5

1–5 Don State Technical University, (Rostov-on-Don, Russia)

3 Research Laboratory "Modeling and Development of Intelligent Technical Systems of the AIC"  (Rostov-on-Don, Russia)

1 nikolay-rt@mail.ru, 2 mpismenskova@mail.ru

In agriculture, image segmentation plays a key role in monitoring the condition of crops, soil, and predicting the optimal time for agricultural operations such as sowing, fertilizing, and harvesting. This technology is also actively used to assess potential yields and early detection of plant diseases. The development of deep learning methods has opened up new opportunities to improve the accuracy and efficiency of such systems, but there are still unsolved problems associated with the development of models that can operate in complex conditions and provide consistently high results. One of the main difficulties is the variety of shooting conditions, including changing lighting, weather, and seasonal fluctuations. Traditional approaches to segmentation are often not flexible enough to adapt to these variables, which leads to a decrease in the accuracy and reliability of forecasts. Thus, despite the progress achieved, the creation of effective and accurate models for segmentation of agricultural images continues to be a relevant and in-demand task that requires further research and development. Solving these problems will significantly improve the productivity and sustainability of agro-industrial complexes.

The objective is reducing agricultural image segmentation error using a novel deep learning model.

The main features of agricultural image segmentation are considered. A deep learning model for agricultural image segmentation is presented. The proposed model has a parallel architecture specifically designed to extract detailed color and texture features from images. The proposed model outperforms existing solutions by an average of 10%-15%. The proposed model for agricultural image segmentation is evaluated in comparison with known methods.

Image segmentation is a critical task in computer vision. Segmentation divides a digital image into several segments, areas. In agriculture, image segmentation is widely used to monitor crops and soil, predict the best time for sowing, fertilizing and harvesting, assess crop yields and detect plant diseases.

Gapon N.V., Puzerenko A.V., Zhdanova M.M., Rudoy D.V., Ibadov R.R. Deep learning based agricultural image segmentation. Technologies of Living Systems. 2025. V. 22. № 2. Р. 95-104. DOI: https://doi.org/10.18127/j20700997-202502-10 (In Russian).

1. Zelensky A., Gapon N., Zhdanova M., et al. Image segmentation based on adaptive quaternion anisotropic gradient for optical inspection applications. In Optical Metrology and Inspection for Industrial Applications XI. 2024. V. 13241. P. 551–557.
2. Gapon N., Puzerenko A., Zhdanova M., et al. Advances in agricultural image segmentation: from traditional techniques to deep learning approaches. International Journal of Ecosystems & Ecology Sciences. 2024. V. 14. № 4. P. 147–166.
3. Voronin V., Semenishchev E., Zelensky A., et al. Real‐time deep learning semantic segmentation for 3-D augmented reality. In Real-time Photonic Measurements, Data Management, and Processing VII. 2023. V. 12772. P. 80–87.
4. Hamuda E., Glavin M., Jones E. A survey of image processing techniques for plant extraction and segmentation in the field. Computers and electronics in agriculture. 2016. V. 125. P. 184–199.
5. Yang J., Wang C., Jiang B., et al. Visual perception enabled industry intelligence: state of the art, challenges and prospects. IEEE Transactions on Industrial Informatics. 2020. V. 17. № 3. P. 2204–2219.
6. Tassis L.M., Souza J.E.T., Krohling R.A. A deep learning approach combining instance and semantic segmentation to identify diseases and pests of coffee leaves from in-field images. Computers and Electronics in Agriculture. 2021. V. 186. P. 106191.
7. Yao N., Ni F., Wu M., et al. Deep learning-based segmentation of peach diseases using convolutional neural network. Frontiers in Plant Science. 2022. V. 13. P. 876357.
8. Chen S., Zhang K., Zhao Y., et al. An approach for rice bacterial leaf streak disease segmentation and disease severity estimation. Agriculture. 2021. V. 11. № 5. P. 420.
9. Kang J., Liu L., Zhang F., et al. Semantic segmentation model of cotton roots in-situ image based on attention mechanism. Computers and electronics in agriculture. 2021. V. 189. P. 106370.
10. Yan J., Yan T., Ye W., et al. Cotton leaf segmentation with composite backbone architecture combining convolution and attention. Frontiers in Plant Science. 2023. V. 14. P. 1111175.
11. Otsu N. A threshold selection method from gray-level histograms. Automatica. 1975. V. 11. P. 23–27.
12. Rosenfeld A. The max Roberts operator is a Hueckel-type edge detector. IEEE Transactions on Pattern Analysis and Machine Intelligence. 1981. V. 1. P. 101–103.
13. Ikonomatakis N., Plataniotis K.N., Zervakis M., Venetsanopoulos A.N. Region growing and region merging image segmentation. Proc Digit Sig Process. 1997. V. 1. P. 299–302.
14. Dhanachandra N., Manglem K., Chanu Y.J. Image segmentation using K-means clustering algorithm and subtractive clustering algorithm. Procedia Computer Science. 2015. V. 54. P. 764–771.
15. Longzhe Q., Enchen J. Automatic segmentation method of touching corn kernels in digital image based on improved watershed algorithm. International Conference on New Technology of Agricultural. 2011. P. 27–29.
16. Boykov Y.Y., Jolly M-P. Interactive graph cuts for optimal boundary and region segmentation of objects in N-D images. Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001. 2001. V. 1. P. 105–112.
17. Wang A., Xu Y., Wei X., Cui B. Semantic segmentation of crop and weed using an encoder-decoder network and image enhancement method under uncontrolled outdoor illumination. Ieee Access. 2020. V. 8. P. 81724–81734.
18. Kang J., Liu L., Zhang F., et al. Semantic segmentation model of cotton roots in-situ image based on attention mechanism. Computers and electronics in agriculture. 2021. V. 189. P. 106370.
19. Pei H., Owari T., Tsuyuki S., Zhong Y. Application of a novel multiscale global graph convolutional neural network to improve the accuracy of forest type classification using aerial photographs. Remote Sensing. 2023. V. 15. № 4. P. 1001.
20. Wu J., Wen C., Chen H., et al. DS-DETR: a model for tomato leaf disease segmentation and damage evaluation. Agronomy. 2023 (2022). V. 12. № 9.
21. Douarre C., Crispim-Junior C.F., Gelibert A., et al. Novel data augmentation strategies to boost supervised segmentation of plant disease. Comput Electron Agric. 2019. V. 15. 165. P. 104967.
22. Espejo-Garcia B., Mylonas N., Athanasakos L., et al. Combining generative adversarial networks and agricultural transfer learning for weeds identification. Biosystems Engineering. 2021. V. 204. P. 79–89.
23. Zhu J.Y., Park T., Isola P., Efros A.A. Unpaired image-to-image translation using cycle-consistent adversarial networks. IEEE International Conference on Computer Vision (ICCV). 2017. P. 2223–2232.
24. Boonpook W., Tan Y., Nardkulpat A., et al. Deep learning semantic segmentation for land use and land cover types using Landsat 8 imagery. ISPRS International Journal of Geo-Information. 2023. V. 12. № 1. P. 14.
25. Li K., Zhang L., Li B., et al. Attention-optimized DeepLab V3+ for automatic estimation of cucumber disease severity. Plant Methods. 2022. V. 18. № 1. P. 109.