In crop detection, ridge structures provide crucial cues for classifying crops and weeds. However, it is difficult to obtain ridge structures for unmanned ground vehicles which can capture images only within a narrow field of view. This study proposes a lightweight algorithm that enables a model to implicitly infer the ridge structure from plant-to-plant spatial relationships and sizes. An object detector first detects each plant. The resulting bounding boxes are treated as pairwise features in the nodes. Metainformation indicating whether two nodes share the same ID is combined with their geometric relationships and encoded as edge features. A graph attention network addresses these relationships to infer and propagate ridge-aware regularities. By understanding the structure only from object relationships, the method compensates for the information lost to the limited field of view without any explicit edge structure input. In the experiments wherein we deliberately introduced a domain shift between the training/validation sets and test set, the proposed method increased the baseline mAP₅₀ from 30.6% to 44.4%. This amounts to an increase of up to 13.8 percentage points. In addition, the proposed method requires only approximately 10 ms/frame on a Jetson AGX Orin to classify plants. This method acquires ridge structures internally without relying on external sensors or hand-tuned thresholds. Thus, it displays potential for in-field agricultural applications such as autonomous weeding.

1. [1] T. Yoshida, T. Fukao, and T. Hasegawa, “Fast detection of tomato peduncle using point cloud with a harvesting robot,” J. Robot. Mechatron., Vol.30, No.2, pp. 180-186, 2018. https://doi.org/10.20965/jrm.2018.p0180
2. [2] M. Kawaguchi and N. Takesue, “A method of detection and identification for axillary buds,” J. Robot. Mechatron., Vol.36, No.1, pp. 201-210, 2024. https://doi.org/10.20965/jrm.2024.p0201
3. [3] J. Tatsuno, K. Tajima, and M. Kato, “Automatic transplanting equipment for chain pot seedlings in shaft tillage cultivation,” J. Robot. Mechatron., Vol.34, No.1, pp. 10-17, 2022. https://doi.org/10.20965/jrm.2022.p0010
4. [4] S. Wolfert, L. Ge, C. Verdouw, and M.-J. Bogaardt, “Big data in smart farming – A review,” Agricultural Systems, Vol.153, pp. 69-80, 2017. https://doi.org/10.1016/j.agsy.2017.01.023
5. [5] A. Olsen et al., “DeepWeeds: A multiclass weed species image dataset for deep learning,” Scientific Reports, Vol.9, Article No.2058, 2019. https://doi.org/10.1038/s41598-018-38343-3
6. [6] D. Steininger, A. Trondl, G. Croonen, J. Simon, and V. Widhalm, “The CropAndWeed dataset: A multi-modal learning approach for efficient crop and weed manipulation,” 2023 IEEE/CVF Winter Conf. on Applications of Computer Vision, pp. 3718-3727, 2023. https://doi.org/10.1109/WACV56688.2023.00372
7. [7] K. Ota, J. Y. Louhi Kasahara, A. Yamashita, and H. Asama, “Weed and crop detection by combining crop row detection and k-means clustering in weed infested agricultural fields,” 2022 IEEE/SICE Int. Symp. on System Integration, pp. 985-990, 2022. https://doi.org/10.1109/SII52469.2022.9708815
8. [8] M. Pérez-Ortiz et al., “Selecting patterns and features for between- and within- crop-row weed mapping using UAV-imagery,” Expert Systems with Applications, Vol.47, pp. 85-94, 2016. https://doi.org/10.1016/j.eswa.2015.10.043
9. [9] P. De Marinis, G. Vessio, and G. Castellano, “RoWeeder: Unsupervised weed mapping through crop-row detection,” Computer Vision – ECCV 2024 Workshops, Part 3, pp. 132-145, 2025. https://doi.org/10.1007/978-3-031-91835-3_9
10. [10] L. P. Osco et al., “A CNN approach to simultaneously count plants and detect plantation-rows from UAV imagery,” ISPRS J. of Photogrammetry and Remote Sensing, Vol.174, pp. 1-17, 2021. https://doi.org/10.1016/j.isprsjprs.2021.01.024
11. [11] T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” 5th Int. Conf. on Learning Representations, 2017.
12. [12] P. Veličković et al., “Graph attention networks,” 6th Int. Conf. on Learning Representations, 2018.
13. [13] Y. Iuchi, A. Koshigoe, S. Nishiwaki, and T. Emaru, “Crop detection method using relative positional relationships for small weeding robots,” 2025 IEEE/SICE Int. Symp. on System Integration, pp. 1357-1362, 2025. https://doi.org/10.1109/SII59315.2025.10871075
14. [14] D. G. Kim, T. F. Burks, J. Qin, and D. M. Bulanon, “Classification of grapefruit peel diseases using color texture feature analysis,” Int. J. of Agricultural and Biological Engineering, Vol.2, No.3, pp. 41-50, 2009. https://doi.org/10.3965/j.issn.1934-6344.2009.03.041-050
15. [15] A. Tellaeche, X. P. Burgos-Artizzu, G. Pajares, and A. Ribeiro, “A vision-based method for weeds identification through the Bayesian decision theory,” Pattern Recognition, Vol.41, No.2, pp. 521-530, 2008. https://doi.org/10.1016/j.patcog.2007.07.007
16. [16] M. Montalvo et al., “Automatic detection of crop rows in maize fields with high weeds pressure,” Expert Systems with Applications, Vol.39, No.15, pp. 11889-11897, 2012. https://doi.org/10.1016/j.eswa.2012.02.117
17. [17] Q. Lü, J. Cai, B. Liu, L. Deng, and Y. Zhang, “Identification of fruit and branch in natural scenes for citrus harvesting robot using machine vision and support vector machine,” Int. J. of Agricultural and Biological Engineering, Vol.7, No.2, pp. 115-121, 2014. https://doi.org/10.3965/j.ijabe.20140702.014
18. [18] R. Girshick, “Fast R-CNN,” 2015 IEEE Int. Conf. on Computer Vision, pp. 1440-1448, 2015. https://doi.org/10.1109/ICCV.2015.169
19. [19] K. He, G. Gkioxari, P. Dollár, and R. Girshick, “Mask R-CNN,” 2017 IEEE Int. Conf. on Computer Vision, pp. 2980-2988, 2017. https://doi.org/10.1109/ICCV.2017.322
20. [20] Z. Cai and N. Vasconcelos, “Cascade R-CNN: High quality object detection and instance segmentation,” IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol.43, No.5, pp. 1483-1498, 2021. https://doi.org/10.1109/TPAMI.2019.2956516
21. [21] N. Carion et al., “End-to-end object detection with transformers,” Proc. of the 16th European Conf. on Computer Vision, Part 1, pp. 213-229, 2020. https://doi.org/10.1007/978-3-030-58452-8_13
22. [22] Sunil G. C. et al., “Field-based multispecies weed and crop detection using ground robots and advanced YOLO models: A data and model-centric approach,” Smart Agricultural Technology, Vol.9, Article No.100538, 2024. https://doi.org/https://doi.org/10.1016/j.atech.2024.100538
23. [23] T. Cover and P. Hart, “Nearest neighbor pattern classification,” IEEE Trans. on Information Theory, Vol.13, No.1, pp. 21-27, 1967. https://doi.org/10.1109/TIT.1967.1053964
24. [24] A. Vaswani et al., “Attention is all you need,” Proc. of the 31st Int. Conf. on Neural Information Processing Systems, pp. 6000-6010, 2017.
25. [25] J. L. Ba, J. R. Kiros, and G. E. Hinton, “Layer normalization,” arXiv:1607.06450, 2016. https://doi.org/10.48550/arXiv.1607.06450
26. [26] D.-A. Clevert, T. Unterthiner, and S. Hochreiter, “Fast and accurate deep network learning by exponential linear units (ELUs),” arXiv:1511.07289, 2016. https://doi.org/10.48550/arXiv.1511.07289
27. [27] S. Ioffe and C. Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” Proc. of the 32nd Int. Conf. on Machine Learning, Vol.37, pp. 448-456, 2015.
28. [28] T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, “Focal loss for dense object detection,” 2017 IEEE Int. Conf. on Computer Vision, pp. 2999-3007, 2017. https://doi.org/10.1109/ICCV.2017.324
29. [29] Y. Tian, Q. Ye, and D. Doermann, “YOLOv12: Attention-centric real-time object detectors,” arXiv:2502.12524, 2025. https://doi.org/10.48550/arXiv.2502.12524
30. [30] H. Zhang et al., “DINO: DETR with improved denoising anchor boxes for end-to-end object detection,” arXiv:2203.03605, 2022. https://doi.org/10.48550/arXiv.2203.03605
31. [31] T.-Y. Lin et al., “Microsoft COCO: Common objects in context,” Proc. of the 13th European Conf. on Computer Vision, Part 5, pp. 740-755, 2014. https://doi.org/10.1007/978-3-319-10602-1_48
32. [32] K. Chen et al., “MMDetection: Open MMLab detection toolbox and benchmark,” arXiv:1906.07155, 2019. https://doi.org/10.48550/arXiv.1906.07155
33. [33] I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” 7th Int. Conf. on Learning Representations, 2019.
34. [34] “The state of food and agriculture 2022: Leveraging automation to transform agrifood systems.” https://openknowledge.fao.org/server/api/core/bitstreams/1c329966-521a-4277-83d7-07283273b64b/content/sofa-2022/technological-change-agricultural-production.html [Accessed July 30, 2025]
35. [35] “White paper on food, agriculture and rural areas (FY2022),” (in Japanese). https://www.maff.go.jp/j/wpaper/w_maff/r4/pdf/zentaiban.pdf [Accessed July 30, 2025]
36. [36] “Ultralytics YOLO (v 8.0.0).” https://github.com/ultralytics/ultralytics [Accessed July 30, 2025]
37. [37] “YOLOv12: Attention-centric real-time object detectors.” https://github.com/sunsmarterjie/yolov12 [Accessed July 30, 2025]