RTrack-DL: A dynamic scene-oriented model for instance identification and continuous tracking of rice seedling rows
Keywords:
rice seedling row instance, seedling row line fitting, merge-and-deduplication strategy, continuous tracking, Kalman filtering, Hungarian algorithmAbstract
Visual navigation is critical for intelligent paddy field operations, where continuous and accurate tracking of rice seedling rows is essential. However, existing methods predominantly address static frame identification and lack robust solutions for continuous tracking in dynamic field environments, particularly under challenging conditions like missing seedlings and irregular planting patterns. To overcome these limitations, this study develops RTrack-DL, a novel dynamic scene-oriented model for instance-level identification and continuous tracking of rice seedling rows. The novelty lies in its dynamic framework that integrates a dedicated merge-and-deduplication strategy (IoU-Plus) to enhance line-fitting robustness against missing seedlings and irregular planting, and fuses the fitted lines with a multi-object tracker to maintain stable row identities and prevent track loss during field operations. The model employs DeepLabv3+ for pixel-level semantic segmentation of row centerlines. Subsequently, connected component analysis and line fitting are applied to obtain individual instances. To handle discontinuities caused by missing seedlings and irregular planting, a merge-and-deduplication strategy based on an IoU-Plus metric is introduced to eliminate redundant lines and enhance fitting robustness. A multi-object tracking module combining Kalman filtering and the Hungarian algorithm ensures consistent identity tracking across frames. The performance of RTrack-DL is evaluated through experiments under three illumination conditions—sunny, cloudy, and overcast—while considering scenarios involving missing seedlings and irregular planting. Results show optimal performance under cloudy conditions, with a fitting accuracy of 99.42% and an angular error of 0.49°, while sunny and overcast conditions yielded slightly reduced accuracy. In tracking tests, RTrack-DL achieved multiple object tracking accuracy scores of 93.5%, 92.1%, and 92.7% under cloudy, sunny, and overcast conditions, respectively, with real-time performance maintained at 15-20 FPS. The proposed framework demonstrates a practical solution for continuous tracking of dense rice seedling rows in dynamic field environments, providing a robust visual perception component for intelligent navigation systems in fully unmanned paddy field management.
Key words: rice seedling row instance; seedling row line fitting; merge-and-deduplication strategy; continuous tracking; Kalman filtering; Hungarian algorithm
DOI: 10.25165/j.ijabe.20261904.10661
Citation: Li Z Q, Li D F, Zheng H, Duan W W, Shan Y C, Xiao M H, et al. RTrack-DL: A dynamic scene-oriented model for
instance identification and continuous tracking of rice seedling rows. Int J Agric & Biol Eng, 2026; 19(4): 243–255.
References
[1] Dadrasi A, Vačkářová D, Salmani F, Nguyen C T, Weinzettel J. Global rice land suitability and adaptation strategies under climate change. J. Environ. Manage., 2025; 394: 127630.
[2] Kotaniemi J, Känsäkoski N, Halbach E, Heikkilä T. A weeding robot for seedling removal. in: 20th IEEEASME Int. Conf. Mechatron. Embed. Syst. Appl. MESA, 2024; pp.1–8. DOI: 10.1109/MESA61532.2024.10704901.
[3] Chen D S, Wang Q L, Lin Y, Ma Z L, Sun L, Yu G H. Path tracking control of paddy field weeder integrated with satellite and visual methods. Comput. Electron. Agric, 2025; 234: 110257.
[4] Feng L Y, Xu C S, Tang H, Wei Z C, Guan X D, Xu J C, et al. Application of navigation technology in agricultural machinery: A review and prospects. Artif. Intell. Agric, 2026; 16(1): 94–123.
[5] Ju J Y, Chen G Q, Lv Z Y, Zhao M Y, Sun L, Wang Z T, et al. Design and experiment of an adaptive cruise weeding robot for paddy fields based on improved YOLOv5. Comput. Electron. Agric., 2024; 219: 108824.
[6] Nunes F, Sousa F. Deep learning soft-decision GNSS multipath detection and mitigation. Sensors, 2024; 24(14): 4663.
[7] Zhang Y W, Liu M N, Chen D, Xu X M, Lu J B, Lai H R, et al. Development and testing of row‐controlled weeding intelligent robot for corn. J. Field Robotics, 2025; 42(3): 850–866.
[8] Han C Y, Wu W B, Luo X W, Li J H. Visual navigation and obstacle avoidance control for agricultural robots via LiDAR and camera. Remote Sensing, 2023; 15(22): 5402.
[9] Martini M, Ambrosio M, Navone A, Tuberga B, Chiaberge M. Enhancing visual autonomous navigation in row-based crops with effective synthetic data generation. Precis. Agric., 2024; 25: 2881–2902.
[10] Navone A, Martini M, Ambrosio M, Ostuni A, Angarano S, Chiaberge M. GPS-free autonomous navigation in cluttered tree rows with deep semantic segmentation. Robot Auton. Syst., 2025; 183: 104854.
[11] Wang S L, Zhang B K, Zhou J, Yang B, He Y. Time–jerk optimal trajectory planning for industrial robots with coupled interpolation function selection. J Field Robotics, 2024; 41(4): 917–941.
[12] Chakhvashvili E, Machwitz M, Antala M, Rozenstein O, Prikaziuk E, Schlerf M, et al. Crop stress detection from UAVs: best practices and lessons learned for exploiting sensor synergies. Precis. Agric., 2024; 25: 2614–2642.
[13] Shammi S, Diepeveen D, Zander S, Jones M G K, Sohel F. Early frost detection in wheat using machine learning from vertical temperature distributions. Comput. Electron. Agric, 2024; 221: 108950.
[14] Zhang Y X, Shen Y, Wang Q, Song C, Dai N, He B. A novel hybrid swarm intelligence algorithm for solving TSP and desired-path-based online obstacle avoidance strategy for AUV. Robot Auton. Syst, 2024; 177: 104678.
[15] Ariza-Sentís M, Vélez S, Martínez-Peña R, Baja H, Valente J. Object detection and tracking in Precision Farming : A systematic review. Comput. Electron. Agric, 2024; 219: 108757.
[16] Kang S, Fan J L, Ye Y K, Li C L, Du D D, Wang J. Maturity recognition and localisation of broccoli under occlusion based on RGB-D instance segmentation network. Biosyst Eng, 2025; 250: 270–284.
[17] Ma Z H, Yin C, Du X Q, Zhao L J, Lin L P, Zhang G F, Wu C Y. Rice row tracking control of crawler tractor based on the satellite and visual integrated navigation. Comput Electron Agric, 2022; 197: 106935.
[18] Liu W, Hu J P, Liu J X, Yue R C, Zhang T F, Yao M J, et al. Method for the navigation line recognition of the ridge without crops via machine vision. Int J Agric & Biol Eng, 2024; 17(2): 230–239.
[19] Li B Y, Xiao C, Wang L G, Wang Y Q, Lin Z P, Li M, et al. Dense nested attention network for infrared small target detection. IEEE Trans. Image Process, 2023; 32: 1745–1758.
[20] Yang Y, Zhou Y, Yue X, Zhang G, Wen X, Ma B, et al. Real-time detection of crop rows in maize fields based on autonomous extraction of ROI. Expert Syst. Appl, 2023; 213(Part A): 118826.
[21] Gong H L, Zhuang W D, Wang X. Improving the maize crop row navigation line recognition method of YOLOX. Front. Plant Sci, 2024; 15: 1338228.
[22] Mohialden Y M, Hussien N M, Salman S A, Alwahhab A B A, Ali M. Enhancing agriculture crop classification with deep learning. Babylon J Artif Intell, 2024; 2024: 20–26.
[23] Li D F, Li B L, Long S F, Feng H Q, Xi T, Kang S, et al. Rice seedling row detection based on morphological anchor points of rice stems. Biosyst Eng, 2023; 226: 71–85.
[24] Wu S L, Chen Z G, Bangura K, Jiang J, Ma X G, Li J Y, et al. A navigation method for paddy field management based on seedlings coordinate information. Comput. Electron. Agric, 2023; 215: 108436.
[25] He J, Dong W H, Tan Q N, Li J N, Song X W, Zhao R M. A variable-threshold segmentation method for rice row detection considering robot travelling prior information. Agriculture, 2025; 15(4): 413.
[26] Chen Z Y, Cai Y H, Liu Y X, Liang Z P, Chen H, Ma R J, et al. Towards end-to-end rice row detection in paddy fields exploiting two-pathway instance segmentation. Comput. Electron. Agric, 2025; 231: 109963.
[27] Zhang S, Wang Y Y, Zhu Z X, Li Z, Du Y F, Mao E R. Tractor path tracking control based on binocular vision. Inf. Process. Agric., 2018; 5(4): 422–432.
[28] Fu D B, Chen Z Y, Yao Z Q, Liang Z P, Cai Y H, Liu C, et al. Vision-based trajectory generation and tracking algorithm for maneuvering of a paddy field robot. Comput. Electron. Agric, 2024; 226: 109368.
[29] Tao H Q, Zhang R R, Zhang L H, Zhang D Z, Yi T C, Wu M Q. Tea harvest robot navigation path generation algorithm based on semantic segmentation using a visual sensor. Electronics, 2025; 14(5): 988.
[30] Hernandez B, Medeiros H. Multi-object tracking in agricultural applications using a vision transformer for spatial association. Comput Electron. Agric, 2024; 226: 109379.
[31] Han X D, Fang P C, Tian Y Y, Yu J H, Cai X H, Roggen D, et al. GRASPTrack: Geometry-reasoned association via segmentation and projection for multi-object tracking. arXiv: 2508.08117, 2025. DOI: 10.48550/arXiv.2508.08117.
[32] Li H W, Lai X D, Mo Y M, He D Q, Wu T. Pixel-wise navigation line extraction of cross-growth-stage seedlings in complex sugarcane fields and extension to corn and rice. Front. Plant Sci, 2025; 15: 1499896.
[33] Zhai W G, Li C C, Fei S P, Liu Y H, Ding F, Cheng Q, et al. CatBoost algorithm for estimating maize above-ground biomass using unmanned aerial vehicle-based multi-source sensor data and SPAD values. Comput. Electron. Agric, 2023; 214: 108306.
[34] Guo Z M, Geng Y H, Wang C, Xue Y, Sun D, Lou Z X, et al. InstaCropNet: An efficient Unet-based architecture for precise crop row detection in agricultural applications. Artif. Intell. Agric, 2024; 12: 85–96.
[35] Zhang X Y, Lu X Y, Zhang Z Q, Yang G F, He Y, Fang H. Simultaneous detection of reference lines in paddy fields using a machine vision-based framework. Comput. Electron. Agric, 2024; 221: 108923.
[36] Yu C H, Feng J Z, Zheng Z Z, Guo J P, Hu Y H. A lightweight SOD-YOLOv5n model-based winter jujube detection and counting method deployed on Android. Comput. Electron. Agric, 2024; 218: 108701.
[37] Chen L C, Zhu Y, Papandreou G, Schroff F, Adam H. Encoder-decoder with atrous separable convolution for semantic image segmentation. 2018; DOI: 10.48550/arXiv.1802.02611.
[38] Kalman R E. A new approach to linear filtering and prediction problems. J. Basic Eng, 1960; 82: 35–45.
[39] Rapado-Rincón D, van Henten E J, Kootstra G. MinkSORT: A 3D deep feature extractor using sparse convolutions to improve 3D multi-object tracking in greenhouse tomato plants. Biosyst. Eng, 2023; 236: 193–200.
[40] Kuhn H W. The Hungarian method for the assignment problem. Nav. Res. Logist, 1955; 2: 83–97.
[41] Jang M, Hwang Y. Tomato pose estimation using the association of tomato body and sepal. Comput. Electron. Agric, 2024; 221: 108961.
[42] Sandler M, Howard A, Zhu M, Zhmoginov A, Chen L-C. MobileNetV2: Inverted residuals and linear bottlenecks. in: 2018 IEEECVF Conf. Comput. Vis. Pattern Recognit, 2018; pp.4510–4520. DOI: 10.1109/CVPR.2018.00474.
[43] Pal A, Leite A C, From P J. A novel end-to-end vision-based architecture for agricultural human–robot collaboration in fruit picking operations. Robot. Auton. Syst, 2024; 172: 104567.
[44] Guerrero J M, Guijarro M, Montalvo M, Romeo J, Emmi L, Ribeiro A, et al. Automatic expert system based on images for accuracy crop row detection in maize fields. Expert Syst Appl, 2013; 40: 656–664.
[45] Chen J, Wu J, Wang Z, Qiang H, Cai G, Tan C, et al. Detecting ripe fruits under natural occlusion and illumination conditions. Comput. Electron. Agric, 2021; 190: 106450.
[46] Zhang Y A, Wang T C, Zhang X Y. MOTRv2: Bootstrapping end-to-end multi-object tracking by pretrained object detectors. arXiv: 2211.09791, 2023. DOI: 10.48550/arXiv.2211.09791.
[47] Zheng Z Y, Li J W, Qin L F. YOLO-BYTE: An efficient multi-object tracking algorithm for automatic monitoring of dairy cows. Comput. Electron. Agric., 2023; 209: 107857.
[48] Jiang G Q, Wang Z H, Liu H M. Automatic detection of crop rows based on multi-ROIs. Expert Syst Appl, 2015; 42(5): 2429–2441.
[49] Zhang S L, Ma Q L, Cheng S K, An D, Yang Z L, Ma B, et al. Crop row detection in the middle and late periods of maize under sheltering based on solid state LiDAR. Agriculture, 2022; 12(12): 2011.
[50] Khan M N, Rahi A, Rajendran V P, Al Hasan M, Anwar S. Real-time crop row detection using computer vision- application in agricultural robots. Front. Artif. Intell, 2024; 7: 1435686.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 International Journal of Agricultural and Biological Engineering
This work is licensed under a Creative Commons Attribution 4.0 International License.
IJABE is an international peer reviewed, open access journal, adopting Creative Commons Copyright Notices as follows.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
