Agri-PIDNet-S: Lightweight semantic segmentation for field-structure and obstacle perception in seedling-stage cotton fields
PLOS One · 02.11.2026 · 21 (11) · e0361992
DOI: 10.1371/journal.pone.0361992 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, online erschienen am 02.11.2026. Verfasst von Yanlin Xin und 5 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer and Information Engineering, Xinjiang Agricultural University, Urumqi, China. Hauptthemen laut MEDLINE: Gossypium, Seedlings, Image Processing, Computer-Assisted. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Educational and Teaching Master Project under the Third Batch of the Tianshan Talent Training Program, Central Government Guiding Local Science and Technology Development Special Fund Project.
Abstract (Originalsprache)
Reliable joint perception of crop rows, furrows, and obstacles is essential for autonomous inter-row operations in seedling-stage cotton fields. However, small and sparsely distributed plants, perspective-convergent rows, irregular furrow boundaries, and low obstacle pixel coverage jointly challenge fine-detail preservation, long-range structural modeling, and boundary discrimination under real-time constraints. This study formulated cotton-field perception as a four-class semantic segmentation task and constructed a dataset of 709 pixel-wise annotated images comprising background, crop rows, furrows, and obstacles. We developed Agri-PIDNet-S, a field-specific lightweight network based on PIDNet-S. A coordinate-aware multi-kernel inverted residual module (CA-MKIR) preserves shallow details and small-obstacle contours, a squeeze-and-excitation-enhanced deep-to-token module (SaE-D2T) models long-range row and furrow structures, and a boundary-guided window-attention refinement Bag (BWR-Bag) improves semantic–boundary fusion. The network was trained with a composite loss incorporating auxiliary boundary supervision. On the cotton-field test set, Agri-PIDNet-S achieved 92.37% mean intersection over union (mIoU) and 96.02% mean Dice coefficient (mDice) in the model-comparison experiment, while repeated training with three seeds yielded an mIoU of 92.35 ± 0.03%. Relative to PIDNet-S under the same comparison protocol, mIoU increased by 3.53 percentage points, while obstacle IoU improved from 79.60% to 88.67%. The model contained 8.048 million parameters, required 19.161 GFLOPs, and achieved 155.67 frames per second. Zero-shot external evaluation on tomato, watermelon, and maize fields yielded a structural mIoU of 86.29% for the background, crop-row, and furrow classes. Under the evaluated conditions, the task-oriented modifications achieved higher segmentation point estimates than PIDNet-S while maintaining computational efficiency; the external results provide preliminary evidence of cross-crop structural transfer.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- College of Computer and Information Engineering, Xinjiang Agricultural University, Urumqi, China.
- Engineering Research Center of Intelligent Agriculture, Ministry of Education, Urumqi, China.
- Xinjiang Agricultural Informatization Engineering Technology Research Center, Urumqi, China.
ORCID-Profile: Yanhong Chen
Förderung
- Educational and Teaching Master Project under the Third Batch of the Tianshan Talent Training Program
- Central Government Guiding Local Science and Technology Development Special Fund Project
Zitieren
APA
Xin, Y., Li, Y., Wei, W., Cao, X., Mao, B., & Chen, Y. (2026). Agri-PIDNet-S: Lightweight semantic segmentation for field-structure and obstacle perception in seedling-stage cotton fields. PLOS One, 21(11), e0361992. https://doi.org/10.1371/journal.pone.0361992
Vancouver
Xin Y, Li Y, Wei W, Cao X, Mao B, Chen Y. Agri-PIDNet-S: Lightweight semantic segmentation for field-structure and obstacle perception in seedling-stage cotton fields. PLOS One. 2026;21(11):e0361992. doi:10.1371/journal.pone.0361992
BibTeX
@article{Xin2026,
author = {Xin, Yanlin and Li, Yongke and Wei, Wuchen and Cao, Xinyu and Mao, Bowen and Chen, Yanhong},
title = {Agri-PIDNet-S: Lightweight semantic segmentation for field-structure and obstacle perception in seedling-stage cotton fields},
journal = {PLOS One},
year = {2026},
volume = {21},
number = {11},
pages = {e0361992},
doi = {10.1371/journal.pone.0361992}
}
EUVIMED