PPNet: a deep learning framework for rice seedling localization and distribution from UAV orthomosaics
4 affiliations
- 1College of Information and Electrical Engineering, China Agricultural University
- 2Research Center of Intelligent Equipment, Beijing Academy of Agriculture and Forestry Sciences
- 3State Key Laboratory of Intelligent Agricultural Power Equipment
- 4Research Center of Information Technology, Beijing Academy of Agriculture and Forestry Sciences
Abstract
Field-scale information on rice seedling distribution is crucial for transplanting quality assessment, site-specific seedling-stage management, and precision agricultural operations. However, seedlings in UAV orthomosaics are small, closely spaced, and easily confused with soil, shallow water, plastic mulch, and shadows, which makes accurate localization challenging. Meanwhile, under high-density planting, existing segmentation-based networks often generate adhesive responses between neighboring seedlings, leading to missed detections and biased counting. To address these issues, this study proposed PPNet, a deep learning framework for rice seedling localization and distribution mapping from UAV orthomosaics. The framework uses root-based point annotation and integrates a rice-point perception module (RPPM), an adjacent decoupling module (ADM), and an aggregated attention module (AAM). These modules enhance seedling-point responses, separate adjacent seedlings, and introduce row-structure information. Experiments were conducted on a UAV orthomosaic dataset containing 3,013 image tiles and 220,553 seedling point annotations from different rice-growing regions. PPNet achieved an F1 score of 0.9029, an AP of 0.8881, a counting MAE of 6.99, and an R2 of 0.9553. The method also generated field-scale seedling point maps and density maps under different planting densities and field backgrounds. The results demonstrate that PPNet can provide reliable seedling localization and counting information for crop establishment assessment and targeted seedling management, supporting intelligent agricultural machinery decision-making in precision farming operations.
Related
- Automated detection, morphometric measurement, and size-class determination of harbor seals (Phoca vitulina) from UAV orthomosaics: a cascaded deep-learning pipeline with stratified validation across platforms, sensors, altitudes, and substrates
- Label-efficient seedling perception via semi-supervised learning and counterfactual spatiotemporal feature rectification
- From localization to distribution: revisiting the functional organization of the extrastriate body area
- A machine-vision framework for electric vehicle charging hole localization based on contour-driven detection and geometric verification
- Subcellularly Resolved Single-Cell Embedding Learning with Transcriptomic data, Protein Structure and Localization Information
BibTeX
@article{shan2026ppnet,
title = {{PPNet: a deep learning framework for rice seedling localization and distribution from UAV orthomosaics}},
author = {Yongchao Shan and Anqi Zhang and Xiaofei An and Jianhui Nie and Zhicheng Zhang and Chunjiang Zhao and Jingping Chen and Weiqiang Fu},
journal = {Frontiers in Plant Science},
year = {2026},
month = oct,
doi = {10.3389/fpls.2026.1932877},
url = {https://www.frontiersin.org/articles/10.3389/fpls.2026.1932877/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - PPNet: a deep learning framework for rice seedling localization and distribution from UAV orthomosaics
AU - Yongchao Shan
AU - Anqi Zhang
AU - Xiaofei An
AU - Jianhui Nie
AU - Zhicheng Zhang
AU - Chunjiang Zhao
AU - Jingping Chen
AU - Weiqiang Fu
PY - 2026
DA - 2026/10/13
JO - Frontiers in Plant Science
DO - 10.3389/fpls.2026.1932877
UR - https://www.frontiersin.org/articles/10.3389/fpls.2026.1932877/full
AB - Field-scale information on rice seedling distribution is crucial for transplanting quality assessment, site-specific seedling-stage management, and precision agricultural operations. However, seedlings in UAV orthomosaics are small, closely spaced, and easily confused with soil, shallow water, plastic mulch, and shadows, which makes accurate localization challenging. Meanwhile, under high-density planting, existing segmentation-based networks often generate adhesive responses between neighboring seedlings, leading to missed detections and biased counting. To address these issues, this study proposed PPNet, a deep learning framework for rice seedling localization and distribution mapping from UAV orthomosaics. The framework uses root-based point annotation and integrates a rice-point perception module (RPPM), an adjacent decoupling module (ADM), and an aggregated attention module (AAM). These modules enhance seedling-point responses, separate adjacent seedlings, and introduce row-structure information. Experiments were conducted on a UAV orthomosaic dataset containing 3,013 image tiles and 220,553 seedling point annotations from different rice-growing regions. PPNet achieved an F1 score of 0.9029, an AP of 0.8881, a counting MAE of 6.99, and an R2 of 0.9553. The method also generated field-scale seedling point maps and density maps under different planting densities and field backgrounds. The results demonstrate that PPNet can provide reliable seedling localization and counting information for crop establishment assessment and targeted seedling management, supporting intelligent agricultural machinery decision-making in precision farming operations.
SN - 1664-462X
ER -