Tree-Guided Transformer for Sensor-Based Ecological Image Feature Extraction and Multitarget Recognition in Agricultural Systems
Sensors · 07.10.2025 · 25 (19) · 6206
DOI: 10.3390/s25196206 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 07.10.2025. Verfasst von Yiqiang Sun und 6 weiteren Autor:innen. Beteiligte Einrichtung: China Agricultural University, Beijing 100083, China. Hauptthemen laut MEDLINE: Agriculture, Image Processing, Computer-Assisted, Ecology. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 48 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Apple Industry Upgrade and Product Capability Enhancement Project, National Key Research and Development Program of China, National Natural Science Foundation of China.
Abstract (Originalsprache)
Farmland ecosystems present complex pest–predator co-occurrence patterns, posing significant challenges for image-based multitarget recognition and ecological modeling in sensor-driven computer vision tasks. To address these issues, this study introduces a tree-guided Transformer framework enhanced with a knowledge-augmented co-attention mechanism, enabling effective feature extraction from sensor-acquired images. A hierarchical ecological taxonomy (Phylum–Family Species) guides prompt-driven semantic reasoning, while an ecological knowledge graph enriches visual representations by embedding co-occurrence priors. A multimodal dataset containing 60 pest and predator categories with annotated images and semantic descriptions was constructed for evaluation. Experimental results demonstrate that the proposed method achieves 90.4% precision, 86.7% recall, and 88.5% F1-score in image classification, along with 82.3% hierarchical accuracy. In detection tasks, it attains 91.6% precision and 86.3% mAP@50, with 80.5% co-occurrence accuracy. For hierarchical reasoning and knowledge-enhanced tasks, F1-scores reach 88.5% and 89.7%, respectively. These results highlight the framework’s strong capability in extracting structured, semantically aligned image features under real-world sensor conditions, offering an interpretable and generalizable approach for intelligent agricultural monitoring.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- China Agricultural University, Beijing 100083, China.
- National School of Development, Peking University, Beijing 100871, China.
- Tsinghua University, Beijing 100084, China.
- State Key Laboratory of Agricultural and Forestry Biosecurity, China Agricultural University, Beijing 100083, China.
ORCID-Profile: Shuo Yan
Förderung
- Apple Industry Upgrade and Product Capability Enhancement Project
- National Key Research and Development Program of China
- National Natural Science Foundation of China
- Pinduoduo-China Agricultural University Research Fund
- 2115 Talent Development Program of China Agricultural University
Zitieren
APA
Sun, Y., Huang, Z., Yang, L., Wang, Z., Ruan, M., Suo, J., & Yan, S. (2025). Tree-Guided Transformer for Sensor-Based Ecological Image Feature Extraction and Multitarget Recognition in Agricultural Systems. Sensors, 25(19), 6206. https://doi.org/10.3390/s25196206
Vancouver
Sun Y, Huang Z, Yang L, Wang Z, Ruan M, Suo J, et al. Tree-Guided Transformer for Sensor-Based Ecological Image Feature Extraction and Multitarget Recognition in Agricultural Systems. Sensors. 2025;25(19):6206. doi:10.3390/s25196206
BibTeX
@article{Sun2025,
author = {Sun, Yiqiang and Huang, Zigang and Yang, Linfeng and Wang, Zihuan and Ruan, Mingzhuo and Suo, Jingchao and Yan, Shuo},
title = {Tree-Guided Transformer for Sensor-Based Ecological Image Feature Extraction and Multitarget Recognition in Agricultural Systems},
journal = {Sensors},
year = {2025},
volume = {25},
number = {19},
pages = {6206},
doi = {10.3390/s25196206}
}
EUVIMED