.png){kind=icon}
.png){kind=icon}
Return
WheatGOAT: Generalizable object-aware tracker via discriminative region semantic learning for wheat ear counting
.png){kind=icon}
PRE
AI
X
Y
L
Z
Y
G
A
G
X
DOI：10.1016/j.aei.2026.104762.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accurate counting of wheat ears is crucial for precision agriculture, aiding in yield estimation, breeding selection, and crop management. However, field conditions, such as occlusion, wheat variety variations, lighting changes, and background interference, make this task challenging. To address these issues, we propose WheatGOAT, a generalizable object-aware tracking framework that incorporates discriminative region semantic learning. This approach improves counting accuracy through three integrated modules. The Object–Background Selector (OBS) isolates wheat ear regions from background noise, enhancing feature refinement. The Object Perception Learner (OPL) boosts inter-class separability by learning semantic features at the region level, improving detection of densely packed and overlapping ears. The Hierarchical Refiner (HR) uses multi-scale contextual information to ensure precise recognition of wheat ears across different sizes and spatial arrangements. By focusing attention on semantically relevant areas, WheatGOAT reduces irrelevant activations and improves localization accuracy in challenging environments. Extensive tests on the GWHD_2021 dataset show WheatGOAT outperforms existing methods, with 15.3% improvements in MAE and 13.9% in RMSE. With strong generalization and semantic alignment, WheatGOAT offers a versatile solution for intelligent crop monitoring, with potential applications in yield estimation, phenological analysis, and precision agriculture. Codes are available at https://wheatgoat.samlab.cn/ .
Keywords:
Wheat ear counting
Object-aware tracking
Discriminative region semantic learning
Precision agriculture
Crop monitoring
Journal
IF:
9.9
Papers:
4.0K
Citations:
1.7W
Organization
Cited Papers