Abstract:

To address the challenges of small target size, dense distribution and occlusion among winter wheat ears in open field environments, this study focused on winter wheat captured by UAV imagery and proposed an improved detection method based on the YOLOv8 model. The SimAM attention mechanism was introduced into the Neck (Neck network) while the GhostNetV2 module was integrated into the C2f module within the Neck. These enhancements improved the representation of spatial and channel features, while maintaining efficient feature fusion and reducing model complexity. As a result, the detection network was better adapted to the complex conditions of open field winter wheat ear detection. In addition, the input image resolution was set to 1280px×1280px to maximize the preservation of critical visual features. The results showed that the improved YOLOv8 model achieved an average precision (AP) of 93.1% and an F1 score of 90.5%, with a model size of only 18.3MB and 9.4 million parameters. Compared to the original YOLOv8, the improved version yield increased of 0.5 percantage point and 0.8 percantage point in AP and F1 score, respectively, while reducing the model size and parameter counted by 3.3MB and 1.7 million parameters. The resulting model is more lightweight and efficient, outperforming the standard YOLOv8 in detecting small, densely distributed and highly occluded winter wheat ears under complex field conditions.

Key words:

Winter wheat ear, YOLOv8, Model light weighting, Target detection, Attention mechanisms