Research on Precision Enhancement of Wheat Pest and Disease Image Recognition Through Deep Learning Feature Fusion

This study proposes a method combining deep learning and feature fusion to improve wheat pest and disease image recognition accuracy. The uneven distribution of pests and diseases leads to varied image features, challenging traditional recognition methods. By using image segmentation to locate affected regions, the study extracts key features (e.g., edges and gradients) with deep learning models. Feature fusion then integrates multi-scale features to capture comprehensive information. The fused features are used to train deep learning models for precise classification. Experiments show that the proposed method achieves 99% - 100% accuracy and near-perfect F1 scores, significantly outperforming traditional methods. This approach provides robust support for early diagnosis and control of wheat pests and diseases.