Apple Leaf Pathology Detection Based on Improved EfficientNetV2

To address the problems of time-consuming and labor-intensive, poor accuracy and inefficiency of apple leaf pathology detection, a lightweight model incorporating the dilated convolution as well as CBAM (Convolutional Block Attention Module, CBAM) attention mechanism is proposed based on the EfficientNetV2 structure. The model improves the EfficientNetV2 main structure by introducing dilated convolutional coordinate attention and coordinates attention CBAM to enhance the semantic information transfer, expand the perceptual field, and improve the model's ability to extract global features. The experimental results show that the model achieves 98.98% of recognition accuracy, which is about 4% higher than the original model. The improved model size is 84.6 MB, which increases the memory capacity by 0.8 MB compared with the original model. To further demonstrate the impact of the improved modules on the network performance, the effectiveness of the module improvement is verified by ablation experiments. Compared with other excellent models, the performance of the improved model is better than other similar classification models.