Tomato Pests and Diseases Identification Tool Based on Improved YOLOv5

In order to accurately and conveniently identify tomato pests and diseases, this paper developed a tomato pests and diseases recognition tool software using PyQt5 on the Windows platform. The software is based on an improved YOLOv5 object detection method, integrates coordinate attention mechanism into the YOLOv5 backbone network to improve network feature extraction ability, introduces depthwise separable convolution to replace traditional convolution and reduce network model size, achieving good training results. The results shows that the accuracy of the tomato pests and diseases identification software can reach 96.8%, and the model size is reduced by 51.4% compared to YOLOv5, achieving a balance between performance and parameter quantity.