Detection and Identification of Peach Leaf Diseases based on YOLO v5 Improved Model

Aiming at the problem of peach tree leaf diseases in modern orchards, this work proposes a lightweight identification method based on the combination of improved YOLO v5 and ShuffleNet to achieve accurate identification of peach tree leaf diseases in the natural environment. In this work, the diseased leaves of peach trees are used as the data set for detection, and the quadratic gradient descent and attention mechanism are added based on traditional YOLO v5. The results show that the accuracy of the improved YOLO v5 is 5.06% higher than that before the improvement. We combined the improved model with two lightweight models, ShuffleNet and MobileNet, to make the model lightweight. The results show that after the lightweight improvement of the model, the size of the model has been significantly reduced, but the accuracy has also dropped slightly. Finally, the improvement of YOLO V5 +ShuffleNet can meet the actual needs of peach tree leaf disease identification, which can effectively solve the problem of modern peach orchard disease control.