Recognition of Pine Wood Affected by Pine Wilt Disease Based on YOLOv5

In order to recognize pine wood affected by pine wilt disease (PWD) efficiently, a dataset of infected pine wood was constructed and a recognition algorithm YOLOv5-PWD based on YOLOv5 was designed. In the mountainous area of Fushun City, Liaoning Province, an unmanned aerial vehicle (UAV) was used to collect images of infected pine wood at three different heights of 200/220m, 100m and 60/80m, and the images were expanded to establish an infected pine wood dataset. Based on the YOLOv5 framework, combined with Moasic data enhancement, Focus unit and spatial pyramid pooling (SPP) unit, a recognition algorithm for pine wood affected by pine wilt disease was designed. The experimental results show that the mean average precision of the dataset at the 100m collection height is the highest. Compared with the Faster R-CNN algorithm, the mean average precision of YOLOv5-PWD is increased by 16.8%, and the recognition speed is increased by 64.5%.