In this paper, the Unmanned Aerial Vehicle(UAV)equipped with a high-resolution RGB digital camera was used to catch high spatial resolution aerial images over the pines. The acquired visible orthophoto images were pretreated with elevation and topo-graphic features and had a classification and identification training with ResNet artificial neural network after extraction of texture infor-mation. Finally, a deep convolutional network-trained model was adopted for intelligent identification of discolored nematodiasis infect-ed pines. The results showed that the average identification accuracy rate was 92.29％with the highest rate reaching 96.51％. The feasi-bility of studying the identification method for discolored pinewood nematodiasis infected pines based on ResNet model wood was syn-chronously verified. The study is aimed to provide reference for pinewood nematodiasis control.
维普
万方数据