A defect recognition method for unmanned aerial vehicle inspection images of transmission lines based on Cascade R-CNN algorithm is proposed.Firstly,preprocess the inspection images, including denoising,enhancement,and normalization,to improve image quality.Then, the Cascade R-CNN algorithm was improved by selecting ResNeXt152 as the backbone network,which enables the network to learn richer feature representations and focus on different types of defect features, thereby utilizing its multi branch approach to enhance detection performance; The second is to introduce a recursive feature pyramid structure for feature fusion and hierarchical prediction, which can simultaneously capture defect information of different scales and improve defect recognition accuracy;Thus, the improved Cascade R-CNN algorithm is applied to identify defects in transmission lines. The experimental results show that the proposed method has high defect recognition accuracy, low complexity, and good recognition effect, which can provide technical support for the safe operation of transmission lines.