Abstract: To achieve automatic identification of defect images in transmission lines by UAV,a method based on prior knowledge faster region convolutional neural network（Faster R-CNN）for transmission line UAV image recognition is proposed. Faster R-CNN image recognition model is improved based on the design of prior boxes and the concept of transfer learning,effectively enhancing the model’s recognition accuracy and generalization capability. The experimental results show that the proposed method can identify faults accurately and quickly under different shooting conditions and fault types,and has excellent identification performance.