Aiming at the problems of low accuracy and long time consuming in the existing inspection image defect detection methods of transmission line UAV, in order to realize the fast and accurate identification of the bird's nest of transmission line tower, an improved YOLO 4 model is proposed for the bird's nest detection of transmission line tower image based on the UAV inspection image acquisition and processing system. The light MobileNetV2 network is used to replace the CSPParkNet53 network, which improves the speed of feature extraction. The average pooling is used in the SPP module to replace the maximum pooling, which improves the detection accuracy of the algorithm for small targets. The attention mechanism CBAM is introduced to enhance feature expression.The feasibility and superiority of the proposed method are verified by experiments.The results show that the proposed method has higher detection accuracy and faster detection speed in the transmission line tower image defect detection compared with the conventional detection methods, the detection accuracy reaches 94.40%, and the detection speed is 60 FPS.This study provides a reference for the development of the defect detection methods of transmission line towers.