Distribution network line project acceptance is a key link in the quality control of distribution network line project, an important factor affecting the safe and stable operation of the distribution network, which directly determines the level of safe operation of the distribution network. In this paper, for the distribution network line manual acceptance time-consuming and laborious, rare quality defects found rate identification rate is low and other issues, to carry out visual positioning and image recognition based on the distribution network drone automated acceptance technology research. In order to optimize the spatial positioning, attitude sensing and target tracking of the UAV, five coordinate systems, including the world coordinate system, body coordinate system, and photocentric coordinate system, are selected for spatial transformation. Based on the visual localization of the UAV, the path planning algorithm for UAV distribution line inspection combined with the path acquisition scheme is proposed. Gaussian denoising and histogram equalization are performed on the UAV inspection collected images, and Sarsa reinforcement learning algorithm is applied to train the samples to improve the automatic identification capability of safety hazards and other security hazards in the distribution network inspection. Visualization and analysis of UAV distribution line inspection path. Combine the distribution network defects dataset for optimal training strategy selection for distribution networks. The automatic identification algorithm for distribution network defects proposed in this paper achieves a mAP value of 79.60% in the target detection experiment. And in multiple dynamic path planning, the UAV nodes are able to accomplish the path planning tasks in different environments.