With the continuous expansion of power demand and China's power grid, the demand for power inspection is also expanding. Traditional power inspection has been unable to meet the needs of the market and industry, and the power inspection drone market is currently very broad. However, at present, the inspection UAV on the market only has the function of photographing, has no autonomous cruise capability, relies on manual flight control and photographing, has no real-time acquisition capability of inspection data, and still needs manual screening of inspection problems after inspection. All in all, the current market lacks a set of unmanned aerial vehicle inspection platforms with autonomous inspection, identification, real-time data processing and real-time data uploading for high-voltage power. After on-the-spot visits to the power industry, when the task of machine patrol is heavy, the number of machine patrol photos that need to be processed is very large, and thousands of photos are considered small. In order to match the inspection photos with the picture information, people in the industry often do not tire of renaming one photo after another, and it is extremely difficult to correct and smooth it after the accident. At present, the industry still lacks a comprehensive data management platform to manage and visualize inspection data. The product positioning of the project research and development is clear, which is to develop a high-voltage power comprehensive inspection platform based on deep learning and 5G edge computing for the pain points of inspection drones currently on the market. An intelligent UAV integrated inspection platform that can replace manual climbing inspection and can efficiently and accurately perform various inspection work of UHV transmission facilities. The drone will conduct autonomous or controlled inspections of power facilities in an outdoor environment where UHV transmission lines are erected. The drone will be equipped with an optoelectronic camera to obtain real-time image information of the target power facility, as well as an edge computing device and deploy a deep learning target detection model to detect the target power facility. In order to realize the identification of insulator pollution, insulator damage, foreign matter in high-voltage power towers and other problems, the project collected a large number of real images taken in actual high-voltage power operation, and manually or semi-manually labeled all images to make data sets into models for training, and the trained models were deployed to edge computing devices through format conversion and code rewriting. In order to realize the remote data transmission of unmanned aerial vehicles and the real-time report information generated after the inspection is completed, 5G communication technology will be used to realize the transmission of data to the web side. At the same time, a comprehensive data management platform is established, which can carry out inspection data processing and provide intuitive inspection data for staff. The construction of the above-mentioned inspection unmanned aerial vehicle and comprehensive data management platform can establish a comprehensive inspection platform with high automation, artificial intelligence participation and no need for a large amount of manual intervention for the inspection of UHV and HV power transmission lines and equipment!