Through the on-site survey video analysis application system, a large number of personnel are required to watch the process with the naked eye and conduct frame by frame screening. Through the system's automated processing, the risks of missed viewing and screening in the manual screening process are solved, and the harm to the body of video surveillance police is reduced. At the same time, by analyzing the information such as portraits, vehicles, features, and behaviors that appear in the video, and combining it with case management, a correlation analysis of the evidence chain is formed. And strengthen the images in the video and use video slicing techniques to identify key node slicing images and data.