| In recent years, with the increase of market demand, people’s security awareness and the improvement of government support, the security industry is in a rapid stage of development. The rapid development of the security industry has brought enormous challenges to video surveillance, which can be found everywhere. All the cameras for video surveillance are working day and night, monitoring and recording to improve social security but producing massive recording video information. How to save video surveillance data has brought great pressure to the user store. And at the same time, how to retrieve the information that we are interested in from the huge amount of video data is a time-consuming and labor-intensive work. Thus, How to solve these problems is a hot spot in modern security industry and analysis of surveillance videos.This paper focuses on the analysis of the original video offline, trying to design and implement a framework of analyzing and monitoring video engineering, which aims to relieve the massive pressure of video surveillance storage data to users and improve the efficiency of retrieval of surveillance videos. The main work of this paper can be divided into four parts:(1) The static and dynamic segmentation of surveillance videos: the static videos of no use are separated from the original surveillance videos and only the dynamic videos containing moving targets and a few static videos are saved. In this way, the saved useful video data is convenient for later evidence and can alleviate the pressure of user storage.(2) Extraction of moving objects in video surveillance: the effect of moving objects extraction has an important influence on the follow-up feature extraction and target tracking. This thesis adopts the background modeling method based on temporal spatial texture features, the combination with the frame difference and level set method in order to extract moving targets.(3) Form a database of objects’ features:extract moving objects’ color features,texture features and moving information.(4) Implement a prototype system of searching video based on the features of moving objects. The system consists of a data generation subsystem and a data query system. The former is responsible for the structural scheme of video, and the latter is responsible for the searching of video. |
