The Research Of Generate Video Synopsis

With the rapid progress of society,people on Social Security made increasing demands.Nowadays,various public places and living quarters have installed surveillance cameras with the rapid development of mass storage technology and surveillance video technology.In this case,a massive video arose.Mass surveillance video can help us understand what has happened more detailedly,but it makes people browsing becomes more time-consuming.Therefore,using computer vision techniques to extract valid information from mass visual information instead of doing so artificially is imminent.Video synopsis can analyze the structure and content of the video,and extract a meaningful portion of the video from the original video and then combining them in some way,and regenerating synopsis to represent the original video.Video synopsis is much shorter than the original video.One need only browsing the video synopsis can understand the original video information,and facilitate the realization of the video content quickly browse and retrieval.Thus it saves a lot of human and financial resources.A method of generating a video synopsis in this paper includes three aspects.First of all,the original video moving target detection.Second,tracked all moving objects and extracted their trajectory data and screening and excluding the interference of the moving target trajectory and the normal trajectory planning of moving target track on the timeline makes the spatial and temporal redundance between them as small as possible.The last,add to the plan complete the movement of the target data pixels corresponding background information and synthesis and produce the final video synopsis.This paper proposes a moving target tree storage structure.For multiple moving targets with shade and adhesion to extract the moving target trajectory tracking.Using the tree structure on the original video related to track moving targets in adhesion,it can maintain the integrity of the multiple moving targets effectively.Experiments show that the video synopsis which use this paper’s method to generate have High compression rate and the processing speed is fast.This method can meet the requirement of real-time processing and have a certain practical prospect.