The Design And Achievement On Static Background Extraction In Video Stream

Multi-objects detection and the video mosaic are the core of picking-up the immobilebackground pictures of video. The investigation on these key technologies has the greatmeaning of improving the reliability and robustness of picking-up the immobile backgroundpictures.The technology of video mosaic is indispensable to it.This paper studies the problems of traditional objects detection and tracking in practicalapplications,and pay attention to four parts of tne video mosaic.in order to apart the movingand the immobile background,achieved background modeling.and dynamic things,and buildbackground modeling.It can stand out the moving things’ outline and cleared know the waywhat the moving line.The main research contents include five aspects.(1)In order to solving the problems of Multi-objects detecting with algorithm ofGaussian mixture models(GMM), studing the background.Lessoned calculatedamount.Improved instantaneity. the proposed algorithms proved more computing efficientand Lower RAM consumption, and can detect objects with low speed or short stay.(2) This paper studies some aspects of technique on the images mosaic.About collectionof video,pick-up the character, the characters matching and panorama exported. Introducedthe method of dealing with edges.The papers adopt the SIFT and RANSAC in order to exportthe panorama fast and exactly.(3) Application of existing video splicing technology, dynamic target trackingtechnology are fused to form a dynamic target can be tracked to the final output of theexistence of the the dynamic target final state of the static panorama. OpenCV-based tool, theapplication of the two studies content to the actual video stream images and real-time trackingof a moving target in a single camera point tracking method to track the target and to outputthe final panoramic image, this paper also applied sport history Figure screening and detectiontechnology for dynamic target, some mobile video streaming with dynamic target targetdetection and a static background extraction software system combined with SIFT andRANSAC algorithm.(4) The system in3D is a new attempt in the existing technology, and not just focus onthe dynamic target, background stitching background modeling and refactoring study, notonly in this projectin certain applications in the field of application in panoramic videosurveillance, Google Street View, the video stream image processing can get pointed in a newdirection of the video stream static background generation system software requirements,functional requirements, the choice of algorithm and architecture, design thinking, the construction ofthe system model to the overall design. Looking to make a summary and expectations of thesystem applications and prospects.(5) on the static background generation system for streaming video software requirements,functional requirements, algorithm and architecture, the design idea, constructs the systemmodel for the overall design. Prospects for the application of the system and the prospect ofthe summary and expectation...