Research On Algorithms Of Object Detection And Tracking In Video Surveillance

The moving object detection and tracking is one of the key technologies to realize intelligent visual surveillance system and the important research subject in the field of application for computer-vision.For many years,the technology of object detection and tracking has made a good progress.This dissertation focuses on the methods of diction and tracking,on the basic of the laboratory's single camera for target detection and tracking experiments,the experimental analysis use the own sports scene made.Main results and contributions of this dissertation are as follows:(1)Introduce some typical object detection algorithms,such as the frame difference method,the background subtraction.To go deep into analysis and research which by understanding their basic principle,research status and their advantages and disadvantages.To research the classification of the moving object tracking,the region-based tracking,the feature-based tracking,the active contour models and the model-based tracking.(2)According to the moving object detection in complex backgrounds,the paper proposes a kind of background modeling,taking the algorithms of the historical frames superposition which has weight on the algorithms fused the advantages of the frame difference method and the background subtraction.To avoid the background interferences and the lacking of the detection effect in environmental catastrophe,this dissertation proposed the corresponding solution.Based on the common shortcoming to environmental catastrophe is the fake background,this paper presented a moving object detection's method based on the color similarity coefficient to resolve the fake background.Meanwhile,explore a kind of improved type of a algorithm based on codebook's background model,under YUV color space.The experiments are used in the comparison because the algorithm has good prospect in the inspection of the effect.(3)Inveitably,video surveillance meet complex features such as the monitoring of the mult-objective.To distinguish the moving objects,realize continuous positioning for moving object,on the basic of getting foreground moving objects from video images.In this paper,use meanshift algorithm in the field of foreground detection which has achieved good effect based on tracking for the face of the color.