Research On Moving Object Detection And Tracking Technology In Video Sequences

Moving object extraction and tracking technology has extensive application prospects, thus a large number of researchers devote themselves in the field and have already achieved many productions.Based on these productions,this thesis makes some researchs and attempts on background modeling,object extraction,shadow detection,object tracking in video sequences, the main tasks are as follows:Acquiring video sequences containing moving objects through the VFW software development kit is the precondition. To extract the moving object accurately, background differential method is selected as the basis for object extraction, OTSU algorithm is applied to determine the threshold value dynamically which is used to binarize the difference image, and acquire the moving object region and background region; This paper extracts background with the median method under the conditions of the analysis of relevant background modeling methods,and selects weight to update background real-timely according to the coordinates of foreground and background. To detect the shadow areas accurately, eight-neighborhood checking criterion is proposed based on the brightness and normalization criterion of RGB color space shadow detection algorithm. The pixel value of shadow area is replaced with the corresponding pixel of background to remove the shadow.Filter out the image noise and non-continuous edges through the morphological processing.Improve object-matching algorithms to construct grade matching model and realize the object tracking with histogram template matching.Experimental results show that the method can extract a clear background, effectively suppress the Interference of shadows and noise to object matching and increase the probability of a correct match. Improved search matching strategy can quickly locate moving objects and improve the matching speed.Experimental results show that the method can extract a clear background, effectively suppress the Interference of shadows and noise to object matching and increase the probability of a correct match. Improved search matching strategy can quickly locate moving objects and improve the matching speed.