Partial Occlusion Of Video Based On Inner Contour Restoration Research And Application Of Target Discrimination Method

In the process of target monitoring and recognition,it is very important to solve the problem of the defect caused by the occlusion of the target,because of the complexity and diversity of the monitoring environment.In this paper,a method that can distinguish the partially occluded objects is proposed,which is based on the inner contour restoration.This method can solve the problem of the detection error caused by the occlusion in the existing target detection algorithm.The main work is as follows:(1)In the case of partial occlusion,the KNN background segmentation device of background subtraction algorithm is used to segment the background,and the moving object foreground is separated from the video stream.The improved Otsu threshold method is used to calculate the high threshold value,and the video frame is binarized,and the wavelet threshold value is used to denoise to complete the image preprocessing.(2)The edge contour is extracted based on morphology,then t-node is detected by the Kona template.Finally,the target contour image we are going to repair is extracted after segmenting the image of t-node detection by the reconstructed watershed segmentation algorithm based on color model.(3)Through the analysis of connected domain,all contour nodes are gained after traversing the image by the marking algorithm based on contour.Then according to the Gestalt completion rules,a judgment mechanism is set up to establish the connectable relationship around the internal contour nodes and group them to complete the repair of the internal contour and identify the occluded objects effectively.Finally,this method is successfully applied to the on-the-job situation detection of the personnel in the duty room.