Research of Moving Target Detection and Tracking Based on Background Difference and CamShift

As one of the hot spots in the field of computer vision research, video object tracking has been widely applied to various fields of science and technology, national defense construction, security monitoring and other national economy, and has great practical value and broad prospects for development. In the aspect of target detection, the motion detection algorithm based on mixed Gaussian background modeling is mainly researched. For Gaussian mixed model parameter initialization, an improved method for initializing the Gaussian background model is proposed. Then, the pixel value of each pixel in the newly input image and the Gaussian distribution of each pixel are detected and matched, and the model is updated. The new Gaussian distribution generation criterion is developed to improve the Gaussian mixture model updating and the foreground pixel detection. For moving target tracking, based on the study of the histogram of the color histogram and the gray gradient histogram, the adaptive multi-feature fusion algorithm is proposed, by calculating the difference between the foreground and the background of each feature of the target, a fusion model of the improved target feature is obtained, which enhances the robustness of the algorithm in multi-vision feature fusion tracking under complex scenes.