Video Background Modeling Based On Improved Gaussian Mixture Model Algorithm

With the development of computer visual technology, the theory and algorithm of video processing make advances. Foreground object extraction is one of the fundamental tasks for video analysis, which is also topic widely researched in the last decade around the world. In this paper, we focus on the background modeling of complex environment and foreground object detection, and place emphasis on Gaussian Mixture Model(GMM), the propose an improved Gaussian Mixture Algorithm and use OpenCV from Intel to test.The Gaussian Mixture Model is a common model in background modeling. It is not the best model in practice that each pixel is modeled a fixed number of Gaussian distributions for the original GMM algorithm. In this paper, an effective adaptive background updating method is presented by choosing the number of components for each pixel in an on-line procedure. And the threshold value is also given a flexible selection to decide whether the pixel is belong to background or not. The algorithm can automatically adapt to the scene.OpenCV is an open source computer visual library developed by Intel, include amount of functions in image processing. This paper will show you how to read, process and store video sequences by using OpenCV. We use the basic framework of video processing to realize two algorithms about foreground object detection and make a comparison.The vast majority of algorithms process images on a pixel-by-pixel basis, where an independent decision is that rich contextual information is not taken into account. We propose a block-based method capable of dealing with noise, illumination variations, and dynamic backgrounds, while still obtaining smooth contours of foreground objects. Specifically, image sequences are analyzed on an overlapping block-by-block basis. A low-dimensional texture descriptor obtained from each block is passed through an adaptive classifier cascade. A probabilistic foreground mask generation approach then exploits block overlaps to integrate interim block-level decisions into final pixel-level foreground segmentation. Experiments show that the proposed approach obtains better results than several prominent methods.