Dynamic Vehicle Detection Algorithm Based on Background Updating and Suppressing

In the vision-based traffic system for the moving vehicle detection, the accuracy of vehicle detection is heavily based on exact acquirement of the background, this paper presents a T-distribution background reconstruction algorithm of moving vehicle detection to obtain background pixels, that is, the background image which doesn´t contain any moving objects is restored by integrating of several background images, and background image is reconstructed and updated constantly while the background is variational. Then the extract moving area is extracted by background suppression method in based on the RGB color space, and considering the environment´s change, the threshold of the three-channel´s noise variance of RGB image is added to the average value of the gray-level´s change. The experimental results show that the background reconstructed can reflect the real background and extract the more integrated complete moving area comparing with the frame-difference method.