A new segmentation method for image sequence of traffic scenes

This paper describes a novel segmentation method for image sequence of traffic scenes recorded by a stationary camera. The process of segmentation includes two stages. In the first stage, the frame difference of the current frame is binarized, and the binary image of the frame difference is processed by morphological operation, then the current frame updates the background according to the binary image processed by morphology, and the background is extracted after dozens of iterative operations of the subsequent frames. In the second stage, the process of every frame is as follows: the noise parameters of the camera are derived from the histogram of the difference image between the current frame and background, and the double thresholds which segment the difference image are acquired from the histogram, and the difference image is binarized by the thresholds, then moving objects are found and the background is updated according to the segmentation result. The experiment results show that the method can realize segmentation effectively and quickly.