The Solution to Background Segmentation and Extraction in Traffic Video Detection

This paper proposes a solution to the background segmentation and extraction in traffic video detection system. Weighted average method, Bimodal method and Median filter are employed to preprocess the video in the proposed solution. The preprocessing provides the satisfied precondition for the backgroud extraction. Furthermore, Gaussian Mixture Model is applied to extract the background from the grayscale image. The experiment results show the validity and the practicality of the solution. Its achievement will lay the foundation for the further work, such as extracting moving vehicle.