Research of Video Tracking Algorithm Based on the Blob Analysis

摘要

To the video that contains the target, a method is proposed to create the background model based on the mixed Gauss. And the target locating method based on the blob analysis and blob filtering, the anti-noise ability and filter robustness of tracking is improved. The kalman filter and the particle filter are separately used to pass and update the foreground target's posterior probability distribution. Finally the kalman filter and the particle filter's are compared and that builds the foundation of the further development.