An intelligent surveillance system based on RANSAC algorithm

Crowd control and management is a very important task in public places. Historically, many crowd disasters happened because of the loss of control of the crowd flow direction. This paper presents an intelligent surveillance system based on RANSAC (Random Sample Consensus) algorithm, which can estimate the crowd flow direction and classify people into different crowd groups. We calculate the optical flow by employing the "pyramidal" Lucas-Kanade(LK) algorithm. Then foreground detection is used to reduce the observation noise. RANSAC provides a simple but effective method to reduce the influence of the outliers in optical flow images, and estimate the crowd flow direction with the inliers. According to the differences between motion direction and position, we classify people into different crowd groups. Experiments on real crowd videos captured at different public places show the effectiveness of the proposed system.