Crowd Security Detection based on Entropy Model
Ying Zhao, Mengqi Yuan, Guofeng Su, Tao Chen
Identifying the terror attack, illegal public gathering or other mass events risks by utilizing cameras is an important concern both in crowd security area and in pattern recognition research area. This paper provides a physical entropy model to measure the crowd security level.The entropy model was created by identifying individuals’moving velocity and the related probability. The individuals are represented by Harris Corners in videos, thus to avoid the time-consuming human recognition task. Simulation experiment and video detection experiments were conducted, verified that in the disordered state, the entropy is higher; while in ordered state, the entropy is much lower; when the crowd security has a sudden change, the entropy will change. It was verified that the entropy is the applicable indicator of crowd security. By recognizing the entropy mutation, it is possible to automatically detect the abnormal crowd behavior and to set the warning alarm.