Modeling and Realizing of the Crowd Event
Ziqing Li · Microcomputer Information · 2010
In the field of video surveillance,the crowd and other abnormal events detection have a broad application prospects,however,the development of related research in China is still relatively slow. The abnormal event detection method based on the Hidden Markov Model are given here,a simple process just as follows:First,for the characteristics of crowd events video sequences,the dual feature vector extraction was completed on the basis of the Gaussian Mixture Model to detect the target; Then,Hidden Markov Model of the crowd events was trained using the Baum-Welch algorithm on condition that rational initial model had existed; In the end,its effectiveness was verified through the real shot of the video sequence.