The algorithm of moving human body detection based on region background modeling
Tao Lei, Yangyu Fan, Liandong Li · 2009
A novel moving human body detection algorithm based on background modeling is proposed. It constructs a moving region with the difference image of adjacent frames, and constructs the second moving region with the difference image of following adjacent frames, so that the regions of background update are acquired by computing the difference between the moving regions and the intersection of front and behind moving regions, and they are updated according to the regions' stability probability. Based on the constructed background model, the moving human body is segmented with the combination of the maximum color difference and the adaptive threshold. The experiment result shows that the novel algorithm can not only establish the background model effectively, but also detect the moving human body perfectly with high robustness and real time in the complex environment.