Mobile target detection of intelligent surveillance system based on modified GVF

The current methods of mobile target detection of video sequences have shortcomings such as poor disturbance-resistance and robustness. To solve this problem, an modified GVF Snake algorithm was presented, new gradient vector field was founded which took full advantage of image gradient information, between-frame motion information and grey-level information of neighbor pixels. Simulation result indicates that this modified method can detect mobile target of video sequences efficiently, and improve the disturbance-resistance ability and robustness.