Moving Object Detection Algorithm of Improved FBS-ABL
An improved FBS-ABL algorithm is proposed to solve the problem that Fast Background Subtraction(FBS-ABL)is prone to ghost-ing in complex environments.The algorithm based on the foreground target detected by FBS-ABL algorithm,after find and match the pixel histogram of foreground object and its neighborhood,then determine whether there is a"ghost"in the target area and eliminate the ghost by changing the pixel values of target area.Meanwhile,reinitializing the background model for area that ghost object may be exist to suppress the ghost.Finally,the improved algorithm is validated through two different scenes,the result show that the proposed algorithm can effectively eliminate ghost and significantly increase the accuracy of moving object detection.
万方数据
