Algorithm of moving object detection with illumination robustness based on confidence

The traditional methods of moving object detection in the illumination changes have no good robustness and accuracy. The background model based on texture feature may lead to cavity and the background model based on color feature is sensitive to illumination and shadow. To solve this problem, this paper proposed a method of moving object detection by fusing color feature and XCS-LBP (eXtended Center Symmetric-Local Binary Pattern) based on confidence. The difference degree of color and the texture between current pixel and its model pixel is calculated respectively, then the overall difference with confidences is calculated to achieve the better results. Our model updating method is similar to that of the ViBe model updating method. Experimental results show that the proposed algorithm can eliminate shadow effectively and have good robustness to the change of illumination, it performs the better foreground segmentation in complex dynamic background.