An Adaptive Threshold Algorithm for Moving Object Segmentation

Connected region detection is usually used to obtain foreground regions from foreground image after moving object detection. In order to remove noise regions and retain true targets, a threshold that limits the circumference of foreground regions should be introduced. The method which uses the same threshold for all surveillance videos cannot handle scene changes. In this case, we propose an adaptive threshold algorithm for moving target segmentation. A strategy based on the combination of background modeling and Grabcut is presented to extract foreground objects and set an initial threshold. On the base of this, we can choose some foreground as samples and classify them by K-means clustering method. Finally, an appropriate threshold could be selected for moving object segmentation according to the classification result. Experimental results show that the proposed method has strong adaptability to various scenes and improves the accuracy oftarget segmentation.