Moving Object Detection Based on Background Subtraction of Block Updates

This article presents a block-updating moving object detection method-background subtraction. This method weights the former two frames of start frame, models initial background, then subtracts current frames and background frame, divides difference image into multiple sub blocks, computes sum and does threshold comparison of total pixels of each block, gets binary image through threshold segmentation, and isolates background area and foreground area. Background area is used to model and update background model. According to the test, this method is insensitive to outside environment and can detect moving object correctly.