Movement target detection method based on block-dividing image

Abstract

The invention discloses a detection method for moving targets in images based on block division. At first, the size of each pixel sub-block when images to be detected are divided into a plurality of pixel sub-blocks is preset, then an area image of a present frame to be detected is divided into corresponding pixel sub-blocks, then a frame is selected as a reference frame from frame images before the present frame or pre-established background images, the absolute difference and sum of pixels of corresponding pixel sub-blocks of area images of the reference frame and the present frame are respectively calculated according to the positions of the corresponding pixel sub-blocks of the area image of the present frame, then the absolute difference and sum of the pixels is compared with a preset moving threshold to find pixel sub-blocks which possibly have moving targets, finally the pixel sub-blocks which possibly have moving targets are statistized and analyzed to judge whether moving targets exist in the area image of the present frame to be detected, thus the video coding process and the motion detection process can be organically combined to effectively decrease the total calculation quantity of video coding and motion detection.

Description

technical field [0001] The invention relates to the field of digital video monitoring, in particular to a method for detecting moving objects in an image based on block division. Background technique [0002] Motion detection is the key technology of the intelligent video surveillance system. Its main purpose is to detect whether there is movement of objects from the series of images captured by the surveillance camera, so as to give an alarm when there is movement. [0003] There are mainly three types of motion detection technologies commonly used at present: [0004] 1. Optical flow method: [0005] Optical flow is a motion-informed description of image brightness. The optical flow method is based on the premise that the adjacent brightness in the image is similar, that is, the brightness in the image changes smoothly. Under this premise, the optical flow method maps the velocity information of the image surface brightness points in the three-dimensional space to an ap...

AI Technical Summary

Problems solved by technology

Although the motion detection method of the inter-frame difference method is simpler than other methods, it still consumes a lot of resources, especially it needs to check whether the pixel value difference of each corresponding pixel point of two adjacent images is less than a certain threshold. Judgment statistics are not conducive to the optimization implementation on DSP

Images