Foreground partition method for overcoming sudden lighting change

Abstract

The invention discloses a foreground partition method for overcoming sudden lighting change. The foreground partition method includes initializing a processed video frame grey map by a mixed Gaussianbackground modeling model; performing mixed Gaussian background modeling according to the quantity of sudden light intensity change to acquire an updated background model; extracting a foreground target according to link domain marking, feature extraction and behavior judgment results. The foreground partition method has the advantages that the quantity of the sudden light intensity change is introduced into the model to process images on the premise of a conventional mixed Gaussian background modeling method, the background model is updated in real time, the problem that target partition is sensitive to sudden change of lighting and environments is solved, the problem that a target stable for a long time is updated into a background and disappears in a conventional mixed Gaussian background model is also solved, and accuracy of foreground partition is further improved. The foreground partition method can be applied to the image processing field.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a foreground segmentation method for overcoming sudden changes in illumination. Background technique [0002] Images and videos are important carriers for people to obtain information through vision. Especially with the rapid development and popularization of network and communication technology, we are entering the information age, and video processing based on image processing plays an increasingly important role in information acquisition and analysis. But pictures and video information are the most difficult to capture, process and display. In order to enable computers to obtain complex and changeable image information through vision like humans, and to liberate humans from visual labor, computer vision came into being. Computer vision is to let the computer reproduce the human visual function, so that the static image or video sequence obtained by the sensor and imaging dev...

AI Technical Summary

Problems solved by technology

[0004] There are many methods for foreground segmentation. Commonly used algorithms are: (1) Inter-frame difference method, which compares video frames at fixed intervals, which is suitable for dynamically changing environments, but due to large-area holes, the integrity of the extracted target (2), the background difference method, through the difference operation between the current video frame and the background frame to realize the detection of the moving target, this method can extract the target better and completely, but the defect is that it is greatly affected by the change of illumination and background ; (3), the optical flow method, the calculation is complicated, it is difficult to meet the real-time performance of motion detection

Images