Enhance Intelligence Video Surveillance with depth and color information

Abstract

Recently, Intelligent Video Surveillance (IVS) mainly use color image to detect people who are entering, leaving and loitering at a fixed area. When detected event is happening, Intelligent Video Surveillance can turn passive into active to notify security guard immediately, and it makes security guard reduce response time. Actually, the color image may be influenced by light and shadow change condition during background removal and foreground extraction. Foreground object is not easily divided up, and it makes detecting object incomplete and may be misjudged. This research uses a color camera to combine with a depth camera to detect the events. Depth information can resolve the light and shadow change condition problem and complicated color background. It divides up the foreground object easier and more completed, and the foreground object is used to detect the events. Experiment result shows that using depth information to divide up the foreground object for detection can enhance effectively accuracy. Although using depth information to detect can resolve the light and shadow change problem and complicated color background problem, depth information has the distance limit of detection. But color information does not have the distance limit. In view of this, I integrate the advantages of color and depth to implement the Intelligent Video Surveillance. The result shows this is better than only using color or only using depth. The system is very suitable for use in the indoor environment.