Design and Implementation of Embedded Video Surveillance System Based on Human Detection

Abstract

In order to get information timely when elderly people fall down and need assistance, a kind of intelligent video surveillance system is designed and implemented. By using this system, the elderly dont need to wear any device. The H.264 video standard, the embedded technology, network technology, video surveillance technology and image processing technology are mixed used to detect human gesture in the video stream monitored. This system updates the background by the Gauss mixture model. The moving target is segmented by using the background subtraction algorithm. Experiments show that the system can be remote monitoring and detecting attitude. When the people falls down the system can automatically alarm, and satisfy the requirement of the timeliness, accuracy and stability.