Intelligent Visual Surveillance System Based on DSP
Shu-man Zheng, Qing Guan, Sheng Xu, Ying Li · 2009
This paper introduces an intelligent visual surveillance system based on TI multimedia digital signal processor (DSP) TMS320DM642, including hardware platform, software platform and algorithm realization. As the core of this system, DSP communicates with host PC via PCI bus. User can set rules via GUI software. The video input from either analog camera or IP camera can be processed by DM642 in real time. If any event breaks the rules, alarm information will be sent to users. Finally, a shadow suppression algorithm based on HSV color information is introduced to reduce the impact of shadow on motion target detection in the sunshine background. Experimental results show that this system can work well outdoors and the CPU load is low enough to meet the requirements of future work such as recognition and tracking.