Dsp Network Video Monitoring System Research Based On Human Face Detection And Position

As the fast development of IT industry, the development of computer technology, embedded technology and the embedded software technology have become hot issues in this society; The new generation intelligent video monitoring control has received more and more popularity. The traditional video monitoring is exercised by human being to find safety hazards or abnormality, therefore such monitoring has its fixed limitations----it is difficult to realize safety monitoring and test management. So the traditional video monitoring system is being eliminated. While the new generation intelligent video analysis monitoring control system based on DSP is causing more and more attention. Such video monitoring system could actively collect, analyze data and exercise alarming, recording and analyzing functions according to the pre-assignments by distinguishing appearance and gestures and other features of monitoring objects so it plays more and more important role in the market. The object of this thesis is to provide technology support to realize new generation intelligent video monitoring system with the DSP processor as the core hardware platform. On this basis, the human face detection arithmetic, human face position and tracking arithmetic in the CCS integration development environment could offer technology foundation for new generation intelligent video monitor.This thesis firstly introduces the basic theory on human face detection, position and tracking arithmetic, designing a hardware overall structure based on DSP real-time video monitor; integrating the real-time features of demanding analysis and video monitor control, designing such important hardware module interface as the video capture interface, video input interface, video output interface and network transmission interface and so on. Furthermore, in software design, this thesis designs software overall framework, on the overall framework, combining with video capture, H263 coding arithmetic it also realizes such core arithmetic as human face detection arithmetic, human face position and tracking arithmetic. The above arithmetic are optimized so that the optimized arithmetic and the tracking arithmetic can run smooth on DSP platform, at last, we should apply TCP/IP NDK derived from TI C6000 series to effective allocation of DSP/BIOS so that the DSP/BIOS which is not the true real-time control system could exercise network transmission, realizing the extract position and tracking of human face detection, thus free from obstacles network real-time monitor control is achieved. The result is perfect and the requirement of this subject is met. To sum up, the final test result shows that: all parts of video monitoring system designed by this plan run smoothly; system is stable and the hardware is completely meet the design requirements, therefore, it lays sounding foundation for the design of new generation intelligent video monitoring system.