Research On The Algorithms And Software Of Video Surveillance System Of Human Detection

With the rapid development of economy, people's standard of living has improved, and the consciousness of security is increasing. Recently, video surveillance system has been developing quickly. Video surveillance system is the most favorite because of its prominent characteristics such as intuitive displaying, real-time performance and reliability.Author is engaged in designing the video surveillance system of human detection based on DSP. It involves man-machine communication module, human detection module, JPEG image compress module and file management module.Based on analyzing the characteristics of image sequences in the real-time surveillance system, a fast and efficient algorithm of moving objects detection is accomplished. Based on studying and analyzing typical objects describing methods, a reliable human character describing system is established. After analyzing intensively the characters of moving human in the specific scene, the equations of character parameters are concluded and the mathematic model of moving human detection is set up. Reliable and efficient algorithm of important area fortification and moving human tracking is designed. Using the JPEG format, the image data is compressed. Based on analyzing the characteristics of NAND Flash memory, an efficient file management system is accomplished. Using the key buttons and monitor, a convenient man-machine communication module is accomplished. All the modules are programmed in C language and assembly language in the CCS compiler.Tested in various circumstances, the results of the experiments show that the video surveillance system of human detection based on DSP can recognize the inbreaking man exactly and quickly, and give an alarm at an abnormal instance. It has the strong anti-jamming ability to the interference caused by non-human factors. All the modules function well. The experiments show that the system is efficient and reliable.