Font Size: a A A

Moving Target Detection And Tracking Based On Video Technology Research

Posted on:2014-02-25Degree:MasterType:Thesis
Country:ChinaCandidate:L L ZhangFull Text:PDF
GTID:2248330395982955Subject:Weapon system and application engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of electronic information, the technology based on the video moving target detection and tracking is more widely applied to many areas, such as transportation, military, industrial manufacturing and public life. Most of the traditional video surveillance systems are based on the PC. They are high cost, poor reliability and difficult to meet real-time requirements of the system. Currently, a simple and effective way, which is embedded the hardware and software of the video surveillance system to a high-performance DSP chip platform, can solve this problem.In this paper, based on digital signal processing technology research, and major research-based the DM642DSP embedded video monitoring system hardware and software design and Implementation, PTZ serial communication, and moving target detection and tracking algorithm realization.According to the system requirements, the paper designed a hardware platform solution which put the TI high performance digital signal processor TMS320DM642as master control chip. This solution includes a camera, PTZ, DM642chip, serial communication, monitor, etc. as one of the monitoring system. Experimental results show that the solution is feasible. Through the real-time operating system DSP/BIOS and internal integrated development tools, the paper developed a based on DM642video acquisition and display driver model which is on the basis of the traditional DSP/BIOS drive. This paper also completed DSP software design for the video driver.According to the requirements of the system and the improvements of the traditional target detection and tracking algorithm, this paper designed background compensation and frame difference method for detecting a moving target and the target feature tracking algorithm for tracking moving targets based on Kalman prediction. The experiments show that the real-time optimization algorithm can better detect and track moving targets.
Keywords/Search Tags:TMS320DM642, Target detection and tracking, Background motioncompensation, Frame difference method, Kalman prediction
PDF Full Text Request
Related items