Detection And Tracking Of Video Moving Target Based On LABVIEW And MATLAB Mixed Programming Techniques

With the rapid development of intelligent video surveillance, video target detection and tracking is becoming an important research topic in the field of computer vision. In the past, scholars payed more attention to theoretical research and engineering practice of specific moving object and static background leading to mature theory and applications, relatively did few studies of non-specific moving target and dynamic background. Scholars gradually distract attention to the latter, in order to solve the emerging theoretical and applied problems. This paper carries out the moving target detection and automatic tracking technology research in the dynamic context, the main works are as follows:In the respect of moving target detection in the dynamic context, mainly do two jobs: First about image inter-frame registration, this paper uses eight neighborhood Harris feature point extraction algorithm and the center-biased rectangular template searching and matching template and uses bilinear interpolation to complete the image of the reference frame matching with the current frame image, to realize background static; Second through summary comparison of several classic moving target detection algorithms, put forward the symmetric differential detection algorithm to meet the system requirements, through binarization, morphological filtering and target area demarcation, extract a moving target more accurately.In automatic tracking of moving targets respect, compare several common target tracking algorithms synthetically, and take into account the tracking lag problem. A Kalman motion estimation for this article and CamShift target tracking hybrid algorithm based on MeanShift thoughts have been used. The algorithm based on the target color characteristics, is with no parameters, good robustness and real-time. By PTZ it can control moving target within the vicinity of the monitoring center of the scene, and can also keep automatic track of the moving target accurately and timely combining with the symmetric differential treatment results to initialize CamShift algorithm search window.Finally, design a single moving target video surveillance system experimental platform. The system’s hardware takes PC as the based controller, DH-300image acquisition card and Enoch-1032-type CMOS camera as video capture device, TRUM-60-P-type traveling wave type rotating ultrasonic motor and the UMD-3II drive as executing components of the basic framework. The system’s software uses LabVIEW and MATLAB mixed programming platform, modular design, and the powerful MATLAB image computing power and Lab VIEW’s vision module IMAQ VISON which has intuitive, friendly, feature-rich image processing interface. Verify the video object monitoring algorithm effectively and gain image data analysis.When a moving target’s rate is within a certain range, the system can detect and track it preferably to meet the design requirements.