Font Size: a A A

Research On The Moving Target Detection And Tracking System

Posted on:2015-02-26Degree:MasterType:Thesis
Country:ChinaCandidate:S M GongFull Text:PDF
GTID:2298330422991968Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
The moving target detection and tracking system identifies and locates themoving target in sequence of images, sends the pilot data to control the movementof the cradle head which can changes the visual axis of the camera to point at thetarget. Moving target detection and tracking technology is one of the most activeembranchment in the field of computer vision and image processing. It is commonlyused in military and daily life, such as missile guidance, automatic surveillance andhuman-computer interaction. To research the moving target detection and trackingsystem, we have to solve the two problems:1) How to quickly and accurately locatethe moving targets from image sequences;2) How to accurately control the cradlehead tracking moving targets. So the researches have done:Firstly, the commonly used methods of moving target detection and trackinghave been studied, such as background ditterencing and coterminous framesdifferencing. Consider to the characteristics of the system, optical flow method isused. As the common optical flow methods have heavy computation burden, itcannot meet the real-time requirements. Also it cannot track the moving objects withhigh speed. So the hierarchical pyramid LK optical flow method is used, it can avoidthe problems. The experiments are done to test the method by the software madewith VC++6.0.Secondly, the image-based visual servo control structure is used. To avoid theproblem that the control signal lags behind the image acquisition time, the Kalmanfilter which is based on the moving object model is used to predict the futureposition of the cradle head. Due to the uncertainty of the cradle head trackingsystem model, the fuzzy controller is researched to adjust the PID parameters andfeedforward parameter. The simulation experiments are done, and compared withthe traditional PID and feedforward method.Finally, in order to build the moving target detection and tracking system, thesoftware and the hardware of the experimental platform are designed. The finalmoving target detection and tracking test is done. But the used camera has a lowspeed, so the test to track the light dot with high speed camera is done.
Keywords/Search Tags:moving target, detection and tracking, optical flow method, Kalmanfilter, fuzzy PID control
PDF Full Text Request
Related items