Research On Moving Targets Detecting And Tracking Technology In Video

Technology of moving object detection and tracking is two key technologies to realize theintelligent video monitoring.Detection and tracking of moving objects in the low-levelmodule of the intelligent video surveillance, is foundation of a variety of follow-up processingtechnologies, such as behavior recognition, behavior classification, behavior understandingand so on, therefore has the important theory significance and the practical value.In order to eliminate effects effectively of the complex dynamic background for movingtarget detection,a background modeling algorithm based on codebook model was researched，and on this basis,based on YUV color space with stronger ability of front and back areadivided,puted forward a kind of new codebook model,new codebook model used a sixelements codeword,new word learning and updating strategy,implemented faster training anda lower storage than previous codeword with nine elements or eight elements.The simulationresults showed that, the new codebook model could effectively reduce the false targetdetection, even in the context of disturbance and light conditions changing, the algorithmbased on new codebook model could be more effective in detecting moving targets.Aiming at the difficulty brought by shelter and background changing for target tracking,focused on the Kalman filter in target tracking algorithm and the tracking target based onmean-shift algorithm foundation,proposed one kind of the target tracking fusion algorithmbased on the YUV color space characteristics.Fusion algorithm combined with the advantagesof the Kalman filter and the mean-shift algorithm,at the same time,combined with a thresholdfunction,accurately calculated the target’s final position,solved the goals lost problem causedby shielding and background changing,improved the accuracy,real-time and robustness of thetarget tracking algorithm.The simulation results validated the effectiveness and accuracy ofthe new method.Developed two softwares on the basis of theoretical reaearch—remote network intelligentvideo monitoring system based on web and short range intelligent video surveillance system based on Qtopia2.2.0,used for practical application scenes testing of moving targets detectionand tracking arithmetics.The two softwares completed the testing work of moving targetdetection and tracking algorithms in the practical application scenes,accumulated experienceand laid a foundation for later real application of the algorithms.