Research On Moving Target Detection In Video Sequences

Based on the video image of moving target detection in computer vision isa research hotspot in the field of. Moving target detection is the basis of theintelligent monitoring, image motion analysis, visual man-machine interactionand many other computer vision problems, by moving target detection can beobtained in moving image motion information, extracted from the video imageof the moving object or target, for subsequent video image analysis and targetmotion understanding played a key role. The moving target detection has beenwidely used in industrial detection, traffic navigation, image processing,intelligent systems and many other fields; therefore, moving target detectionalgorithm research has important practical significance.This paper briefly introduces the moving target detection technique basedon content, Including image segmentation, mathematical morphology processingand image segmentation methods, the traditional target detection algorithm aredescribed, the optical flow method, frame difference method and the backgrounddifference method, And the advantages and disadvantages and applicationenvironment of the three methods are discussed in comparison，and focusing onthe background difference method is discussed, for the static backgroundsituations of moving target detection algorithm research and improvement.Based on the original traditional background difference method of moving targetdetection algorithm, joined the idea of genetic algorithm, the original method isimproved, This paper proposes a new algorithm of moving target detection, themoving target detection is more accurate and real-time, the most importantcharacteristic is compared with the traditional method of memory required isreduced a lot, count the number of frames, update speed, suitable for real-timemonitoring. Also the method of video image sequence to do the experiment, thispaper introduces the method of the reliability and validity of the method, withthe other two methods were compared, confirmed in the real time and memoryfootprint, this method has a better effect. For moving targets detection in the presence of shadow problem, put forward a kind to use color feature and shapefeature of combining shadow removal algorithm, in the shadow of types basedon the discriminant, the finding of multiple point, use mathematical method thatdividing line, shadow and object segmentation technology to remove theshadows.