An Extended Mean Shift Tracking Algorithm With Scale Adaptive

As the increasingly development of artificial intelligence, digital image processing, patternrecognition and other areas, intelligent video monitoring is increasingly becoming a wideapplication of comprehensive discipline. And in this field,the target tracking technology is a kind ofenergy from the image signal real-time identification of target, extract target location information,automatic tracking target motion technology.So it is an important research significance of thesubject.At present there are many methods in the target tracking technology, based on the previousresearchs,I put forward an adaptive Mean shift tracking algorithm..In this paper, it’s introduces the probability density histogram which based on the R, G, B colorand texture of the object, and bring in a new type of texture feature isLBP8,1texture feature. Aimto target model description,combine the texture and color features together, only to join the texturefeature and the color features simply,the computational complexity of this method is very high, So itis not suitable for the need of daily application. Therefore, this use of the five kinds of texturefeature model to represent objects texture. The experimental results show that this method cansignificantly more reflect the characteristics of target, and algorithm complexity doesn’t improve toomuch.Secondly based on the traditional Mean Shift tracking algorithm is susceptible to be sheltered,background information, etc. This paper bring in background weighted factor, make an adaptiveMean shift tracking algorithm.and acquire a good result. Through the test shows that the method fortarget tracking has a very good effect.In order to improve the adaptive tracking algorithm, this paper puts forward through thesimilarity comparison update bandwidth of the new method. Calculate the similarity function of thecentral pixels and the edge pixels in current frame and the central pixels in previous frame12,12.by comparing the function of both judge the target scale changes, to realize scale update.Experiment results show that the scale update method in the process of tracking is efficient.