Design And Implementation Of Vision Tracking Platform Based On OpenCV

This paper consists of lots of useful method to deal with different kinds of object-tracking problem,implemented by the Open CV,which developed by Intel Corporation..Actually,object-tracking refers to computer vision,pattern recognition,image processing.according to diverse classification method,object-tracking belongs to different categories, Tracking algorithm used to solve this problem.particle fiter is algorithm of predicting motion status.combined with color-feature used to represent object to track object.particle filter method should establish prediction model to estimate motion status to solve object-tracking problem.this paper raise a new version of particle filter based on color-feature.To use motion model to represent motion status.Mean Shift use core fuction to analysis image feature.tracking result focus on feature-most area.in this area there is local maximum value.when used mean-shift algorithm to track object.feature like color-feature,luminance intensity needed to be found. Mean Shift need to calculate constantly to find the local maximum value,and this local maximum just represent the tracking result.CAMShift,based on color-feature,usually convert the RGB color space to HSV color space.Then transform feature to color diagram.Next,to figure out possiblity of every segment of diagram.Then calculate every pixel’s possiblity of Channel H,and get possibility projection.then estimate object’s location of next frame,according to the pre-frame.Open CV is a platform for developers in the filed of computer vision,which developed by Intel,and used widely between researchers.there is supporting function,and microsoft platform to accomplish this vision platform.Lastly,used above algorithm to show tracking result.and summarize them.