Motion Tracking System Based On OpenCV

The Technology of the motion tracking which is a hotspot in computer vision,is combined with artificial intelligence,Morphology and picture processing,and is widely used in many fields.How to get more clearly image and obtain the target objects accurately has become a research hotspot.Based on open source computer vision on Mac OS X,this paper uses Python to design integrated framework of the system.This paper introduces critical funcitions and filtering mode like median filtering algorithm and Gaussian filtering to get more clear goals.In recent years,through comparatively analysis of common algorithms,motion tracking system could be created by target detection and corner extraction So as to solve the problem of large computation and low efficiency.Primary algorithm motion tracking is background cut method,inter-frame difference method and optical flow,and Harris corner detection has become a major method of image feature extraction,This system will transform the image into grayscale for more convenient feature extraction,then determine the interest area to grab the target and analyse the corner point,and finally confirm new corner point by polynomial expansion to locate the position of the target in next frame.The new algorithm makes the system development in robustness and adaptability.