The Research Of Moving Object Recognition In Specific Background

Nowadays the moving object recognition technology is a important research in the field of artificial intelligence and pattern recegnition,the fast and stable face detection algorithm becomes the new host research spot.Image recognition and tracking of moving object is a process of detecting-recognition-tracking, which is based on dynamic image analysis and combine methods of image recognition and image tracking.This article chooses the moving object as the target.It puts forward a method which carry out image recognition and tracking on moving object in the simple background and makes comparison and study on algorithm of object recognition and tracking.It makes progress on algorithm of image preprocessing, image segmentation, feature extracting, image recognition, tracking and predicting, which is based on analysis of system error,and A new fast face detection algorithm which based on skin color model and distance transform is proposed.First,skin pixels are distinguished from the non-skin pixels by using the clutter feature of skin color in the YCrCb color space and the essence position of face area in whole picture can be found.Then figure out the length and width using distance transform and the geometrical feature of human face,so the accurate face position can be located.It aslo make a progress on image tracking process by improving traditional mean shift arithmetic.The Mean Shift algorthm for segmentation is mainly used in static images, and has the characteristic of fast and effective iteration. In this thesis, the method of segmentation based on the Mean Shift combined with the adaptive thresholding segmentation algorithm is applied to segment moving sequence images,and a moving object is selected in order to tracking automatically. This paper based on OpenCV and CAMShift algorithm to realize the self-motion tracking of colorful sequential lmaging.