Human Movement Tracking And Head Positioning System Design And Implementation,

With the development of modern society and people's increasing sense of security, human identity recognition is indispensable in more and more public places such as railway stations, airports, banks, administrations, communities and so on. Biometrics recognition, a novel kind of identity recognition technologies, can identify human identity according to the physiological or behavioral characters and is better than traditional recognition methods in accuracy and speed. Biometric features used currently include fingerprint, face and iris, etc.In this dissertation, a subject on face recognition of national 863 Project that a new method and technique on the basis of Visual Perception and Cognition mechanism Multimode Character Features authentication will be discussed. As previous work of this subject, human motion tracking and head locating based on video will directly affect the performance of face detection.The topics are mainly discussed in this dissertation as follows: detection of moving target, tracking of moving objects and head locating.Object detection which separates the moving object from the sequent images is the foundation of subsequent treatments. Basing on the analysis of the three commonly algorithm of movement target extraction (background difference,inter-frame difference and optical flow method) at this stage, and taking the application occasion into account , background difference method is used. In the object region obtained by the method discussed above, some gaps still exist, and noises are also found in the background. With the combining of mathematical morphology and connectivity detection (scale filter), we can reduce the noise and smooth out gaps.During the phase of moving target classification, the dispersedness and area information of the moving target is used to recognize human in this dissertation.On the research of object tracking, by studying several related algorithm, this dissertation improves the extend Kalman filter to tracking targets. It compares centroid position and area to replace complicated template matching, thus this method improves the tracking speed, as well as handles the cluttered scene simply and effectively. Human head is located by a "convex" shape between head and torso. Avoiding the interference of external factors such as the skin color, neck, contour, etc., this algorithm achieves high rate of success and is much more accurate.Last, in this dissertation, a complete moving human tracking and head location system is designed and implemented. The experiment results show that the algorithms of complicated human motion tracking and head locating propesed in the dissertation are useful. It established a solid foundation for the future research.