| In recent years, video surveillance in schools, airports, shopping, malls, financial, and other areas of residential quarters are widely used. However there are still some deficiencies for such system. Such as school violence video surveillance is only limited to passive monitoring system to provide video images. The thesis focuses on the application of video surveillance of the human face location algorithm which can effectively improve the system efficiency and intelligence.The purpose of face detection and localization is to automatically detect the video, locate, track and identify the human faces which are the hot research areas of pattern recognition and computer graphic processing. The technology can be widely used in image retrieval, video coding, biometric authentication, computer security, electronic commerce. Although researches on such areas have been taken on many years, as the face detection and positioning have its own complexity, many researchers are still painstaking to the solution. This thesis improved the YCbCr color space (i.e. YUV) fast location algorithm, the AdaBoost face detection algorithm and the Mean Shift algorithm, and implemented these algorithms by C++ programs. We have achieved face detection algorithm software, from static images to video real-time image, with higher accuracy, and processing speed. We also achieved an automatic face tracking software for video acquisition, image acquisition and real-time face tracking.This thesis fully investigated the face detection and localization algorithm theory and method. based on the human face detection and localization algorithm,it is applied to real security in theory, designed with a passed effective detection, identification, location of video surveillance systems, simulation results show that the human face detection technology for video surveillance, can effectively and accurately detect faces in video images to capture the face of effective information Monitoring System to improve the current level. |
PDF Full Text Request