Staff Identification And Tracking OpenCV Research And Implementation

With the rapid development of computer, network and communication technology, the human-computer interaction becomes an important part for people’s daily life as time goes by. The traditional human-computer interaction methods, such as microphone, tablet, mouse and keyboard gradually meet bottlenecks during the human-computer interaction. In this situation, human should meet the requirements of computer with the previous disciplines, it can’t meet people’s requirements at present, the research of new methods for the human-computer interaction becomes more and more important. Hand, as the widely used and most flexible part of our body, and sign language is said to be the second language for our daily life. So the research of recognition and tracking for people’s hands become the focus at present.The core research work in this thesis is the OpenCV based recognition and tracking for people’s hand. First, we discuss the basic concept, theories, technologies and core problems of recognition and tracking for people’s hands. Second, researching the methods for recognition and tracking for people’s hands based on the theories mentioned before. There are4parts in this thesis. First, the preprocess of ROI area for a hand, include Gaussian filtering and morphological operation, second, build a color model for skin, include the introduction and exchange in RGB, YCrCb and HSV model, third, template matching, to recognize the hand based on the background of skin color based method, fourth, hand tracking, we introduce a method called Camshift tracking algorithm, combine the recognition technology to solve the problem of semi-automatic tracking exist in the Camshift tracking algorithm. The core content in this thesis includes the hand recognition algorithm based on the combination of skin color and template matching, and the Camshift tracking algorithm. The highlight of research work in this thesis is below:(1) In order to increase the recognition rate, we introduce a skin color and template matching based hand recognition method based on background of the traditional skin color based hand recognition. The experiment results show that this new methods increase the recognition accuracy under complex situation. (2) In order to solve the problems in the semi-automatic tracking of Camshift algorithm, we come up with a new method combine the hand tracking and recognition technology, optimize the Camshift algorithm, solve the problems in the semi-automatic tracking of Camshift algorithm to achieve real-time recognition and tracking for people’s hands.