Research On Gesture Recognition And Its Application

Hand Gesture recognition technology is a technology that translates the hand gesture into its' represent meaning through the computer. Due to the diversity, ambiguity in a different time and space, gesture recognition becomes a challenging research topic. This paper mainly do a research on gesture recognition and its' application, and apply this technology to a UAV control system and a guide system of warehouse.First of all, a comparative research on gesture recognition which based on sensor technology and image processing technology is done in this paper. After giving an analysis of the cost and the convenience of the gesture recognition, the image processing technology is chosen to do the further research.Secondly, the basic principle and method of gesture recognition technology are discussed in this paper. The basic steps of gesture recognition technology can be divided into gesture area segmentation, feature analysis an gesture recognition. And some key technologies for the each step are also analyzed in this paper.Thirdly, different skin models are compared in this paper, and a new skin model named HS-CrCb model which based on the advantages and disadvantages of the HSV, YCrCb skin model are proposed in this paper, and the model achieves a good effect.Then, methods to calculate the number of fingers of a static gesture are proposed in this paper. The convex defect is used to establish a simple hand model, and the shape characteristics are used to get the corresponding number of fingers.At last, after the research on the static hand gesture, a further research on the UAV control system and warehouse guide system are done in this paper. Hidden Markov Model is used to design the UAV control system. After analyzing the technical requirements of the guide system, a new method to design guide system is proposed in this paper, which combines the Hidden Markov Model(HMM) and Support Vector Machine(SVM) together.Through a large number of experiments, the static gesture recognition and the dynamic gesture recognition technology achieves a high identification accuracy, which has a certain value on other engineering applications.