Research Of Interactive Technology In Wall Projection Interactive Based On Infrared

With the development of computer and multimedia technology, people need more and more requirement and demand for human-computer interaction. People no longer satisfied to interact with the computer by the peripheral, but the need for a more natural way of interaction. In order to research and design a simple structure, easy realization, low cost large screen interactive system,which can be used in games, entertainment,advertising and other interactive field. This paper uses the interactive equipment infrared camera,projector, screen and other advanced equipment to realize a wall projection interactive system based on infrared. This paper focuses on the key technology of moving human target detection, target contour extraction and human motion recognition, human-computer interaction. The main results in this paper are as follows:(1) Aiming at the problem of moving human target detection in the scene, this paper puts research and forward the method of background subtraction based on improved mixed Gauss background model. Firstly, make median filter to the video image to reduce the influence of noise on the detection results;then establish the mixed Gauss background model, by improving the mixed Gauss background model updating rate to make the former N frame update rate more quickly than the later update rate, thus we can establish the mixed Gauss background model more quickly; after this we can detect the motion region by making the subtraction between video image and background model; Finally, we need to make the corrosion expansion process on the motion region to improve the accuracy of detect results. The experimental results show that, the method is feasible and effective.(2) Aiming at the moving human contour extraction problem, this paper puts research and forward the method of moving human contour extraction based on improved Sobel edge detection operator. Firstly, make the background point contacted with the moving human target into human body by the morphological dilation algorithm, So that the moving human target boundary expandes to the outward, and connect the human target contour edge; then analyze the disadvantage of the Sobel edge detection operator of detecting only twohorizontal and vertical edges, we put forward the improved method by increasing six direction template;finally realize the effective detection and extraction of moving human target contour. The experimental results show that, the method is feasible and effective. (3) For human motion recognition problem, this paper puts research and forward the human activity recognition method based on Hidden Markov Model (HMM). Firstly, we research the basic theory of HMM, including its mathematical background, definition of the model and three problems need to solve by HMM; then we establish the HMM of human behavior; after this we estimate and train the parameters in the model;finally, we classify the target human behavior. The experimental results show that, the method is feasible and effective.(4) We design and implement a projection interaction system based on computer, infrared camera, projection projector and other advanced equipment. In order to achieve natural and friendly interaction with the scene, and thus we can improve the user experience, we design and implementation the wall projection interactive system based on infrared through the moving human target detection, moving human contour extraction, human motion recognition method. The system is divided into five important module of visual simulation, moving human target detection, behavior recognition, interaction, audio effects etc. Finally we complete the wall projection interactive system based on infrared. After implementation,the system has already chieved good results in practical operation.