| With the rapid development of the multimedia technology, computer technology and large screen display technology, and low cost of hardware, Multimedia teaching and lectures have been widely used, major colleges and schools have more and more multi-media classrooms. However, through the keyboard and mouse operation, the speaker should be bound in front of the computer during multi-media teaching has became more and more inconvenience. As a result, many researchers began to study a new method of interpersonal interaction in order to adapt the requirements of current multi-media teaching and lectures reporting. In this demand, laser pointer technology has began to appear and been widely used. But laser pointer can only instruct the content and turn page to front or back in a lecture, it's interaction is not strong,and operation is inconvenient, at the same time it also can not carry out priority mark, and so on. Followed the development of high-end electronics, interactive white board has became a major research direction of human-computer interaction, It turns a big screen into a large touch screen,so it can be achieved any writing, mark on the big screen using by a specific laser pointer.The rare operation functions of laser pointer, interactive white board needs high-quality hardware, and it's technical is complicated, at the same time it need a specially pen。So in this paper we discusses a new method of human-computer interaction which using the image recognition technology. In recent years, Image recognition technology has became a major direction of the computer vision technology. In this paper we adopt a generic USB camera to capture real-time sequence image information of the projection screen, in order to get the laser pointer's position in the projection screen. We use image recognition technologies to process these images. Then user can remote the mouse operation using by a laser pointer.This will achieve human-computer interaction.First of all, this article introduces the basics of digital image processing and image pre-processing operation. Secondly, analyzing the video capture technology and moving target detection method.In this paper, we will using DirectShow technology. USB camera can capture 30 frames per second. It can better meet the system requirements. At the same time, we analyze image tracking methods in detail which have been used in this article, such as inter-frame different method, the background different method. Then we discuss the identification and location methods of laser pointer, and put forward a new positioning algorithm to improve the positioning accuracy of laser pointer. We also pointed out that the sequences image which captured by the USB camera will come forth geometric distortion, and calibrate these image in order to further improve the accuracy of the laser pointer's positions in the system. Finally, we describe the structure, working principle of the large screen human-machine interaction system which based on the image recognition technology, we also expect the improvements of this system and the next work we will do. |
