Research On Hand Detection And Virtual-Real Synthesis Technology In The Remote Game System

With the rapid development of computer and image processing technology, augmented reality has become a frontier research topic. Augmented reality can merge the images collected by a camera with the virtual objects generated by computer accurately, so as to provide users with a more realistic scene. Therefore, the research of a remote game system based on augmented reality has a potential research value. As Chinese chess is a common two player game, this paper mainly studies the remote chess game system based on augmented reality technology.After a camera shooting the scenes of playing chess, Firstly, it transmit the rival’s scene data to the local computer in real time after obtained and compressed encoding. Then obtain your owner chess’s location and moving information. Finally superpose the designed virtual pieces to the image transmitted from the rival according to the location information. Eventually, we realize that two people are able to share a chessboard while playing chess. At the same time, they can also see each other in the scene while enhancing the entertainment.This topic is mainly to study some key technologies of the remote game system, such as the identification of chessboard area in a complex scene, space point position, chess movement position based hand detection, the real-time integration between virtual chess and chessboard screen of the opponent, and etc.In this paper, the main research results are as follows:1. The traditional identification method of chessboard is that shoot and obtain the chess images in the boars area, so it is quite simple to identify the board area in the scene. However, the remote chess game system based on the augmented reality not just acquire images in board area, but the larger scenes including the board area. According to the need of long-distance chess game system, we need to identify and localize the chessboard correctly in a complex scene. We proposes a chessboard detection in a complex scene based on the technology of clustering analysis and template matching, and provide the support for some work in the late.2. In order to fuse our designed3d chess into the scene, the grid points on the board need to be located effectively. This paper proposes a position technology based on the straight line detection. For the purpose of eliminating the detected redundant straight, we put forward a method based on the distance parameters to obtain an automatic parameters to classify and integrate the redundant straight, and eventually get a10*9linear model. We obtain the checkerboard grid point location information according to compute the intersection of line and finally we achieve in fusing our designed3d virtual chess into a real-world scene. We fuse our designed3d virtual chess into a real-world scene successfully.3. In the process of chess game, we need to determine which chess we moved and its locations before and after being moved, and then control the movement of virtual chessboard. During the process of the chess movement, it is the hand which is the main moving object in the area of chessboard. According to the above problem, this paper proposes a method that distinguishes the images before and after the movement according to detect whether there is a hand area appearing in the board region. Finally we identify the moving location of chess by processing the chess images before and after movement and achieve in controlling the movement of the virtual chess.