Electronic Whiteboard Positioning Technology Based On Computer Vision Research

Electronic whiteboard is an advanced teaching or meeting tool.With computers,projectors and other tools,it can freely mark, interactive response, flexible operation,remote conferencing. China has strongly advocated to speed up the process ofeducational information.But in the market competitiveness of domestic enterprisesand capabilities of independent research,our companies lag behind foreign companies.More importantly, real-time,accuracy and robustness of the positioning technology arelack.This paper designs a new type of interactive whiteboard, based on availableresearchs.And this paper proposes a algorithm that improves real-time,accuracy androbustness of the system.The main contributions are as follows:(1) Architecture design for the whiteboard positioning systemIn this paper,the CCD camera device captures the trajectory of the electronic pen.By video image processing,computer vision and positioning algorithm,the system cancapture the location of the light point in the screen.In the spcific software support,anew interactive whiteboard is established. This whiteboard can change the arbitrarylarge screen into the large touch screen with functions of the electronic whiteboardand touching.And it can work without the additional screen and cost low.(2) Design the correction algorithmIn this paper,based the whiteboard of monocular vision system, the internalreference point correction of the camera and the angle correction method achievecamera correction. The experiments show that the algorithm is fast and error issmalland is suitable for practical application and promotion.(3) Study on the light point detectionThis paper stydies the optical flow method and inter-frame difference,background difference.As well as this paper uses the improved background differencemethod to detect the light point.In practice,we use ellipticalfitting method forprocessing the motion blur of light point.Finally,the experiments show the results.(4) Algorithm of real-time tracking and prediction the track of moving point This paper proposes a hypercomplex correlation matching and Kalman filteralgorithm that can real-time track and predict the track of moving point.The algorithmuses hypercomplex for color images registration,and the registration results are usedto correct the filter parameters.The experiments show that the new algorithm canidentify directly the color target and high accuracy.