Research On Multi-channel Surface Projection Technology

The multi-projection display system has a wide range of applications in the fields of virtual scene exhibition,simulation training and etc.The frame is composed of multiple single projection frames,and the projection scene is diverse,witch is not limited to the plane.Therefore,it is necessary to correct the geometric distortion of the spliced frame,also the highlight and color change during multi-channel fusion.The existing geometric correction technology can solve the geometric distortion problem,but most of them rely on the hardware system with a high cost on equipment and a low universality of the projection screen.Then how to get rid of the limitations of hardware and increase the universality of projection can be a challenging task.Based on the “Full Visual Virtual Simulation Project” of the Graduate Innovation and Entrepreneurship Fund Project and the “Multichannel Radial Projection Correction and Fusion Technology Research” of the University Science and Technology Innovation Fund,this paper do the experiment with the ship driving digital simulation platform and the Wuhan University of Technology's virtual driving laboratory projection equipment.The view distribution scheme based on the ship driving digital simulation platform divides the 3D vision into several data display channels,and based on this,the paper analyzes the related research work from the aspects of improving the universality of the geometric correction method,the dependence of the geometric correction on the hardware,and the optimization problem when considering the multi-channel fusion.Then propose the idea of realizing universality expansion by replacing the projection screen function and the method of reconstructing the surface by the spline function.Main work and academic contributions are as below:(1)Irregular surface correction based on adjustment relationship between pixels.The algorithm automatically fits the projection screen function by substituting the relationship between the corner points of the feature points and different coordinate systems,then calculates the position of the offset points according to the position offset function to complete the geometric correction.(2)Quadratic quasi-uniform B-spline surface correction based on shape parameters.The algorithm performs rough correction on the feature points of the surface by the fitting of the spline to the surface;then fine-corrects the splicing between adjacent channels according to the shape parameters;finally,the dirty rectangle algorithm is selected to optimize the desktop copy to complete the full view scene projection correction.(3)Optimized processing for multi-channel fusion including edge weakening and color correction.It mainly includes Alpha brightness adjustment based on pixel point and highlight coincidence band distance relationship,and color correction based color correction matrix.Finally,this paper uses Wuhan University of Technology's virtual driving laboratory equipment and camera to carry out real scene experiments.And then verify the feasibility of the program through the objective evaluation index of image fusion effect and the subjective effects.