Construction And Optimization Of Large-scale Real 3D Display

The three-dimensional display technology is one of the important development directions of the future display technology. It can provide observers the depth cue of 3D scene, which not found in 2D display, to make 3D scene more real. In 1996, M. Levoy and P. Hanrahan proposed light field rendering technique for 3D scene reconstruction, which use the light field distribution of a real object to regenerate spatial 3D display. It results in vigorous development of 3D display devices based on light field reconstruction. The advantages include large viewing field, naked eye observation and viewable for a number of people at the same time, making it an advanced 3D display technology.In this paper, computer graphics and light field reconstruction principle are combined expected to construct a large-scale real 3D display system. And then make a detailed analysis of design requirements of display zone in cylindrical space. The mapping relationship is figured out based on system structure. In the final, the experiment results are shown in this paper.With the promising development of three-dimensional (3D) display techniques, imaging performance improvement will be the focus of the research in the future. Here we discuss a series of optimization methods on the basis of proposed system, especially on the problem of luminance uniformity and calibration. The luminance uniformity is mainly caused by the alignment of projectors. Solution to this problem is adding one projector every five projectors to light the dark zone, and then employ the adjustment method proposed by former researcher. In this way, the luminance uniformity of system can be effectively increased.Secondly, it is rather difficult and inefficient to calibrate each projector manual-separately in multi-projectors system. Combined with the traditional 2D image calibration, we proposed an innovative automatic geometrical calibration method for multi-projector-type light field 3D display. The calibration framework based on image mapping is detailed, which transfers the calibration of 3D scene into the calibration of 2D image in the diffuser interface. Given that we have taken the property of rays into consideration, we believe that the calibration algorithm can be applied on most projector-type 3D displays, not only based on light field reconstruction, but multi-view reconstruction.Finally, the present work is summarized and the future work is prospected.