Research And Implement On Digital Image 3D Stereoscope Display Algorithm

Digital images of three-dimensional display technology is the same scene taken two or more different perspectives of digital image processing as the eyes can achieve a disparity in the images and produce three-dimensional visual effects technology. High-precision three-dimensional digital image display, not only has wide range of application needs, but also is the technology base of a dynamic video-based three-dimensional image display. Stereoscopic display of digital images including stereo matching and view synthesis these two key technologies. This article features for complex three-dimensional digital image display applications demand of high precision, implemented a more in-depth study on three-dimensional matching algorithm, the view synthesis algorithm and digital image processing software for stereoscopic display technology designed. The main work and research results are as follows:1. Aim to the problem of existing global stereo matching algorithm assuming that the depth of whole image continuous everywhere, while the stereo matching accuracy is too poor for image with depth discontinuous, this paper proposed a stereo matching algorithm based on image segmentation called SGC. The SGC algorithm builds the MRF model using the result of image segmentation; thereby the edge information of the disparity map can be kept in the continuity (smoothness) constraints. Moreover, to improve the disparity accuracy, a novel energy function is well designed to restrict the depth-continuity of an image and applied to the Graph Cut algorithm to describe the image feature. The experiments show that our SGC algorithm can reflect the depth information more exactly than the existing algorithm and achieve a high-precision disparity by avoiding the error arisen from the continuity constraints.2.studies the pixel mapping ,shadowing determine and holes filling methods for view synthesis using Depth Image-based Rendering(DIBR) techniques, aim to the problem of blurred vision comes with bilinear interpolation, designs and implementations of a hole filling algorithm based on depth judgment and function approximation. This algorithm considered in detail the true scene structure, estimate the range of holes regional pixel disparity starting from the principles of holes formation to determine the ownership of an hole area, at last estimate the hole pixel value according to the relevant pixel. Experiment results show that our algorithm can reflect the complex scene structure more exactly than bilinear interpolation algorithm, and basically eliminate the blurred vision caused by bilinear interpolation. Based on the above research, this paper developed a stereoscopic display of digital image processing software using OpenCV vision library, through the NVIDIA 3D VISION SYSTEM, on the LCD monitors with no less than refresh rate of 100Hz,many people can have a free view to the complex scene high-quality digital image three-dimensional display together.