Computer virtual reconstruction of a three dimensional scene in integral imaging

Abstract

To solve the problem of rapid perception of the collected content and non-contact measurement in integral imaging, computer virtual reconstruction algorithm of a three dimensional scene is proposed. Firstly, calculate the combined disparity map of an elemental image array using the region-based iterative matching algorithm according to the distribution characteristics of the homologous pixels in the elemental image array; then calculate the spatial coordinates of the reconstructed object points in line with the triangulation principle; and at last, delete error points and reduce the data redundancy using the data simplification algorithm based on the scanning line, and then the reconstructed three dimensional scene is obtained. The experimental results indicate that the method can not only reconstruct a clear and complete three dimensional scene, restore the relative position of the objects, but also measure the objects' sizes in the three dimensional scene.