Depth Computing based on Stereo Matching with Singular Value Decomposition

Abstract

The system is aimed to develop the 3D model reconstruction from multi stereo images. In this paper, an efficient depth computing based on stereo matching algorithm is introduced based on feature extraction with morphological method and efficient matching method. To reconstruct the 3D model, the depth map is created by the algorithm with Gaussian correlation based on the singular value decomposition (SVD) approach is used to project the image point. An efficient surface reconstruction algorithm is developed for generating the 3D model of an object. The objective of the system is to reconstruct the 3D model of multi stereo images to reduce the time complexities and occlusion errors of the scene and object.