3D Reconstruction System with Mesh Model and Texture Mapping

We describe and demonstrate an algorithm for 3D Reconstruction that output a beautiful model with textures. Taking a set of images as input, we detect features from images by SIFT, estimate camera parameters and 3D points positions by Bundle Adjustment, and use PMVS to acquire a point cloud. PMVS algorithm can output accurate object and scene models with fine surface detail. Then, we propose a novel and cheap triangulation method suited for the point cloud generated from PMVS. Our triangulation method make full use of the advantage of BFS(Breadth-First Search), with choosing points constraints in optical condition. Current algorithms end up when the point cloud is acquired, our algorithm has a complete reconstruction work flow to acquire an attractive model with textures, which appears great better than point cloud.