We propose a fast 3D model acquisition system that aligns intensity and depth images, and reconstructs a textured 3D mesh. 3D views are registered with shape alignment based on intensity gradient constraints and a global registration algorithm. This approach improves the registration when range image is acquired from noisy cameras. We reconstruct the 3D model using an efficient algorithm which takes advantage of a new data structure: linked voxel space. Finally, we present experiments to test the accuracy of our approach on real stereo images.
