We examine the use of 3D scanned objects in computer graphics applications. We consider the requirements for the types and resolution of data required. We also identify some outstanding research issues in this area.

Automatic 3D Registration using Range and Colour Data refers to the process of automatically joining a pair of overlapping colour 3D images acquired using a laser range finder. Currently there are no methods that can automatically and reliably register overlapping 3D images. This Thesis presents a solution to the problem of automatic registration, which uses both the 3D data, along with co-registered 2D colour information. Features are found in the 2D data, but both 2D and 3D descriptors are attached to these feature points. The correspondences between these feature points in the two overlapping images are determined, and the transformation between the overlapping range images is computed from the best correspondences. The combination of 2D and 3D data exploits both the object geometry, as well as the object colour. Both the feature finding and correspondence process are automatic, and do not require any user intervention. A number of experiments are conducted to demonstrate the effectiveness of the approach.