Automatic and improved camera information processing for a 3D photo- modelling system

In this study, we propose a process of photo capturing, improved camera calibration, and improved camera alignment to generate a three-dimensional (3D) textured model of an object from its multiple two-dimensional (2D) images for presentation on a website. The proposed photo capturing can represent the object and mat photos in a controlled environment and can easily remove the background color. The mat photos are used in camera calibration to evaluate the camera parameters. The object photos and camera parameters are then used in model reconstruction to generate a 3D mesh model. Camera alignment is employed to obtain the bottom information of the object by flipping the object to take another layer of photos. The accuracy and efficiency of camera calibration and alignment are both improved for automatic generation. Finally, we describe an integrated software system for modelling to demonstrate the entire process, which can yield a highquality 3D textured model for object presentation. Six realistic examples were presented to demonstrate the feasibility of the proposed system.