High-precision texture mapping on 3D free-form objects

Cultural heritage objects such as historical buildings and sculptures are being lost every year due to such causes as earthquakes and fires, so that the need of preserving such objects digitally is becoming widely recognized. Digital preservation is carried out by measuring both 3D shape and optical property of an object of interest. The shape is usually measured with a laser scanner while the optical property is measured with a digital camera. It has been found to be difficult to match textures on the object surface exactly. Matching is carried out usually via camera parameters associated with the digital camera employed. Such camera parameters are usually derived based upon object silhouettes or perimeters. This paper proposes utilization of both silhouettes and textures for a more precise match of textures on the object surface. Furthermore, it proposes an algorithm to derive an optimal set of camera positions to realize a constant overlapping ratio. The proposed methods were applied to two examples, resulting in excellent matching. (C) 2006 Wiley Periodicals, Inc.