An automated and accurate procedure for texture mapping from images

The possibility of deriving 3D digital models characterized by having and accurate description of real objects and surfaces is nowadays possible using both laser scanning surveys and Photogrammetric or Computer Vision methodologies. In conjunction with the development of these techniques also the request of photorealistic models raised up. Texture mapping allows this result mapping the image information onto the digital surface. In many cases commercial software applications require an high amount of manual modeling to achieve satisfactory results. In this work we present a novel texture mapping approach aiming in reducing at minimum the human intervention. In particular we are focusing on two different aspects: filtering out occluded areas from the texturing phase and the correction of color and brightness among images to reduce sharp texture transitions.