Automated Texture Extraction from Multiple Images to Support Site Model Refinement and Visualization

Abstract

Texture mapping has wide and important applications in visualization and virtual reality. Surface texture extraction from a single image su ers from perspective distortion, data de ciency, and corruption caused by shadows and occlusions. In this paper, a system is developed for automated acquisition of complete and consistent texture maps from multiple images in order to support subsequent detailed surface analysis and scene rendering. Given camera and light source parameters for each image, and a geometric model of the scene, the textures of object surfaces are systematically collected into an organized orthographic library. Occlusions and shadows caused by objects in the scene are computed and associated with each retrieved surface. A \Best Piece Representation " algorithm is designed to combine intensities from multiple views, resulting in a unique surface intensity representation. Detailed surface structures, such as windows and doors, are extracted from the uniquely represented surface images to rene the geometric model. Experiments show successful applications of this approach to model re nement and scene visualization.