Extraction of surface texture data from low quality photographs to aid the construction of virtual reality models of archaeological sites

Abstract

A tool has been designed and implemented to use information extracted from photographs captured using uncalibrated cameras (so-called casual photographs) to fill the occlusions which occur in three-dimensional models of photogrammetrically captured sites. Capturing the geometry of archaeological sites by photogrammetric means is relatively expensive and, because of the layouts typical of such sites, usually results in a degree of occlusion. Occlusions are filled by extracting texture and calculating hidden geometry from casual photographs with the support of three-dimensional geometric data gleaned from the photogrammetric survey. The essential philosophy underlying the tool is to segment each occlusion into surfaces which may be approximated using curves and then use known geometry in the region of the occlusion to calculate the most probable locations of the junctions of such surface segments. The tool is primarily a combination of existing techniques for pre-filtering and calibrating the casual photograph, boundary detection and ultimately texture adjustment. The technique implemented for calculating the locations of occluded comers using minimisation of least square errors is new. The tool has been applied to occlusions of the various configurations that are expected to be typical of archaeological sites and has been found to deal well with such features and to provide accurate patches from typical data sets. It is also shown that the three-dimensional geometric model is clearly improved by the filling-in of the occlusion.