Validation of 3-D Reconstruction from Sequence of Images

A number of approaches have been proposed in the literature for reconstruction of 3-D objects from sequence of images. Yet, very few studies have been reported on the quantification/validation of the accuracy of these reconstructions. In addition, no design has been reported for a generic vision platform that can allow various modality imaging. The purpose of this paper is two folds: 1) propose a vision platform that lend itself for acquisition of calibrated sequence of images, and concurrently obtain a direct 3-D reconstruction by laser scanning; and 2) develop and implement different approaches for 3-D reconstructions from sequence of images. 3-D reconstructions will be evaluated against the 3-D scanning generated from a laser scanner. Validation of the reconstructions is made by pair-wise comparison with the 3-D scanning results. In addition, other validation approaches are suggested if the ground truth is not available. Preliminary studies with the proposed vision platform show a good promise for its use in the validation of various 3-D reconstructions.