Multiple-Patch Matching in the Object Space for Aerotriangulation
Loading...
Date
1994-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Ohio State University. Division of Geodetic Science
Abstract
Aerotriangulation has been a well-established photogrammetric procedure for years. With recent developments in digital photogrammetry, it is possible to automate the selection and transfer of tie points as well as the measurement. This increases the efficiency, accuracy and reliability of aerotriangulation projects. This work is concerned with finding an automated method for accurately measuring the tie points. Although accurate matching has been the subject of numerous research studies in the fields of photogrammetry and computer vision, its application for measuring tie points in aerotriangulation adds some new aspects that have not been fully addressed previously. Such aspects are, for example, lack of exterior orientation parameters, very high accuracy requirements, matching only one point at each area and consistent matching of more than two overlapping image patches. The method proposed in this dissertation is based on matching warped image patches, or matching in the object space. The warping procedure requires the knowledge of the object surface around each prospective point. The surface is reconstructed by a hierarchical, iterative approach. Both the surface and the set of orientation parameters of all the photographs in the block are improved throughout the iterative procedure. Experiments were performed to demonstrate the feasibility of the proposed method and to study its effects. The results show that where man-made objects exist, matching warped images outperforms the traditional image space matching techniques. The accuracy of the results is comparable to manual measurements even when relatively low resolution images are used.