Automatic Photo Orientation via Matching with Control Patches

Traditional aerial-triangulation has long relied on control points orientating the photo models into a ground-based coordinate system, reducing the distortion effect when tying photos by imperfect photo measurements, and calibrating camera inner parameters. Field surveying preparing for adequate number of control points and manual measurements of the control points afterwards on the photos cost considerably both in labour work and expense. During the past decade, digital photogrammetry by integrating the pros of image processing technique and rapid computation via the computer into photogrammetric discipline has made database-supplied control points as well as the automation of control point measurement truly possible. In this study, we employ the control patches, which have been created in previous photogrammetric projects with each feature point known in the object space as the centre of each patch, serving for the candidates of the control points that are likely to be found in the newly taken photos by utilizing image matching technique. Among others, the successful implementation of the above idea lies in the underlying factors: (1). The quantity and quality of control patches; (2). Predicting the control patches and projecting them onto new photos; (3). Alignment of control patches with respect to the new photos; (4). Generating the equivalent ground elements of control patches versus the new photos for the purpose of correlation; (5). Developing effective matching methods and matching strategy; (6). Refining the exterior orientation parameters. Furthermore, whenever at least three matched control patches succeed in a single photo, it follows that single photo orientation is applicable. The experiments under this work suggest the potential efficiency of automatic control point measurement from control patch database and photo orientation by the proposed workflow. * Corresponding author.