AUTOMATING QUALITY CONTROL FOR AERIAL MAPPING USING DENSE POINT CLOUDS

Quality control (QC) in aerial mapping is a predominantly manual process, carried out in various steps throughout the entire data processing chain. One of these steps is the verification of triangulation results, which comprises interactive measurements of control points for absolute accuracy as well as relative offsets in-between overlapping imagery. We present a way to automate this relative QC, using photogrammetrically derived point clouds. With panchromatic intensity and full color information added, these point clouds become “image point clouds”. Computed pair-wise in overlapping areas, the combined geometric and radiometric information of such image point clouds enables the robust derivation of three-dimensional offsets. These are presented to the user along with reliability measures in condensed reports that provide comprehensive assessment of the relative geometric accuracy within a flight recording or block. This automated method, called “Shear Analysis”, is used at North West Geomatics Ltd. (North West) to evaluate Leica ADS line-scanner data before and after aerial triangulation. We outline the combined geometric/radiometric image point cloud matching approach and verify its results in comparison to manual measurements that are traditionally used in QC. The automatic Shear Analysis is discussed in detail for an ADS block from North West’s production, comparing the orientations from Direct Georeferencing with two triangulation results, one of them based on manual image point measurements and the other using automatic point measurement (APM).