Accuracy Study of Airborne Laser Scanning Data with Photogrammetry

This paper describes an accuracy study of airborne laser scanning data obtained by the Airborne Topographic Mapper (ATM) laser system over Ocean City, Md. The ATM is a conical scanning laser altimeter developed by NASA for precise measurement of surface elevation changes in polar ice sheets, ocean beaches and drainage systems. First, we determine the “internal" accuracy of the system by comparing data from different flight missions. This is followed by a comparison of the merged laser data sets with surface elevations obtained by photogrammetry. Large-scale aerial photographs have been acquired over the test area and an aerial triangulation was performed to determine the exterior orientation parameters. The comparison consists of several experiments that were performed with the digitized photographs and the laser points. First we determine how well the laser points agree with the visible surface as defined by two overlapping images (stereopsis). This is accomplished by backprojecting the laser points to the images based on their exterior orientation parameters. The location of the laser points in the images serve as initial approximations for image matching. We use an adaptive least-squares matching procedure with a variable template size. A non-zero matching vector indicates discrepancies between laser points and photogrammetry. The purpose of the second experiment is to estimate the horizontal accuracy of laser points. One way to accomplish this is to extract linear features and to compare them. Linear features in laser point data sets can only be determined indirectly, e.g. by intersecting planar surface patches. In contrast, linear features in aerial images can be determined directly by an edge operator. We used the Canny operator to extract edges in the images and feature-based matching to find corresponding edges in the stereopair. After describing the procedure, experimental results are reported.