Coarse orientation of terrestrial laser scans in urban environments

The use of terrestrial laser scanners is becoming increasingly popular. For the acquisition of larger scenes, it is usually necessary to align all scans to a common reference frame. While there are methods using direct measurement of the orientation, due to simplicity and costs, mostly artificial targets are used. This works reliably, but usually adds a substantial amount of time to the acquisition process. Methods to align scans using the scan data itself have been known for a long time, however, being iterative, they need good initial values.In this paper, we investigate two different methods targeted at the determination of suitable initial values. The first one is based on a symbolic approach, using corresponding features to compute the orientation. The second one is based on an iterative alignment scheme originally proposed in the robotics domain. To assess the performance of both methods, a set of 20 scans has been acquired systematically along a trajectory in a downtown area. Reference orientations were obtained by a standard procedure using artificial targets. We present the results of both methods regarding convergence and accuracy, and compare their performance.