Towards Automated LiDAR Boresight Self-calibration

This paper focuses on practical aspects when performing boresight calibration in airborne laser scanning using rigorous methodology implemented in LIBOR software. LIBOR technique, introduced by (Skaloud and Lichti, 2006), is based on expressing the boresight calibration parameters within the direct-georeferencing equation separately for each target point and conditioning a group of points to lie on a common planar surface. Although there is no need for a priori information about the plane parameters as these are part of the unknowns, good estimation requires implication of various planar features that differ in slope and orientation. Such conditions are typically fulfilled in residential-urban areas where the presence of planes in form of roof-tops is abundant. These are identified by grouping points belonging to the same surface into a distinct class separately for each flight line and finding class-correspondences among the flight lines. We present an automated approach for this selection process that stems from intrinsic geometry of curved surfaces. This classification is followed by additional fine-filtering for returns from features as chimneys, antennas etc. The presented discussion focuses on practical examples with data from continuously-rotating and oscillating-mirror systems. These findings show that good automated point selection is possible and acts as a pre-requisite to robust estimates of all boresight angles with accuracy that is several times superior to the system noise level.