Methods for measuring height and planimetry discrepancies in airborne laserscanner data

tems is limited by the laser power and receiver sensitivity. Airborne laserscanning (or lidar) has become a very important Together with a narrow opening angle for minimizing occlutechnique for the acquisition of digital terrain model data. sions, this results in laserscanner datasets consisting of many Beyond this, the technique is increasingly being used for the parallel strips with a width of several hundred meters in most acquisition of point clouds for 3D modeling of a wide range of cases. objects, such as buildings, vegetation, or electrical power lines. Over the last few years, airborne laserscanning has gained As an active technique, airborne laserscanning offers a high a lot of attention and has become a leading technique for the reliability even over terrain with poor image contrast. The acquisition of digital terrain model data. Its advantages are its precision of the technique is often specified to be on the order fast and efficient manner of data acquisition as well as its high of one to two decimeters. By reason of its primary use in digital reliability and precision potential. Since the early nineties, airterrain modeling, examinations of the precision potential of borne laserscanninghas been usedfor special taskssuch as,for airborne laserscanning have so far been concentrated on the example, beach erosion monitoring. Meanwhile, it is replacing height precision. With the use of the technique for general 3D conventional stereo-imaging-based photogrammetric techreconstruction tasks and the increasing resolution of niques andis acceptedas a generaltool for digitalterrain model laserscanner systems, the planimetric precision of data acquisition. The Netherlands were the first country to laserscanner point clouds becomes an important issue. generate a nationwide digital terrain model purely based on In addition to errors in the laser distance meter and the laserscanner data (e.g., Wouters and Bollweg, 1998). deflecting mirror system, the error budget of airborne While early systems offered data rates on the order of 2 to 7 laserscanning instruments is strongly influenced by the GPS/ kHz (i.e., 2000 to 7000 3D surface points per second), modern INS systems used for sensor pose (position and orientation) systems come with data rates of 25 up to 83 kHz. This gain in determination. Errors of these systems often lead to the temporal resolution has not only increased the efficiency of deformation of laserscanner data strips and may become data acquisition, but has also opened a whole range of new evident as discrepancies in the overlap region between application fields to airborne laserscanning. It broadens the neighboring strips in a block of laserscanner data. The paper scope of the technique beyond the pure acquisition of digital presents least-squares matching implemented on a TIN terrain models to a more general tool for the acquisition of point structure as a general tool for the determination of laser- clouds for 3D modeling of a wide range of objects. High resoluscanner strip discrepancies in all three coordinate directions, tion laserscanner data have proven to be a valuable source for using both height and reflectance data. Practical problems of the automatic generation of 3D building models (Haala and applying matching techniques to 2.5D laserscanner point Brenner, 1997; Maas and Vosselman, 1999). Further examples clouds are discussed and solved, and the success of the of new application fields are the determination of forest stand technique is shown on the basis of several datasets. Applying parameters and corridor mapping. least-squares matching techniques to dense laserscanner data in a TIN structure, strip discrepancies can be determined with The precision potential of the technique is often specified as 1 to 2 decimeters. Due to the primary application field in the centimeter precision for the height coordinate and decimeter precision for the planimetric coordinates.