Dr. Thomas Melzer, Institute for Photogrammetry and Remote Sensing, received the Best Paper Award 2007 for his article

"Non-parametric segmentation of ALS point clouds using mean shift"

in Journal of Applied Geodesy. Band 1, Heft 3, Seiten 159–170, ISSN (Online) 1862-9024, ISSN (Print) 1569-3988, November 2007

Abstract:

Segmentation is a key task in the processing of 3D point clouds as obtained from airborne laser scanners (ALS). However, most of the segmentation techniques currently employed require prior gridding of the data and thus do not respect the inherently three-dimensional geometry of more intricate structures such as power lines. By contrast, the mean shift procedure, a filtering and clustering approach which has recently found much interest in the image processing community, works directly on the original 3D point cloud; also, mean shift is a non-parametric technique (i.e., it does not depend on any geometric model assumptions) and can thus also be applied to vegetation structures. In this paper, we will give a self-contained derivation of the mean shift procedure, and discuss how it can be used to obtain a classification or segmentation of an unstructured 3D point cloud. Two application examples shall further illustrate its usefulness to ALS data processing.

Figure 1 (2 pictures on the left): (a) Part of the Piestingtal scan, the color coding reflects the result of the mean shift clustering. (c) Side view of the computed vegetation clusters, which also demonstrates the challenging topography of the project area. Note that the pylons have also been classified as vegetation.

Figure 2 (4 pictures on the right): (a) Aerial photograph of the test area near Schönbrunn palace in Vienna. (b)–(c) show a planimetric view of the corresponding ALS point cloud, where the colors reflect the result of the mean shift segmentation. (b) Only the 3D coordinates of the points were used as features. (c) The 3D coordinates, the pulse width and the amplitude of the echo signal were used as features. (d) 3D view of (b).