Results on Range Image Segmentation for Service Robots

Abstract

This report presents an experimental evaluation of a plane extraction method using various line extraction algorithms. Four different algorithms are chosen, which are well known in mobile robotics and computer vision. Experiments are performed on two sets of 25 range images either obtained by simulation or acquired by a proprietary 3D laser scanner. The segmentation outcome of the simulated range images is measured in terms of an average segment classification ratio. Moreover, the speed of the method is measured to conclude on the suitability for service robot applications.

An Adaptive Technique for Accurate Feature Extraction from Regular and Irregular Image Data

“…Pratt [1] considered three major areas of error associated with the determination of an edge: missing valid edge points; failure to localise edge points; and classification of noise fluctuations as edge points. Pratt therefore introduced the FoM technique as one that balances these three types of error, defined as I is the ideal number of edge pixels, d is the separation distance of a detected edge point normal to a line of ideal edge points, and α is a scaling factor, most commonly chosen to be 9 1 , although this value may be adjusted to penalize edges that are localized but offset from the true edge position. Initially we evaluated the gradient operators with respect to their use Figure 2 presents comparative results on a vertical ramp edge using the Prewitt and Sobel operators (denoted P and S) and the proposed 3 3× , 5 5× and 7 7 × Bilinear-Gaussian gradient operators (denoted as BG3, BG5 and BG7).…”

“…This is largely because range imagery can be used to obtain reliable descriptions of 3-D scenes; a range image contains distance measurements from a selected reference point or plane to surface points of objects within a scene [4], allowing more information about the scenes to be recovered [3]. However, techniques that can be used for feature extraction directly from more than one image type and in particular that can be directly applied to both range and intensity images are not readily available [2,9]. Typically approaches for feature extraction in intensity or range images can only be applied principally to that specific image type; none of the typical approaches can be readily applied to both image types without image preprocessing.…”

Extraction of Geometrical Features in 3D Environments for Service Robotic Applications

“…An assessment of four different line extraction algorithms to separate points from different planar surfaces in service robotics has recently been published, [16]. In [3] two region-based algorithms for planar segmentation of range images were presented; they are both capable of processing probabilistic data in order to do 3D SLAM based upon planar features.…”

Roboternavigation in Büros und Wohnungen

“…Obwohl mö glich (Nuchter et al, 2005;Gachter, Nguyen, Siegwart), ist dies jedoch mechanisch sehr aufwendig, die Sensoren bauen groß und sind mit dem rotierenden Laserstrahl nicht unbedenklich. Inzwischen sind auch industrielle Produkte verfü gbar, z.…”