Journal of Sensor Technology

Volume 2, Issue 3 (September 2012)

ISSN Print: 2161-122X   ISSN Online: 2161-1238

Google-based Impact Factor: 1.38  Citations  

Lookup Table Hough Transform for Real Time Range Image Segmentation and Featureless Co-Registration

Full-Text HTML  XML Download Download as PDF (Size: 1231KB)  PP. 148-154  
DOI: 10.4236/jst.2012.23021    5,803 Downloads   8,805 Views   Citations

ABSTRACT

The paper addresses range image segmentation, particularly of data recorded by range cameras, such as the Microsoft Kinect and the Mesa Swissranger SR4000. These devices record range images at video frame rates and allow for acquisition of 3-dimensional measurement sequences that can be used for 3D reconstruction of indoor environments from moving platforms. The role of segmentation is twofold. First the necessary image co-registration can be based on corresponding segments, instead of corresponding point features (which is common practice currently). Secondly, the segments can be used during subsequent object modelling. By realisising that planar regions in disparity images can be modelled as linear functions of the image coordinates, having integer values for both domain and range, the paper introduces a lookup table based implementation of local Hough transform, allowing to obtain good segmentation results at high speeds.

Cite this paper

B. Gorte and G. Sithole, "Lookup Table Hough Transform for Real Time Range Image Segmentation and Featureless Co-Registration," Journal of Sensor Technology, Vol. 2 No. 3, 2012, pp. 148-154. doi: 10.4236/jst.2012.23021.

Copyright © 2019 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.