Title: A new 6D ICP algorithm with color segmentation: based adaptive sampling
Authors: Kim, Eung-Su
Sung-In, Choi
Park, Soon-Yong
Citation: WSCG 2015: poster papers proceedings: 23rd International Conference in Central Europe on Computer Graphics, Visualization and Computer Visionin co-operation with EUROGRAPHICS Association, p. 35-39.
Issue Date: 2015
Publisher: Václav Skala - UNION Agency
Document type: konferenční příspěvek
URI: wscg.zcu.cz/WSCG2015/CSRN-2503.pdf
ISBN: 978-80-86943-67-1
ISSN: 2464-4617
Keywords: ICP z bodu do roviny;ICP;3D registrace;segmentace barev
Keywords in different language: point-to-plane ICP;ICP;3D registration;color segmentation
Abstract: In ICP-based algorithms, the closest points are considered as the corresponding points. However, this method fails to find matching points accurately when the initial position of the point clouds is not sufficiently close. In this paper, we propose a new method to solve this problem using six-dimensional (6D) distance, which consists of color information and three-dimensional (3D) distance, and color distribution matching. First, before finding the corresponding points using this method, a Gaussian filter is applied on the input color image. A color based image segmentation is done on that image and then 𝑛 number of samples are randomly chosen from each segment. This process is applied in order to improve the computational time and performance. Second, corresponding point candidates are searched by solving a local minima problem using 6D distance. Then the color distribution matching is applied on these candidates to find the final corresponding point. Several experiments are conducted to evaluate the proposed method and the experimental results prove it has improved over the conventional methods.
Rights: © Václav Skala - Union Agency
Appears in Collections:WSCG 2015: Poster Papers Proceedings

Files in This Item:
File Description SizeFormat 
Kim.pdfPlný text788,77 kBAdobe PDFView/Open

Please use this identifier to cite or link to this item: http://hdl.handle.net/11025/29563

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.