Premium
Automated Registration of 3D Scans Using Geometric Features and Normalized Color Data
Author(s) -
GómezGarcíaBermejo Jaime,
Zalama Eduardo,
Feliz Raúl
Publication year - 2013
Publication title -
computer‐aided civil and infrastructure engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.773
H-Index - 82
eISSN - 1467-8667
pISSN - 1093-9687
DOI - 10.1111/j.1467-8667.2012.00785.x
Subject(s) - point cloud , computer science , iterative closest point , computer vision , artificial intelligence , range (aeronautics) , point (geometry) , process (computing) , convergence (economics) , cultural heritage , automation , image registration , color correction , mathematics , image (mathematics) , geography , engineering , mechanical engineering , geometry , archaeology , aerospace engineering , economic growth , economics , operating system
Laser range sensors are playing an increased role in construction. These devices are used to collect a large number of points from different locations and then, those points are registered in a common framework. This article describes a new procedure for the registration of point clouds, especially suited to the fields of architecture and cultural heritage. Often, in these fields, the registration of point clouds is subject to errors due to the fact that an important number of points do not lie on particular geometric features. In this article, an accurate and efficient approach for 3D data registration based on Iterative Closest Point (ICP) algorithm is proposed, which takes advantage of the color data acquired along with range data. Points suitable for registration are selected according to their local geometry and/or color properties, thus a significant improvement on performance convergence and processing time is obtained. The algorithm performs an automatic, on‐the‐flight estimation of the overlapping region, taking into account possible color differences produced by lighting changes through the measurement process. The proposed approach has been tested on real scanned data from cultural heritage buildings and compared to other approaches, showing a better performance in terms of automation degree, accuracy, and speed.