Share Email Print
cover

Journal of Applied Remote Sensing

Feature-constrained registration of building point clouds acquired by terrestrial and airborne laser scanners
Author(s): Hangbin Wu; Marco Scaioni; Hanyan Li; Nan Li; Minfeng Lu; Chun Liu
Format Member Price Non-Member Price
PDF $20.00 $25.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Point-cloud registration is usually accomplished on the basis of several corresponding features to compute the parameters of the transformation model. However, common point features are difficult to select because airborne laser scanner (ALS) and terrestrial laser scanner (TLS) point clouds of the same object have be aligned due to the different sensing positions and sampling modes. Taking building profile features as objects, a registration method based on feature constraints is proposed here. The standard six-parameter rigid-body transformation adopted for alignment of laser scans is replaced by a two-step transformation: horizontal registration based on a two-dimensional similarity transformation and vertical registration based on a simple vertical shift. First, the feature-line and feature-plane equation parameters are obtained from both the airborne and terrestrial point clouds. Second, the plane transformation parameters are computed after projecting the extracted features onto a horizontal reference plane. Finally, the elevation transformation parameter is calculated by comparing the heights of flat features. The ALS and TLS datasets of two buildings (Shanghai Pudong International Conference Center and Shanghai Ocean Aquarium, China) were used to evaluate the robustness and accuracy. The results show that the proposed feature-constrained method works well for registration between two datasets. Five checkpoints and one overlap zone for the Pudong International Conference Center were selected to evaluate the accuracy and resulted in accuracies of 0.15 to 0.5 m in the horizontal direction and 0.20 m in the vertical direction.

Paper Details

Date Published: 18 July 2014
PDF: 22 pages
J. Appl. Remote Sens. 8(1) 083587 doi: 10.1117/1.JRS.8.083587
Published in: Journal of Applied Remote Sensing Volume 8, Issue 1
Show Author Affiliations
Hangbin Wu, Tongji Univ. (China)
Marco Scaioni, Tongji Univ. (China)
Hanyan Li, Georgia Institute of Technology (United States)
Nan Li, Tongji Univ. (China)
Minfeng Lu, Tongji Univ. (China)
Chun Liu, Tongji Univ. (China)


© SPIE. Terms of Use
Back to Top