An automatic registration method for adjustment of relative elevation discrepancies between lidar data strips

Jaebin Lee, Dongyeop Han, Kiyun Yu, Joon Heo, Ayman Habib

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Despite the recent development in light detection and ranging (LIDAR) systems, discrepancies between strips in overlapping areas persist because of systematic errors. During the past decade, several methods have been developed for compensating for errors, such as checking the coincidence of conjugate features extracted from overlapping LIDAR strips or comparing interpolated range and intensity images. However, these approaches rely upon the ability to detect and extract suitable conjugate features within the overlap area and/or during the preprocessing of raw LIDAR data in, for example, interpolation or segmentation. Such procedures make the overall process complex and may impose limitations on the development of an automated method. Furthermore, some of the preprocessing techniques, such as raster interpolation, may induce errors in raw LIDAR data when dealing with large-scale coverage over an urban area. This paper therefore presents an automated approach, working with raw LIDAR data without the restrictions associated with using conjugate features and without any preprocessing. We present an approach using changes in local height variations that occur within the overlap area between two neighboring strips. In this case, local height variations of the LIDAR data in the overlap area increase if there are discrepancies. This scheme can be helpful in determining an appropriate transformation for the adjustment of discrepancies between neighboring LIDAR strips in a way that minimizes the local height variation. A contour tree (CT) was used to represent the local height variations and to find an appropriate initial transformaunction. The iterative closest point (ICP) procedure was then applied to refine the function parameters. Following transformation, LIDAR strips were registered with each other, and the discrepancies were measured again to determine whether they had been resolved. The statistical evaluation of the results revealed that the discrepancies were significantly reduced.

Original languageEnglish
Pages (from-to)115-134
Number of pages20
JournalGIScience and Remote Sensing
Volume47
Issue number1
DOIs
Publication statusPublished - 2010 Jan 1

Fingerprint

lidar
interpolation
detection
method
registration
raster
segmentation
urban area

All Science Journal Classification (ASJC) codes

  • Earth and Planetary Sciences(all)

Cite this

Lee, Jaebin ; Han, Dongyeop ; Yu, Kiyun ; Heo, Joon ; Habib, Ayman. / An automatic registration method for adjustment of relative elevation discrepancies between lidar data strips. In: GIScience and Remote Sensing. 2010 ; Vol. 47, No. 1. pp. 115-134.
@article{cef58a10bb804ce18aca2d63258287cd,
title = "An automatic registration method for adjustment of relative elevation discrepancies between lidar data strips",
abstract = "Despite the recent development in light detection and ranging (LIDAR) systems, discrepancies between strips in overlapping areas persist because of systematic errors. During the past decade, several methods have been developed for compensating for errors, such as checking the coincidence of conjugate features extracted from overlapping LIDAR strips or comparing interpolated range and intensity images. However, these approaches rely upon the ability to detect and extract suitable conjugate features within the overlap area and/or during the preprocessing of raw LIDAR data in, for example, interpolation or segmentation. Such procedures make the overall process complex and may impose limitations on the development of an automated method. Furthermore, some of the preprocessing techniques, such as raster interpolation, may induce errors in raw LIDAR data when dealing with large-scale coverage over an urban area. This paper therefore presents an automated approach, working with raw LIDAR data without the restrictions associated with using conjugate features and without any preprocessing. We present an approach using changes in local height variations that occur within the overlap area between two neighboring strips. In this case, local height variations of the LIDAR data in the overlap area increase if there are discrepancies. This scheme can be helpful in determining an appropriate transformation for the adjustment of discrepancies between neighboring LIDAR strips in a way that minimizes the local height variation. A contour tree (CT) was used to represent the local height variations and to find an appropriate initial transformaunction. The iterative closest point (ICP) procedure was then applied to refine the function parameters. Following transformation, LIDAR strips were registered with each other, and the discrepancies were measured again to determine whether they had been resolved. The statistical evaluation of the results revealed that the discrepancies were significantly reduced.",
author = "Jaebin Lee and Dongyeop Han and Kiyun Yu and Joon Heo and Ayman Habib",
year = "2010",
month = "1",
day = "1",
doi = "10.2747/1548-1603.47.1.115",
language = "English",
volume = "47",
pages = "115--134",
journal = "GIScience and Remote Sensing",
issn = "1548-1603",
publisher = "Bellwether Publishing, Ltd.",
number = "1",

}

An automatic registration method for adjustment of relative elevation discrepancies between lidar data strips. / Lee, Jaebin; Han, Dongyeop; Yu, Kiyun; Heo, Joon; Habib, Ayman.

In: GIScience and Remote Sensing, Vol. 47, No. 1, 01.01.2010, p. 115-134.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An automatic registration method for adjustment of relative elevation discrepancies between lidar data strips

AU - Lee, Jaebin

AU - Han, Dongyeop

AU - Yu, Kiyun

AU - Heo, Joon

AU - Habib, Ayman

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Despite the recent development in light detection and ranging (LIDAR) systems, discrepancies between strips in overlapping areas persist because of systematic errors. During the past decade, several methods have been developed for compensating for errors, such as checking the coincidence of conjugate features extracted from overlapping LIDAR strips or comparing interpolated range and intensity images. However, these approaches rely upon the ability to detect and extract suitable conjugate features within the overlap area and/or during the preprocessing of raw LIDAR data in, for example, interpolation or segmentation. Such procedures make the overall process complex and may impose limitations on the development of an automated method. Furthermore, some of the preprocessing techniques, such as raster interpolation, may induce errors in raw LIDAR data when dealing with large-scale coverage over an urban area. This paper therefore presents an automated approach, working with raw LIDAR data without the restrictions associated with using conjugate features and without any preprocessing. We present an approach using changes in local height variations that occur within the overlap area between two neighboring strips. In this case, local height variations of the LIDAR data in the overlap area increase if there are discrepancies. This scheme can be helpful in determining an appropriate transformation for the adjustment of discrepancies between neighboring LIDAR strips in a way that minimizes the local height variation. A contour tree (CT) was used to represent the local height variations and to find an appropriate initial transformaunction. The iterative closest point (ICP) procedure was then applied to refine the function parameters. Following transformation, LIDAR strips were registered with each other, and the discrepancies were measured again to determine whether they had been resolved. The statistical evaluation of the results revealed that the discrepancies were significantly reduced.

AB - Despite the recent development in light detection and ranging (LIDAR) systems, discrepancies between strips in overlapping areas persist because of systematic errors. During the past decade, several methods have been developed for compensating for errors, such as checking the coincidence of conjugate features extracted from overlapping LIDAR strips or comparing interpolated range and intensity images. However, these approaches rely upon the ability to detect and extract suitable conjugate features within the overlap area and/or during the preprocessing of raw LIDAR data in, for example, interpolation or segmentation. Such procedures make the overall process complex and may impose limitations on the development of an automated method. Furthermore, some of the preprocessing techniques, such as raster interpolation, may induce errors in raw LIDAR data when dealing with large-scale coverage over an urban area. This paper therefore presents an automated approach, working with raw LIDAR data without the restrictions associated with using conjugate features and without any preprocessing. We present an approach using changes in local height variations that occur within the overlap area between two neighboring strips. In this case, local height variations of the LIDAR data in the overlap area increase if there are discrepancies. This scheme can be helpful in determining an appropriate transformation for the adjustment of discrepancies between neighboring LIDAR strips in a way that minimizes the local height variation. A contour tree (CT) was used to represent the local height variations and to find an appropriate initial transformaunction. The iterative closest point (ICP) procedure was then applied to refine the function parameters. Following transformation, LIDAR strips were registered with each other, and the discrepancies were measured again to determine whether they had been resolved. The statistical evaluation of the results revealed that the discrepancies were significantly reduced.

UR - http://www.scopus.com/inward/record.url?scp=77949889811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77949889811&partnerID=8YFLogxK

U2 - 10.2747/1548-1603.47.1.115

DO - 10.2747/1548-1603.47.1.115

M3 - Article

AN - SCOPUS:77949889811

VL - 47

SP - 115

EP - 134

JO - GIScience and Remote Sensing

JF - GIScience and Remote Sensing

SN - 1548-1603

IS - 1

ER -