Urban Point Cloud Mining Based on Density Clustering and MapReduce

Harith Aljumaily, Debra Laefer, Dolores Cuadra

Research output: Contribution to journalArticle

Abstract

This paper proposes an approach to classify, localize, and extract automatically urban objects such as buildings and the ground surface from a digital surface model created from aerial laser scanning data. To achieve that, the approach involves three steps: (1) dividing the original data into smaller, more manageable pieces using a method based on MapReduce gridding for subspace partitioning, (2) applying the DBSCAN algorithm to identify interesting subspaces depending on point density, and (3) grouping of identified subspaces to form potential objects. Validation of the method was conducted in an architecturally dense and complex portion of Dublin, Ireland. The best results were achieved with a 1-m3-sized clustering cube, for which the number of classified clusters most closely equaled that which was derived manually (correctness=84.91%, completeness=84.39%, and quality=84.65%).

Original languageEnglish (US)
Article number04017021
JournalJournal of Computing in Civil Engineering
Volume31
Issue number5
DOIs
StatePublished - Sep 1 2017

Fingerprint

Antennas
Scanning
Lasers

Keywords

  • Big data
  • Building extraction
  • Clustering classification approaches
  • Density-based spatial clustering of applications with noise (DBSCAN) algorithm
  • Light detection and ranging (LiDAR)
  • MapReduce

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Computer Science Applications

Cite this

Urban Point Cloud Mining Based on Density Clustering and MapReduce. / Aljumaily, Harith; Laefer, Debra; Cuadra, Dolores.

In: Journal of Computing in Civil Engineering, Vol. 31, No. 5, 04017021, 01.09.2017.

Research output: Contribution to journalArticle

@article{c91380852aa94d04973b3b6b1943bce2,
title = "Urban Point Cloud Mining Based on Density Clustering and MapReduce",
abstract = "This paper proposes an approach to classify, localize, and extract automatically urban objects such as buildings and the ground surface from a digital surface model created from aerial laser scanning data. To achieve that, the approach involves three steps: (1) dividing the original data into smaller, more manageable pieces using a method based on MapReduce gridding for subspace partitioning, (2) applying the DBSCAN algorithm to identify interesting subspaces depending on point density, and (3) grouping of identified subspaces to form potential objects. Validation of the method was conducted in an architecturally dense and complex portion of Dublin, Ireland. The best results were achieved with a 1-m3-sized clustering cube, for which the number of classified clusters most closely equaled that which was derived manually (correctness=84.91{\%}, completeness=84.39{\%}, and quality=84.65{\%}).",
keywords = "Big data, Building extraction, Clustering classification approaches, Density-based spatial clustering of applications with noise (DBSCAN) algorithm, Light detection and ranging (LiDAR), MapReduce",
author = "Harith Aljumaily and Debra Laefer and Dolores Cuadra",
year = "2017",
month = "9",
day = "1",
doi = "10.1061/(ASCE)CP.1943-5487.0000674",
language = "English (US)",
volume = "31",
journal = "Journal of Computing in Civil Engineering",
issn = "0887-3801",
publisher = "American Society of Civil Engineers (ASCE)",
number = "5",

}

TY - JOUR

T1 - Urban Point Cloud Mining Based on Density Clustering and MapReduce

AU - Aljumaily, Harith

AU - Laefer, Debra

AU - Cuadra, Dolores

PY - 2017/9/1

Y1 - 2017/9/1

N2 - This paper proposes an approach to classify, localize, and extract automatically urban objects such as buildings and the ground surface from a digital surface model created from aerial laser scanning data. To achieve that, the approach involves three steps: (1) dividing the original data into smaller, more manageable pieces using a method based on MapReduce gridding for subspace partitioning, (2) applying the DBSCAN algorithm to identify interesting subspaces depending on point density, and (3) grouping of identified subspaces to form potential objects. Validation of the method was conducted in an architecturally dense and complex portion of Dublin, Ireland. The best results were achieved with a 1-m3-sized clustering cube, for which the number of classified clusters most closely equaled that which was derived manually (correctness=84.91%, completeness=84.39%, and quality=84.65%).

AB - This paper proposes an approach to classify, localize, and extract automatically urban objects such as buildings and the ground surface from a digital surface model created from aerial laser scanning data. To achieve that, the approach involves three steps: (1) dividing the original data into smaller, more manageable pieces using a method based on MapReduce gridding for subspace partitioning, (2) applying the DBSCAN algorithm to identify interesting subspaces depending on point density, and (3) grouping of identified subspaces to form potential objects. Validation of the method was conducted in an architecturally dense and complex portion of Dublin, Ireland. The best results were achieved with a 1-m3-sized clustering cube, for which the number of classified clusters most closely equaled that which was derived manually (correctness=84.91%, completeness=84.39%, and quality=84.65%).

KW - Big data

KW - Building extraction

KW - Clustering classification approaches

KW - Density-based spatial clustering of applications with noise (DBSCAN) algorithm

KW - Light detection and ranging (LiDAR)

KW - MapReduce

UR - http://www.scopus.com/inward/record.url?scp=85032210646&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85032210646&partnerID=8YFLogxK

U2 - 10.1061/(ASCE)CP.1943-5487.0000674

DO - 10.1061/(ASCE)CP.1943-5487.0000674

M3 - Article

AN - SCOPUS:85032210646

VL - 31

JO - Journal of Computing in Civil Engineering

JF - Journal of Computing in Civil Engineering

SN - 0887-3801

IS - 5

M1 - 04017021

ER -