Improving LiDAR-based tree species mapping in Central European mixed forests using multi-temporal digital aerial colour-infrared photographs

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Digital colour-infrared (CIR) aerial photographs, which have been collected routinely in many parts of the world, are an invaluable data source for the monitoring and assessment of forest resources. Yet, the potential of these data for automated individual tree species mapping remains largely unexplored. One way to maximize the usefulness of digital CIR aerial photographs for individual tree species mapping is to integrate them with modern and complementary remote sensing technologies such as the light detection and ranging (LiDAR) system and 3D segmentation algorithms. In this study, we examined whether multi-temporal digital CIR orthophotos could be used to further increase the accuracy of airborne LiDAR-based individual tree species mapping for a temperate mixed forest in eastern Germany. Our results showed that the texture features captured by multi-temporal digital CIR orthophotos under different view-illumination conditions were species-specific. As a consequence, combining these texture features with LiDAR metrics significantly improved tree species mapping accuracy (overall accuracy: 77.4%, kappa: 0.68) compared to using LiDAR data alone (overall accuracy: 69.3%, kappa: 0.58). Among various texture features, the average gray level in the near-infrared band was found to contribute most to the classification. Our results suggest that the synergic use of multi-temporal digital aerial photographs and airborne LiDAR data has the potential to accurately classify individual tree species in Central European mixed forests.
Original languageEnglish
Article number101970
Pages (from-to)1-10
Number of pages10
JournalInternational Journal of Applied Earth Observation and Geoinformation (JAG)
Volume84
Early online date12 Sep 2019
DOIs
Publication statusE-pub ahead of print/First online - 12 Sep 2019

Fingerprint

mixed forest
photograph
Antennas
Color
Infrared radiation
aerial photograph
Textures
orthophoto
texture
forest resource
Remote sensing
temperate forest
segmentation
Lighting
detection
near infrared
Monitoring
remote sensing
monitoring

Keywords

  • ITC-ISI-JOURNAL-ARTICLE

Cite this

@article{0f0b5fa5e55c4c2caeb6f8fc6f8e821c,
title = "Improving LiDAR-based tree species mapping in Central European mixed forests using multi-temporal digital aerial colour-infrared photographs",
abstract = "Digital colour-infrared (CIR) aerial photographs, which have been collected routinely in many parts of the world, are an invaluable data source for the monitoring and assessment of forest resources. Yet, the potential of these data for automated individual tree species mapping remains largely unexplored. One way to maximize the usefulness of digital CIR aerial photographs for individual tree species mapping is to integrate them with modern and complementary remote sensing technologies such as the light detection and ranging (LiDAR) system and 3D segmentation algorithms. In this study, we examined whether multi-temporal digital CIR orthophotos could be used to further increase the accuracy of airborne LiDAR-based individual tree species mapping for a temperate mixed forest in eastern Germany. Our results showed that the texture features captured by multi-temporal digital CIR orthophotos under different view-illumination conditions were species-specific. As a consequence, combining these texture features with LiDAR metrics significantly improved tree species mapping accuracy (overall accuracy: 77.4{\%}, kappa: 0.68) compared to using LiDAR data alone (overall accuracy: 69.3{\%}, kappa: 0.58). Among various texture features, the average gray level in the near-infrared band was found to contribute most to the classification. Our results suggest that the synergic use of multi-temporal digital aerial photographs and airborne LiDAR data has the potential to accurately classify individual tree species in Central European mixed forests.",
keywords = "ITC-ISI-JOURNAL-ARTICLE",
author = "Y. Shi and Tiejun Wang and A.K. Skidmore and Marco Heurich",
year = "2019",
month = "9",
day = "12",
doi = "10.1016/j.jag.2019.101970",
language = "English",
volume = "84",
pages = "1--10",
journal = "International Journal of Applied Earth Observation and Geoinformation (JAG)",
issn = "1569-8432",
publisher = "Elsevier",

}

TY - JOUR

T1 - Improving LiDAR-based tree species mapping in Central European mixed forests using multi-temporal digital aerial colour-infrared photographs

AU - Shi, Y.

AU - Wang, Tiejun

AU - Skidmore, A.K.

AU - Heurich, Marco

PY - 2019/9/12

Y1 - 2019/9/12

N2 - Digital colour-infrared (CIR) aerial photographs, which have been collected routinely in many parts of the world, are an invaluable data source for the monitoring and assessment of forest resources. Yet, the potential of these data for automated individual tree species mapping remains largely unexplored. One way to maximize the usefulness of digital CIR aerial photographs for individual tree species mapping is to integrate them with modern and complementary remote sensing technologies such as the light detection and ranging (LiDAR) system and 3D segmentation algorithms. In this study, we examined whether multi-temporal digital CIR orthophotos could be used to further increase the accuracy of airborne LiDAR-based individual tree species mapping for a temperate mixed forest in eastern Germany. Our results showed that the texture features captured by multi-temporal digital CIR orthophotos under different view-illumination conditions were species-specific. As a consequence, combining these texture features with LiDAR metrics significantly improved tree species mapping accuracy (overall accuracy: 77.4%, kappa: 0.68) compared to using LiDAR data alone (overall accuracy: 69.3%, kappa: 0.58). Among various texture features, the average gray level in the near-infrared band was found to contribute most to the classification. Our results suggest that the synergic use of multi-temporal digital aerial photographs and airborne LiDAR data has the potential to accurately classify individual tree species in Central European mixed forests.

AB - Digital colour-infrared (CIR) aerial photographs, which have been collected routinely in many parts of the world, are an invaluable data source for the monitoring and assessment of forest resources. Yet, the potential of these data for automated individual tree species mapping remains largely unexplored. One way to maximize the usefulness of digital CIR aerial photographs for individual tree species mapping is to integrate them with modern and complementary remote sensing technologies such as the light detection and ranging (LiDAR) system and 3D segmentation algorithms. In this study, we examined whether multi-temporal digital CIR orthophotos could be used to further increase the accuracy of airborne LiDAR-based individual tree species mapping for a temperate mixed forest in eastern Germany. Our results showed that the texture features captured by multi-temporal digital CIR orthophotos under different view-illumination conditions were species-specific. As a consequence, combining these texture features with LiDAR metrics significantly improved tree species mapping accuracy (overall accuracy: 77.4%, kappa: 0.68) compared to using LiDAR data alone (overall accuracy: 69.3%, kappa: 0.58). Among various texture features, the average gray level in the near-infrared band was found to contribute most to the classification. Our results suggest that the synergic use of multi-temporal digital aerial photographs and airborne LiDAR data has the potential to accurately classify individual tree species in Central European mixed forests.

KW - ITC-ISI-JOURNAL-ARTICLE

UR - https://ezproxy2.utwente.nl/login?url=https://library.itc.utwente.nl/login/2020/isi/tiejunwang_imp.pdf

U2 - 10.1016/j.jag.2019.101970

DO - 10.1016/j.jag.2019.101970

M3 - Article

VL - 84

SP - 1

EP - 10

JO - International Journal of Applied Earth Observation and Geoinformation (JAG)

JF - International Journal of Applied Earth Observation and Geoinformation (JAG)

SN - 1569-8432

M1 - 101970

ER -