Deep learning-based approach for landform classification from integrated data sources of digital elevation model and imagery

Sijin Li, Liyang Xiong, Guoan Tang, Josef Strobl

Publikation: Beitrag in FachzeitschriftArtikelForschungBegutachtung

Abstract

Landform classification is one of the most important aspects in geomorphological research, dividing the Earth'ssurface into diverse geomorphological types. Thus, an accurate classification of landforms is a key procedure indescribing the topographic characteristics of a given area and understanding their inner geomorphological formation processes. However, landform types are not always independent of one another due to the complexityand dynamics of interior and external forces. Furthermore, transitional landforms with gradually changing sur-face morphologies are widely distributed on the Earth's surface. With this situation, classifying these complexand transitional landforms with traditional landform classification methods is hard. In this study, a deep learning(DL) algorithm was introduced, aiming at automatically classifying complex and transitional landforms. Thisalgorithm was trained to learn and extract landform features from integrated data sources. These integrateddata sources contain different combinations of imagery, digital elevation models (DEMs), and terrain derivatives.The Loess Plateau in China, which contains complex and transitional loess landforms, was selected as the studyarea for data training. In addition, two sample areas in the Loess Plateau with complex and transitional loess hill and ridge landforms were used to validate the classified landform types by using the proposed DL method.Meanwhile, a comparative analysis between the proposed DL and random forest (RF) methods was also con-ducted to investigate their capabilities in landform classification. The proposed DL approach can achieve thehighest landform classification accuracy of 87% in the transitional area with data combination of DEMs and im-ages. In addition, the proposed DL method can achieve a higher accuracy of landform classification with betterdefined landform boundaries compared to the RF method. The classified loess landforms indicate the differentlandform development stages in this area. Finally, the proposed DL method can be extended to other landformareas for classifying their complex and transitional landforms
OriginalspracheEnglisch
Seiten (von - bis)2-14
FachzeitschriftGeomorphology
Jahrgang354
DOIs
PublikationsstatusVeröffentlicht - 2020

Systematik der Wissenschaftszweige 2012

  • 105 Geowissenschaften

Zitieren

@article{c84e610955bd4acf9affbeac9c777976,
title = "Deep learning-based approach for landform classification from integrated data sources of digital elevation model and imagery",
abstract = "Landform classification is one of the most important aspects in geomorphological research, dividing the Earth'ssurface into diverse geomorphological types. Thus, an accurate classification of landforms is a key procedure indescribing the topographic characteristics of a given area and understanding their inner geomorphological formation processes. However, landform types are not always independent of one another due to the complexityand dynamics of interior and external forces. Furthermore, transitional landforms with gradually changing sur-face morphologies are widely distributed on the Earth's surface. With this situation, classifying these complexand transitional landforms with traditional landform classification methods is hard. In this study, a deep learning(DL) algorithm was introduced, aiming at automatically classifying complex and transitional landforms. Thisalgorithm was trained to learn and extract landform features from integrated data sources. These integrateddata sources contain different combinations of imagery, digital elevation models (DEMs), and terrain derivatives.The Loess Plateau in China, which contains complex and transitional loess landforms, was selected as the studyarea for data training. In addition, two sample areas in the Loess Plateau with complex and transitional loess hill and ridge landforms were used to validate the classified landform types by using the proposed DL method.Meanwhile, a comparative analysis between the proposed DL and random forest (RF) methods was also con-ducted to investigate their capabilities in landform classification. The proposed DL approach can achieve thehighest landform classification accuracy of 87{\%} in the transitional area with data combination of DEMs and im-ages. In addition, the proposed DL method can achieve a higher accuracy of landform classification with betterdefined landform boundaries compared to the RF method. The classified loess landforms indicate the differentlandform development stages in this area. Finally, the proposed DL method can be extended to other landformareas for classifying their complex and transitional landforms",
author = "Sijin Li and Liyang Xiong and Guoan Tang and Josef Strobl",
year = "2020",
doi = "https://doi.org/10.1016/j.geomorph.2020.107045",
language = "English",
volume = "354",
pages = "2--14",
journal = "Geomorphology",
issn = "0169-555X",
publisher = "Elsevier B.V.",

}

Deep learning-based approach for landform classification from integrated data sources of digital elevation model and imagery. / Li, Sijin; Xiong, Liyang; Tang, Guoan; Strobl, Josef.

in: Geomorphology, Jahrgang 354, 2020, S. 2-14.

Publikation: Beitrag in FachzeitschriftArtikelForschungBegutachtung

TY - JOUR

T1 - Deep learning-based approach for landform classification from integrated data sources of digital elevation model and imagery

AU - Li, Sijin

AU - Xiong, Liyang

AU - Tang, Guoan

AU - Strobl, Josef

PY - 2020

Y1 - 2020

N2 - Landform classification is one of the most important aspects in geomorphological research, dividing the Earth'ssurface into diverse geomorphological types. Thus, an accurate classification of landforms is a key procedure indescribing the topographic characteristics of a given area and understanding their inner geomorphological formation processes. However, landform types are not always independent of one another due to the complexityand dynamics of interior and external forces. Furthermore, transitional landforms with gradually changing sur-face morphologies are widely distributed on the Earth's surface. With this situation, classifying these complexand transitional landforms with traditional landform classification methods is hard. In this study, a deep learning(DL) algorithm was introduced, aiming at automatically classifying complex and transitional landforms. Thisalgorithm was trained to learn and extract landform features from integrated data sources. These integrateddata sources contain different combinations of imagery, digital elevation models (DEMs), and terrain derivatives.The Loess Plateau in China, which contains complex and transitional loess landforms, was selected as the studyarea for data training. In addition, two sample areas in the Loess Plateau with complex and transitional loess hill and ridge landforms were used to validate the classified landform types by using the proposed DL method.Meanwhile, a comparative analysis between the proposed DL and random forest (RF) methods was also con-ducted to investigate their capabilities in landform classification. The proposed DL approach can achieve thehighest landform classification accuracy of 87% in the transitional area with data combination of DEMs and im-ages. In addition, the proposed DL method can achieve a higher accuracy of landform classification with betterdefined landform boundaries compared to the RF method. The classified loess landforms indicate the differentlandform development stages in this area. Finally, the proposed DL method can be extended to other landformareas for classifying their complex and transitional landforms

AB - Landform classification is one of the most important aspects in geomorphological research, dividing the Earth'ssurface into diverse geomorphological types. Thus, an accurate classification of landforms is a key procedure indescribing the topographic characteristics of a given area and understanding their inner geomorphological formation processes. However, landform types are not always independent of one another due to the complexityand dynamics of interior and external forces. Furthermore, transitional landforms with gradually changing sur-face morphologies are widely distributed on the Earth's surface. With this situation, classifying these complexand transitional landforms with traditional landform classification methods is hard. In this study, a deep learning(DL) algorithm was introduced, aiming at automatically classifying complex and transitional landforms. Thisalgorithm was trained to learn and extract landform features from integrated data sources. These integrateddata sources contain different combinations of imagery, digital elevation models (DEMs), and terrain derivatives.The Loess Plateau in China, which contains complex and transitional loess landforms, was selected as the studyarea for data training. In addition, two sample areas in the Loess Plateau with complex and transitional loess hill and ridge landforms were used to validate the classified landform types by using the proposed DL method.Meanwhile, a comparative analysis between the proposed DL and random forest (RF) methods was also con-ducted to investigate their capabilities in landform classification. The proposed DL approach can achieve thehighest landform classification accuracy of 87% in the transitional area with data combination of DEMs and im-ages. In addition, the proposed DL method can achieve a higher accuracy of landform classification with betterdefined landform boundaries compared to the RF method. The classified loess landforms indicate the differentlandform development stages in this area. Finally, the proposed DL method can be extended to other landformareas for classifying their complex and transitional landforms

U2 - https://doi.org/10.1016/j.geomorph.2020.107045

DO - https://doi.org/10.1016/j.geomorph.2020.107045

M3 - Article

VL - 354

SP - 2

EP - 14

JO - Geomorphology

JF - Geomorphology

SN - 0169-555X

ER -