Landslide Susceptibility Mapping Considering Landslide Local-Global Features Based on CNN and Transformer

被引:16
作者
Zhao, Zeyang [1 ]
Chen, Tao [1 ,2 ,3 ,4 ]
Dou, Jie [2 ]
Liu, Gang [5 ]
Plaza, Antonio [6 ]
机构
[1] China Univ Geosci, Sch Geophys & Geomat, Wuhan 430074, Peoples R China
[2] China Univ Geosci, Badong Natl Observat & Res Stn Geohazards, Wuhan 430074, Peoples R China
[3] Minist Nat Resources, Key Lab Natl Geog Census & Monitoring, Wuhan 430079, Peoples R China
[4] Chengdu Univ Technol, State Key Lab Geohazard Prevent & Geoenvironm Prot, Chengdu 610059, Peoples R China
[5] Chengdu Univ Technol, Coll Earth Sci, Chengdu 610059, Peoples R China
[6] Univ Extremadura, Escuela Politecn, Dept Technol Comp & Commun, Hyperspectral Comp Lab, Caceres 10071, Spain
基金
中国国家自然科学基金;
关键词
Terrain factors; Feature extraction; Transformers; Geology; Convolutional neural networks; Task analysis; Surfaces; Convolutional neural network (CNN); landslide local-global features; landslide susceptibility mapping (LSM); transformer; RANDOM FOREST;
D O I
10.1109/JSTARS.2024.3379350
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Landslide susceptibility mapping (LSM) is a crucial step in quantitatively assessing landslide risk, essential for geologic hazards prevention. With the rapid development of deep learning models, convolutional neural networks (CNNs), and transformer architectures have been applied to LSM. However, these models still face the challenges of suboptimal mapping accuracy and limited capacity for multilevel landslide features extraction. In this study, we present a CNN-transformer local-global features extraction network (CTLGNet) that combines the strengths of both CNN and transformer models to effectively extract both landslide local and global features. We apply this model to LSM in two regions: the Three Gorges Reservoir area and Jiuzhaigou. To begin, nine landslide conditioning factors are selected and analyzed to construct the landslide dataset for LSM. Subsequently, the dataset is randomly split into training, validation, and test datasets in a 6:2:2 ratio to attain LSM results. Then, CTLGNet is compared to CNN, residual neural network, densely connected convolutional network, vision transformer, and fractional Fourier image transformer using various evaluation metrics. The results demonstrate that CTLGNet exhibits exceptional landslide prediction and generalization capabilities, outperforming the other five models across all evaluation metrics except Recall, with AUC values of 0.9817 and 0.9693 for the two regions, respectively. The LSM results indicate that CTLGNet can effectively extract both landslide local and global features to achieve landslide localization and detail capture. Overall, our proposed framework excels in extracting multilevel landslide features and holds great potential for widespread application.
引用
收藏
页码:7475 / 7489
页数:15
相关论文
共 50 条
[41]   Effects of non-landslide sampling strategies on machine learning models in landslide susceptibility mapping [J].
Gu, Tengfei ;
Duan, Ping ;
Wang, Mingguo ;
Li, Jia ;
Zhang, Yanke .
SCIENTIFIC REPORTS, 2024, 14 (01)
[42]   High-resolution landslide mapping and susceptibility assessment: Landslide temporal variations and vegetation recovery [J].
Ali, Muhammad Zeeshan ;
Chen, Kejie ;
Shafique, Muhammad ;
Adnan, Muhammad ;
Zheng, Zhiwen ;
Zhang, Wei ;
Qing, Zhanhui .
ADVANCES IN SPACE RESEARCH, 2024, 74 (08) :3668-3690
[43]   Mapping landslide susceptibility and types using Random Forest [J].
Taalab, Khaled ;
Cheng, Tao ;
Zhang, Yang .
BIG EARTH DATA, 2018, 2 (02) :159-178
[44]   Advanced data mining techniques for landslide susceptibility mapping [J].
Ibrahim, Muhammad Bello ;
Mustaffa, Zahiraniza ;
Balogun, Abdul-Lateef ;
Hamonangan Harahap, Indra Sati ;
Ali Khan, Mudassir .
GEOMATICS NATURAL HAZARDS & RISK, 2021, 12 (01) :2430-2461
[45]   Landslide Susceptibility Mapping using Machine Learning Algorithm [J].
Hussain, Muhammad Afaq ;
Chen, Zhanlong ;
Wang, Run ;
Shah, Safeer Ullah ;
Shoaib, Muhammad ;
Ali, Nafees ;
Xu, Daozhu ;
Ma, Chao .
CIVIL ENGINEERING JOURNAL-TEHRAN, 2022, 8 (02) :209-224
[46]   Improving Spatial Agreement in Machine Learning-Based Landslide Susceptibility Mapping [J].
Adnan, Mohammed Sarfaraz Gani ;
Rahman, Md Salman ;
Ahmed, Nahian ;
Ahmed, Bayes ;
Rabbi, Md. Fazleh ;
Rahman, Rashedur M. .
REMOTE SENSING, 2020, 12 (20) :1-23
[47]   An integrated approach based landslide susceptibility mapping: case of Muzaffarabad region, Pakistan [J].
ul Basharat, Mubeen ;
Khan, Junaid Ali ;
Abdo, Hazem Ghassan ;
Almohamad, Hussein .
GEOMATICS NATURAL HAZARDS & RISK, 2023, 14 (01)
[48]   Study on landslide susceptibility mapping based on rock-soil characteristic factors [J].
Yu, Xianyu ;
Zhang, Kaixiang ;
Song, Yingxu ;
Jiang, Weiwei ;
Zhou, Jianguo .
SCIENTIFIC REPORTS, 2021, 11 (01)
[49]   Application of environmental variables in statistically-based landslide susceptibility mapping: A review [J].
Zhao, Xin ;
Zhao, Zhifang ;
Huang, Faming ;
Huang, Jiangcheng ;
Yang, Zhiquan ;
Chen, Qi ;
Zhou, Dingyi ;
Fang, Liuyang ;
Ye, Xian ;
Chao, Jiangqin .
FRONTIERS IN EARTH SCIENCE, 2023, 11
[50]   Integration of Analytical Hierarchy Process and Landslide Susceptibility Index Based Landslide Susceptibility Assessment of the Pearl River Delta Area, China [J].
Zhang, Haoran ;
Zhang, Guifang ;
Jia, Qiwen .
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2019, 12 (11) :4239-4251