Deep learning landslide extraction based on multi-temporal image difference local spatial information

Sheng Miao; Yu Qu; Xirong Liu

doi:10.1117/12.2681307

8 June 2023 Deep learning landslide extraction based on multi-temporal image difference local spatial information

Sheng Miao, Yu Qu, Xirong Liu

Proceedings Volume 12707, International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023); 127074B (2023) https://doi.org/10.1117/12.2681307
Event: International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023), 2023, Changsha, China

Abstract

Landslide as a natural disaster, it is important to obtain accurate information of landslide spatial distribution. The current method of landslide extraction is mainly based on the spectral features, spatial structure and texture features of landslides on remote sensing images, which is difficult to accurately extract the spatial distribution of landslides for surface structures with similar spectral features (bare soil, villages, exposed rocks, roads, water bodies, etc.) and does not make full use of local spatial information features and causes loss of landslide boundary information. To address the above problems, this paper proposes a method to extract landslides based on the fusion of local spatial information with multitemporal image differences. The high-resolution images before and after the landslide are used to calculate the difference images, extract each band and combine the temporal difference images to combine the bands and use them as input features for deep learning. The ASPP module is introduced to the U-net network model for expanding the local spatial information perception field, enhancing the preservation of landslide boundary information, and guiding the attention mechanism of feature down-sampling for extracting multi-scale contextual information. The evaluation shows that the method can ensure the detection error of landslide area range is less than 8.6%, the extraction accuracy is more than 95%, and the Kappa is 0.90. The experimental results prove the feasibility and accuracy of the method in rainfall landslide detection.

Citation Download Citation

Sheng Miao, Yu Qu, and Xirong Liu "Deep learning landslide extraction based on multi-temporal image difference local spatial information", Proc. SPIE 12707, International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023), 127074B (8 June 2023); https://doi.org/10.1117/12.2681307

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Network landslides

Feature extraction

RGB color model

Contour extraction

Deep learning

Remote sensing

Buildings

Show All Keywords

Keywords/Phrases

Search In:

Publication Years