River shoreline extraction and bank collapse recognition based on Chinese GF satellite

Extracting river information from high-resolution satellite remote sensing images is of great significance for monitoring and warning of river bank collapses. This article takes the example of the bank collapse with a length and depth of about 100m in Xiaopan of Jiayu County, Hubei Province in Decem...

Full description

Saved in:
Bibliographic Details
Published in2024 5th International Conference on Geology, Mapping and Remote Sensing (ICGMRS) pp. 182 - 186
Main Authors Chen, Kebing, Sun, Sirui
Format Conference Proceeding
LanguageEnglish
Published IEEE 12.04.2024
Subjects
Accuracy
Data mining
GF satellite
Image edge detection
Image resolution
Landslides
remote sensing monitoring
river bank collapse
Rivers
Satellites
semantic segmentation
Online AccessGet full text

Cover

More Information
Summary:Extracting river information from high-resolution satellite remote sensing images is of great significance for monitoring and warning of river bank collapses. This article takes the example of the bank collapse with a length and depth of about 100m in Xiaopan of Jiayu County, Hubei Province in December 2021, and establishes a water extraction model based on DeepLabV3+image semantic segmentation technology. The application ability of domestic GF No.1 and No.2 satellites in identifying river bank collapses is discussed. The research results indicate that the DeepLabv3+model can extract water edge information well and the extraction results are reliable. By comparing long-term image images, it can reflect the situation of bank collapse on both sides of the river. The resolution of domestically produced Gaofen-1 and Gaofen-2 satellites is sufficient for accurate identification of larger scale landslides, and further exploration is necessary to enhance their ability to identify smaller scale landslides.
DOI:10.1109/ICGMRS62107.2024.10581359