Very high-resolution satellite-derived bathymetry using panchromatic and multispectral image fusion

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Oceanography

Background:

• Satellite-derived bathymetry offers high spatial resolution for shallow waters.
• High costs of very high-resolution (VHR) imagery restrict VHR bathymetric data acquisition.
• Panchromatic and multispectral image fusion (pan-sharpening) can enhance image resolution.

Purpose of the Study:

• To develop and evaluate a novel VHR bathymetric mapping approach using pan-sharpening.
• To assess the effectiveness of the random forest algorithm incorporated with pan-sharpening (RF-P) for VHR bathymetry.
• To provide a cost-effective and precise method for VHR bathymetric mapping in optically shallow waters.

Main Methods:

• Developed a new VHR bathymetric mapping approach: random forest algorithm incorporated with pan-sharpening (RF-P).
• Utilized multispectral and panchromatic images from the GF-2 satellite and ICESat-2 elevation data.
• Applied the RF-P method for VHR bathymetric mapping in Ganquan Island.

Main Results:

• The RF-P method demonstrated superior performance in VHR bathymetric mapping.
• Achieved a mean absolute error (MAE) of 1.18 m and a root mean square error (RMSE) of 1.40 m.
• RF-P reduced MAE by 0.48 m and RMSE by 0.60 m compared to the log-ratio model.

Conclusions:

• The RF-P approach is an effective method for VHR bathymetric mapping.
• This technique offers a cost-effective, rapid, and precise solution for shallow water bathymetry.
• The study highlights the potential of image fusion for improving VHR bathymetric data acquisition.