Innovative Polarimetric Interferometric Synthetic Aperture Radar Land Cover Classification: Integrating Power,...

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Earth Observation

Background:

• Traditional land cover classification methods inadequately utilize multi-source SAR data.
• Limited classification types and low accuracy are persistent challenges in current methods.

Purpose of the Study:

• To develop an improved PolInSAR land cover classification method.
• To enhance classification accuracy and detail by fusing power, polarimetric, and interferometric information.

Main Methods:

• Fusing power, polarimetric, and interferometric information from PolInSAR data.
• Differentiating land cover types by analyzing scattering characteristics and elevation.
• Subdividing vegetation into detailed categories based on height and type using anisotropy parameters.

Main Results:

• Achieved an overall classification accuracy of 90.2% with a Kappa coefficient of 0.876.
• Successfully distinguished between strong and weak scattering areas, shadows, water bodies, vegetation, buildings, and bare soil.
• Further refined vegetation classification into four detailed sub-categories.

Conclusions:

• The proposed PolInSAR data fusion method significantly improves land cover classification accuracy and detail.
• This approach offers a more comprehensive understanding of land cover composition compared to traditional methods.