An Improved Linear Spectral Emissivity Constraint Method for Temperature and Emissivity Separation Using...

Area of Science:

• Remote Sensing
• Geophysics
• Thermodynamics

Background:

• Hyperspectral thermal infrared data analysis is crucial for understanding Earth's surface processes.
• Accurate separation of land surface temperature (LST) and land surface emissivity (LSE) is a persistent challenge.
• Existing methods like the linear spectral emissivity constraint (LSEC) rely on simplified emissivity models.

Purpose of the Study:

• To develop an improved method for LST and LSE separation.
• To enhance the accuracy of LST and LSE estimation using hyperspectral thermal infrared data.
• To introduce a novel approach that refines emissivity shape estimation.

Main Methods:

• Integration of a pre-estimate shape (PES) method with the LSEC algorithm.
• Development of a new piecewise scheme for LST and LSE separation, termed PES-LSEC.
• Validation using simulated datasets and in situ experimental data.

Main Results:

• The PES-LSEC method demonstrated superior accuracy in both LST and LSE estimation compared to the LSEC method.
• Simulated data showed low root-mean-square errors (RMSEs) for LST (0.07 K) and LSE (0.0045) with minimal at-ground error.
• In situ data validation yielded an LST RMSE of 0.9 K and an average LSE accuracy of 0.01.

Conclusions:

• The PES-LSEC method significantly improves LST and LSE separation accuracy.
• This method effectively preserves essential emissivity spectral features (crest and trough information).
• PES-LSEC offers a more robust approach for analyzing hyperspectral thermal infrared data.