Feature Wavelengths for Quantifying Methane Concentrations Using Shortwave Infrared Hyperspectral Imaging: A...

Area of Science:

• Environmental Science
• Spectroscopy
• Remote Sensing

Background:

• Methane (CH4) is a potent greenhouse gas.
• Accurate CH4 concentration monitoring is crucial for climate change mitigation.
• Hyperspectral imaging (HSI) offers potential for CH4 quantification.

Purpose of the Study:

• Identify optimal spectral regions in the short-wave infrared (SWIR) for CH4 quantification using HSI.
• Validate existing remote sensing wavelengths for CH4 detection.
• Compare the accuracy of single-band vs. multiband approaches.

Main Methods:

• Collected HSI data (1010-2495 nm) of CH4 at varying concentrations (0-2.5%) under controlled conditions.
• Employed Partial Least-Squares Regression (PLSR) on full spectral bands.
• Analyzed regression coefficients and PLS weights to pinpoint key spectral regions and wavelengths.
• Developed and tested new PLSR models using identified regions and specific wavelengths.

Main Results:

• Multiband spectral regions and two-band combinations yielded higher accuracy than single wavelengths.
• Optimal spectral regions for CH4 quantification were identified, including specific combinations like 1648 + 1670 nm.
• The ranking of effective spectral regions was: full 266-band > 1648 + 1670 nm > full 128-band > 2150-2243 nm > 1010-1185 nm.
• Investigated and noted potential overlaps with water vapor (H2O) absorption bands.

Conclusions:

• Multiband spectral analysis significantly enhances CH4 quantification accuracy via HSI.
• The identified SWIR spectral regions provide a basis for improved remote sensing of methane.
• Further validation in complex environmental settings is recommended to confirm optimal wavelengths.