Assessing Scientific and Industry Grade SWIR Airborne Imaging Spectrometers for CH 4 Mapping

This study focus on the evaluation of scientific and industry-grade hyperspectral airborne sensors for the mapping of methane (CH4) emissions in the SWIR range. An imaging dataset from areas with known CH4 emissions was processed using the classic matched filter technique, and a new CH4 index. The airborne sensors were evaluated based on sensor design (spectral sampling and band centers), effectiveness of image processing, and impact of the signal-to-noise ratio (SNR) on CH4 mapping. The gas plume was mapped only in the images acquired with scientific-grade sensors. Results demonstrated that superior performance could be achieved when the position of band centers are closely located to the center of diagnostic CH4 absorption features. The impact of SNR was examined using a noise simulation, adding white noise to simulate images with varying SNR levels. Results indicate that the noisier signal of the industry-grade sensor is probably what prevented mapping the CH4 plume in this dataset. Simulations also demonstrated that as densest the plumes lower is the impact of SNR. Combined, results indicate that an imaging spectrometer with a scientific-grade SNR and band centers properly positioned to match the main CH4 features would improve the mapping of CH4 plumes with airborne sensors operating in the SWIR range.