Detection of Gaseous Plumes using Basis Vectors

Area of Science:

• Remote Sensing
• Spectroscopy
• Data Analysis

Background:

• Detecting weak gaseous plumes via thermal imaging faces challenges.
• Current methods either rely on chemical spectral libraries (e.g., matched filter) or data-driven transformations (e.g., PCA, Fourier transform).
• Library-dependent methods risk missing uncatalogued chemicals, while data-driven methods may not fully exploit signal physics.

Purpose of the Study:

• To present a generalized least squares detection method for gas spectra.
• To introduce a novel detection method utilizing basis vectors.
• To compare the efficacy of these two methods using synthetic hyperspectral data.

Main Methods:

• Generalized least squares (GLS) detection applied to gas spectra.
• A new detection algorithm based on basis vectors.
• Application and comparison of both methods on synthetic hyperspectral datasets.

Main Results:

• The generalized least squares method was described for gas spectra detection.
• A novel basis vector detection method was presented.
• Detection images from both methods were compared on synthetic hyperspectral data, highlighting performance differences.

Conclusions:

• The study evaluates advanced methods for gaseous plume detection.
• Basis vector methods offer a promising alternative to traditional spectral library approaches.
• Further research can refine these techniques for enhanced gas detection capabilities.