Mid-Wave Infrared Spectral Imaging for Gas Detection Based on Dictionary Learning and Sparse Reconstruction

Mid-wave infrared (MWIR) spectral imaging is a powerful technique for regional-scale gas detection, however, existing instruments suffer from bulky volume, long scanning time, and complex structures which impede their practical deployment. This necessitates the development of lightweight, high-efficiency MWIR spectral imaging techniques tailored to gas monitoring scenarios. To address these challenges, a snapshot-compatible MWIR computational spectral imaging method based on dictionary learning and sparse reconstruction is proposed. The broadband thin-film filters with high transmittance in the target band are employed for spectral compressed measurements, and the corresponding encoded images are captured by the detector. Subsequently, the orthogonal matching pursuit algorithm is utilized for spectral reconstruction, enabling the efficient recovery of the target spectral data cube. To ensure the accurate reconstruction ofMWIRgas absorption characteristics, a phased dictionary construction strategy is designed, comprising basic dictionary learning and MWIR feature enhancement. The simulation reconstruction results for blackbody radiation spectra, non-training smooth spectra, and absorption spectra of 4 typical gases verify the feasibility of the proposed method. Experiments are conducted based on sequential simulation of the proposed snapshot architecture. For real-world spectral reconstruction tests, the method achieves a peak signal-to-noise ratio (PSNR) greater than 26.336 dB and a spectral angle mapper (SAM) less than 0.157. These results confirm the effectiveness of the proposed method, which provides a feasible technical solution for the development of lightweight MWIR spectral imaging systems for gas detection applications.