FBRDLR: Fast blind reconstruction approach with dictionary learning regularization for infrared microscopy spectra

•A fast blind spectral reconstruction method is proposed for infrared spectrum.•Dictionary learning regularization is proposed to preserve spectral details.•Alternation minimization algorithm is described to solve the proposed model.•The robustness of the method is verified by some simulation and real experiments. Infrared (IR) spectra are the fingerprints of the molecules, and the spectral band location closely relates to the structure of a molecule. Thus, specimen identification can be performed based on IR spectroscopy. However, spectrally overlapping components prevent the specific identification of hyperfine molecular information of different substances. In this paper, we propose a fast blind reconstruction approach for IR spectra, which is based on sparse and redundant representations over a dictionary. The proposed method recovers the spectrum with the discrete wavelet transform dictionary on its content. The experimental results demonstrate that the proposed method is superior because of the better performance when compared with other state-of-the-art methods. The method the authors used remove the instrument aging issue to a large extent, thus leading the reconstruction IR spectra a more convenient tool for extracting features of an unknown material and interpreting it.