Infrared Spectroscopy with Variable Decomposition Level Dual-Tree Complex Wavelet Transform for Quantification of Air Pollutants

Open-Path Fourier Transform Infrared Spectroscopy (OP-FTIR) is a key active remote sensing technology for detecting and identifying atmospheric pollutants. However, accurately characterizing gaseous pollutants under open-path conditions presents significant challenges, primarily due to the real-time estimation of background spectra. To address this issue, a variable decomposition level dual-tree complex wavelet transform method is proposed for adaptive background spectrum estimation. The algorithm dynamically adjusts the decomposition level based on spectral band characteristics and incorporates nonlinear compression with L1 regularization for enhanced stability under variable environmental conditions. The method was evaluated through both controlled gas cell experiments and field measurements. Compared with high-order Legendre polynomial fitting, it achieved lower average root-mean-square errors for C2H4 and C3H6 by 7.04 and 19.37%, respectively, indicating improved retrieval accuracy. In addition, it reduced the relative uncertainties for key background gases such as CO2 and CO across three distinct field sites, demonstrating enhanced robustness under variable environmental conditions. This approach provides a generalizable strategy for background estimation in complex open-path environments, supporting more accurate and scalable OP-FTIR-based gas monitoring.