Detecting and identifying spectral anomalies using wavelet processing

An automated method integrating wavelet processing and techniques from multivariate statistical process control (MSPC) is presented, providing a means for the simultaneous localization, detection, and identification of disturbances in spectral data. A defining property of the wavelet transform is its ability to map a one-dimensional chemical spectrum into a two-dimensional function of wavelength and scale. Therefore, unlike the traditional MSPC approach where disturbance detection is carried out in the original wavelength domain by using a single principal component analysis (PCA) model, detection employing wavelet transform processing results in the generation of multiple models within the wavelength-scale domain. Provided that the spectral disturbance can be localized within a subregion of the wavelength-scale domain through an advantageous choice of basis set, the method described allows the identification of the underlying disturbance. The utility of the proposed method in localizing, detecting, and identifying spectral disturbances is demonstrated by using real near-infrared measurements, suggesting its general applicability in spectroscopic monitoring of chemical processes.