Under this Phase I SBIR program, ORINCON proposes to develop and demonstrate digital signal processing (DSP) algorithms to provide substantial improvements in the identification and quantification of chemical compounds using Fourier Transform Infrared (FTIR) Spectroscopy. The algorithm we propose is conceptually founded on wavelet techniques. We will take advantage of the fact that the wavelet representation of the IR spectrum offers greater flexibility in enhancing low signal-to-noise ratios because the wavelet naturally resolves noise and spectral features into different domains. The algorithm itself is based on work described in Ref. [1], where an advanced wavelet technique is used to denoise astronomical spectra and remarkably recovers weak signal features that are drowned in noise so as not to be seen in the spectrum. In addition, the FTIR application of remote sensing is generally performedin a nonlinear environment due to dynamic background and instrumentation fluctuations. We propose the testing of wavelet pattern classification algorithms to aid in the identification of chemical compounds in this environment. In Phase II, algorithms, techniques, and procedures developed and demonstrated during Phase I will be optimized, implemented, and tested on a computer platform that runs ORINCON's commercial, real-time signal processing software.
Benefits: Potential commercialapplications for the algorithms developed under this program are numerous. The use of wavelet techniques in signal processing is becoming increasingly popular. Its inevitable insertion into the field of spectroscopy will have a profound effect on such fields as industrial chemical sensing, biological applications, medicine, and analytical chemistry.
