A comparative study of spectral detectors

Spectral matching algorithms can be used for the identification of unknown spectra based on a measure of similarity with one or more known spectra. This paper presents a limited evaluation comparing different spectral matching algorithms utilized to identify signatures from a spectral library containing item of interest. We modify the maximum likelihood correlation coefficient estimate of a spectral change detection scheme so that it could perform as a spectral detector. We compare the results of this adaptation to other well-known spectral detection algorithms: spectral correlation mapper; adaptive matched filter; and adaptive coherence/cosine estimator. Our detector is shown to be an alternative to other correlation-based methods increasing detection performance while being considerable more computationally efficient than the covariance-based methods (O(n) versus O(n3)). As such, our detector is likely more viable in real-time applications.