Apart from shape and size, color is one of the most important characteristics in the discrimination and recognition of objects. The introduction of color information in optical pattern recognition is usually made via the multi-channel correlation technique, which decomposes the source and the target color images in three RGB channels. In this work we propose a new method for optical color pattern recognition based on the use of linear models that describe both the surface and the illuminant spectra, and we face the extension of correlation matched filter operations designed for pattern recognition. Different scenes were captured with a CCD camera and three correlation operations were used to test the model. The results show that the coefficient method derived from the linear description of images can discriminate polychromatic objects by optical correlation and leads to results that are almost independent of any spectral change in the illuminant. The discrimination capability of this method is clearly an improvement upon that obtained with the RGB multi-channel decomposition and is slightly better than other approaches used in optical correlation which are based on uniform color spaces. The method allows the use of more than three “color” components in optical pattern recognition, which can lead to better spectral surface description and accurate color object recognition.