This paper presents the results of identification of a license plate at daytime and dusk. We introduce the novel technique for processing multispectral images within geoinformation systems utilizing the advanced spectral analysis. This innovative method significantly enhances the accuracy of license plate recognition across varying lighting conditions. Comparison of the laboratory and field experiments has been carried out to validate our approach. Research indicates that utilizing multi-band data effectively aids in uncovering significant features that are often overlooked when processing standard RGB images. Overall, this technique has the potential to greatly enhance the performance of video surveillance systems, contribute to improving road safety, prevent potential infractions, and significantly improve traffic management efficiency.