Research on the Extraction of Characteristic Information and Similarity Evaluation Method for Relative Intensity Information between Characteristic Peaks in Raman Spectra

Automatic identification technology of Raman spectra is crucial for the accuracy and stability of qualitative Raman detection analysis and remains a technical challenge. The Relative Intensity Information between Characteristic Peaks (RIICP) in a substance's Raman spectrum is unique to the substance. This characteristic can significantly improve the identification accuracy and stability of Raman spectra. However, as this characteristic information represents an overall characteristic, it is highly susceptible to interference signals. Any disturbance to an individual peak will significantly change this characteristic. In this work, a quantization method of sequence array with gradient depth is proposed for the RIICP. This method can accurately extract and quantify this characteristic information while ensuring high suppression and tolerance capabilities to interference signals, thereby providing critical data for corrections in spectral identification. A similarity evaluation method based on Euclidean distance is designed for the quantized RIICP, which can accurately calculate the similarity of this dimensional feature. Finally, this paper utilizes the calculation data obtained from the proposed RIICP method to refine and adjust the identification results derived from the reverse matching method. Experimental results demonstrate that the proposed method has strong distinguishing capability, which can significantly improve the credibility and reliability of Raman spectrum identification results.