Target spectral matching method

Abstract

The invention relates to a target spectral matching method. The target spectral matching method sequentially comprises the following steps: 1, acquiring a reference spectrum A={a1, a2, ..., an|ai>= 0, i epsilon 1, ...n} <T>, wherein ai is the spectral value of the reference spectrum A, n is the band number of the reference spectrum, and the band number of a test spectrum B is equal to that of the reference spectrum; 2, building a vector C={c1, c2, ..., cn|ci is not equal to 0, ci=c1, i epsilon 1, ...n} <T>, wherein ci is a real number constant; 3, building a linear equation system matrix M: M=|A C| or M=|C A|; 4, building a linear equation system: Mx=B; 5, solving a linear equation system: x=(M<T>M)<-1>M<T>B, wherein the solution corresponding to A is the content of the reference spectrum in the test spectrum B or the similarity degree of the test spectrum and the reference spectrum; the solution corresponding to C is the noise mean value and the system droop in the test spectrum. According to the invention, the calculation result is free from limitation on the value range, and the value is subject to linear variation according to the content of the reference spectrum in a target spectrum. The target spectral matching method has stronger anti-noise performance and the calculation result is free from influence of gaussian noise.

Description

technical field [0001] The invention relates to a target spectrum matching method, in particular to a target spectrum matching method with a higher value range and stronger anti-noise performance. Background technique [0002] Spectral matching refers to comparing a reference spectrum with a test (target) spectrum with the same number of bands and corresponding band wavelengths, and using a calculated value to represent the similarity or difference between the test spectrum and the reference spectrum. The larger the result value of the matching calculation, the greater the similarity between the test spectrum and the reference spectrum, and the smaller the difference. Spectral matching is often used to determine the size of the substance content corresponding to the reference spectrum contained in the target corresponding to the test spectrum. Commonly used spectral matching methods include spectral angle matching and spectral similarity matching. [0003] Spectral angle m...

AI Technical Summary

Problems solved by technology

The spectral similarity matching method ignores the fact that the spectral mean is an important feature of spectral data when calculating the spectral correlation coefficient, and the part whose value range is less than 0 has no practical significance from the perspective of determining the substance content

Images