CN-121982460-A - Method and device for global matching of mineral spectra based on multi-band feature vector

CN121982460ACN 121982460 ACN121982460 ACN 121982460ACN-121982460-A

Abstract

The application provides a method and a device for global matching of mineral spectra based on a multi-band feature vector. The method comprises the steps of continuously removing an input spectrum to obtain a target spectrum, determining an effective band and a central wavelength of the target spectrum based on a preset threshold strategy, determining absorption characteristic parameters corresponding to the effective band, constructing and determining to obtain a first target comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band, determining a second target comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of a preset reference spectrum, determining basic Euclidean distance and augmented Euclidean distance of the first target comprehensive absorption characteristic vector and the second target comprehensive absorption characteristic vector, and determining minerals corresponding to the reference spectrum with the minimum basic Euclidean distance or augmented Euclidean distance as target minerals. By the method, the multi-band feature vectors are fused, and mineral identification is performed based on the basic Euclidean distance or the extended Euclidean distance, so that the accuracy of mineral identification can be improved.

Inventors

Bu Jiancai
Xiao Songchun
XU ZHIBIN
XIAO YUNYUN
GAO YONG
ZHANG LIJUN
CHEN YONG
LUO WENXIANG

Assignees

湖南省遥感地质调查监测所

Dates

Publication Date: 20260505
Application Date: 20251219

Claims (10)

1. A method of global matching of mineral spectra based on a multi-band feature vector, the method comprising: Acquiring an input spectrum, and performing continuous system removal processing on the input spectrum to obtain a target spectrum; Determining an effective band of the target spectrum and a center wavelength of the effective band based on a preset threshold strategy; determining absorption characteristic parameters corresponding to the effective band of the target spectrum, wherein the absorption characteristic parameters comprise absorption band center wavelength, absorption depth, absorption half-width, absorption area, absorption symmetry and absorption slope; Constructing a first initial comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the target spectrum; Performing standardization processing on the first initial comprehensive absorption feature vector to obtain a first target comprehensive absorption feature vector; Determining a second target comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the preset reference spectrum; determining a basic Euclidean distance and an augmented Euclidean distance of the first target comprehensive absorption feature vector and the second target comprehensive absorption feature vector; and determining the mineral corresponding to the reference spectrum with the minimum basic Euclidean distance or the minimum augmented Euclidean distance as the target mineral.
2. The method of claim 1, wherein said constructing a first initial integrated absorption feature vector from absorption feature parameters of an effective spectral band-stop of the target spectrum comprises: sequentially extracting the absorption depth, the absorption half-width, the absorption symmetry, the absorption slope and the absorption area of each effective band based on the sequence from small to large of the central wavelength of each effective band of the target spectrum, and constructing a five-dimensional feature vector; and splicing the five-dimensional feature vectors of each effective band to obtain a first initial comprehensive absorption feature vector.
3. The method of claim 1, wherein determining the second target integrated absorption feature vector from the absorption feature parameters of the effective spectral band of the predetermined reference spectrum comprises: Acquiring an effective band of a predicted reference spectrum, and a center wavelength of the effective band of the reference spectrum; according to a preset wavelength tolerance threshold, the center wavelength of the effective band of the preset reference spectrum is matched with the center wavelength of the effective band of the target spectrum; Determining an absorption characteristic parameter of a matched absorption band in response to the existence of the matched absorption band of the target spectrum and the preset reference spectrum; constructing a second initial comprehensive absorption characteristic vector according to the absorption characteristic parameters of the matched absorption band; and carrying out standardization processing on the second initial comprehensive absorption feature vector to obtain a second target comprehensive absorption feature vector.
4. A method according to claim 3, wherein said determining the second target integrated absorption feature vector from the absorption feature parameters of the effective band of the preset reference spectrum further comprises: And in response to the absence of a matching absorption band between the target spectrum and the preset reference spectrum, determining a pseudo-absorption characteristic parameter of the preset reference spectrum according to the central wavelength of the effective band of the target spectrum.
5. The method of claim 1, wherein the determining the base euclidean distance and the augmented euclidean distance for the first target synthetic absorption feature vector and the second target synthetic absorption feature vector comprises: determining a basic Euclidean distance between the target spectrum and the preset reference spectrum according to the first target comprehensive absorption feature vector and the second target comprehensive absorption feature vector; determining the ratio of the number of the matched absorption bands to the total number of absorption bands of the target spectrum as a matching rate; And determining the extended Euclidean distance between the target spectrum and the preset reference spectrum according to the first target comprehensive absorption feature vector, the second target comprehensive absorption feature vector and the matching rate.
6. The method of claim 5, wherein the basic euclidean distance satisfies: Wherein, the Represents the basic Euclidean distance between the ith target spectrum and the jth preset reference spectrum, p represents the total dimension number of the absorption characteristic parameter, Characteristic parameter values representing the kth dimension in the integrated absorption characteristic vector of the ith target spectrum, And the characteristic parameter value of the kth dimension in the comprehensive absorption characteristic vector of the jth preset reference spectrum is represented.
7. The method of claim 5, wherein the augmented euclidean distance satisfies: Wherein, the Represents the augmented Euclidean distance between the ith target spectrum and the jth preset reference spectrum, p represents the total dimension number of the absorption characteristic parameter, Characteristic parameter values representing the kth dimension in the integrated absorption characteristic vector of the ith target spectrum, Characteristic parameter values representing the kth dimension in the integrated absorption characteristic vector of the jth preset reference spectrum, The matching rate of the absorption band between the ith target spectrum and the jth preset reference spectrum.
8. An apparatus for global matching of mineral spectra based on a multi-band eigenvector, the apparatus comprising: The acquisition module is used for acquiring an input spectrum and carrying out continuous system removal processing on the input spectrum to obtain a target spectrum; The first determining module is used for determining an effective band of the target spectrum and the center wavelength of the effective band based on a preset threshold strategy; the second determining module is used for determining absorption characteristic parameters corresponding to the effective band of the target spectrum, wherein the absorption characteristic parameters comprise absorption band center wavelength, absorption depth, absorption half-width, absorption area, absorption symmetry and absorption slope; The construction module is used for constructing a first initial comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the target spectrum; The processing module is used for carrying out standardization processing on the first initial comprehensive absorption characteristic vector to obtain a first target comprehensive absorption characteristic vector; the third determining module is used for determining a second target comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the preset reference spectrum; A fourth determining module, configured to determine a basic euclidean distance and an extended euclidean distance of the first target comprehensive absorbing feature vector and the second target comprehensive absorbing feature vector; and a fifth determining module, configured to determine, as a target mineral, a mineral corresponding to a reference spectrum with a minimum basic euclidean distance or an extended euclidean distance.
9. An electronic device, comprising: a memory configured to store instructions, and A processor configured to recall from the memory the instructions and when executed enable the method of multi-band feature vector based mineral spectrum global matching according to any one of claims 1 to 7.
10. A machine-readable storage medium having instructions stored thereon for causing a machine to perform the method of multi-band feature vector based global matching of mineral spectra according to any of claims 1 to 7.

Description

Method and device for global matching of mineral spectra based on multi-band feature vector Technical Field The invention relates to the technical field of mineral identification, in particular to a method and a device for global matching of mineral spectra based on multi-band eigenvectors. Background The traditional rock and ore identification and geological mapping work mainly depends on manual field investigation and indoor rock and ore identification, and has the advantages of complex working procedure, large workload, long period, and easily influenced by terrain and environmental conditions, and cannot meet the requirements of high efficiency and intellectualization of geological work in new era. In recent years, with the rapid development of detection technology, detection instruments are continuously changed, and advanced test technologies such as X-ray fluorescence spectrum, X-ray diffraction, infrared spectrum, raman spectrum, electron microscope scanning and the like are increasingly widely applied to rock and mineral identification work, so that the accuracy of rock and mineral identification is greatly improved, the scale range of rock and mineral identification is widened, but the methods are generally complex in equipment, high in use cost and high in operation technology difficulty, are limited to sample tests in laboratories, are difficult to directly apply to field investigation work, and are greatly limited in application in the field of geography. Disclosure of Invention The application aims to provide a method and a device for global matching of mineral spectra based on multi-band feature vectors, so that the accuracy of mineral identification is improved. In a first aspect, there is provided a method of global matching of mineral spectra based on a multi-band feature vector, the method comprising: Acquiring an input spectrum, and performing continuous system removal processing on the input spectrum to obtain a target spectrum; Determining an effective band of the target spectrum and a center wavelength of the effective band based on a preset threshold strategy; determining absorption characteristic parameters corresponding to the effective band of the target spectrum, wherein the absorption characteristic parameters comprise absorption band center wavelength, absorption depth, absorption half-width, absorption area, absorption symmetry and absorption slope; Constructing a first initial comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the target spectrum; Performing standardization processing on the first initial comprehensive absorption feature vector to obtain a first target comprehensive absorption feature vector; Determining a second target comprehensive absorption characteristic vector according to the absorption characteristic parameters of the effective band of the preset reference spectrum; determining a basic Euclidean distance and an augmented Euclidean distance of the first target comprehensive absorption feature vector and the second target comprehensive absorption feature vector; and determining the mineral corresponding to the reference spectrum with the minimum basic Euclidean distance or the minimum augmented Euclidean distance as the target mineral. Optionally, the constructing a first initial integrated absorption feature vector according to the absorption feature parameters of the effective spectrum band of the target spectrum includes: sequentially extracting the absorption depth, the absorption half-width, the absorption symmetry, the absorption slope and the absorption area of each effective band based on the sequence from small to large of the central wavelength of each effective band of the target spectrum, and constructing a five-dimensional feature vector; and splicing the five-dimensional feature vectors of each effective band to obtain a first initial comprehensive absorption feature vector. Optionally, the determining the second target comprehensive absorption feature vector according to the absorption feature parameter of the effective band of the preset reference spectrum includes: Acquiring an effective band of a predicted reference spectrum, and a center wavelength of the effective band of the reference spectrum; according to a preset wavelength tolerance threshold, the center wavelength of the effective band of the preset reference spectrum is matched with the center wavelength of the effective band of the target spectrum; Determining an absorption characteristic parameter of a matched absorption band in response to the existence of the matched absorption band of the target spectrum and the preset reference spectrum; constructing a second initial comprehensive absorption characteristic vector according to the absorption characteristic parameters of the matched absorption band; and carrying out standardization processing on the second initial comprehensive absorption feature vec