Metal sorting system using laser induced breakdown spectroscopy and operating method thereof

Patent metadata

Field	Value
Publication number	US-11543357-B2
Application number	US-201916534824-A
Country	US
Kind code	B2
Filing date	Aug 7, 2019
Priority date	Aug 8, 2018
Publication date	Jan 3, 2023
Grant date	Jan 3, 2023

Abstract

Disclosed is an operating method of a metal sorting system using laser induced breakdown spectroscopy (LIBS), which may include: analyzing a metal component distribution for various metals using LIBS library information; setting multiple clusters according to the metal component distribution; performing first regression component analysis with respect to spectral data of a metal sample; calculating a probability that the spectral data will belong to each of the set multiple clusters using the first regress component analysis result; performing second regression component analysis with respect to the spectral data which belong to each cluster; and discriminating a type of metal sample by a weighted sum of the calculated probability and the second regression component analysis result.

First claim

What is claimed is:

1. An operating method of a metal sorting system using laser induced breakdown spectroscopy (LIBS), wherein the metal sorting system comprises a signal processing device, the method comprising:
analyzing, by the signal processing device, a metal component distribution for various metals using LIBS library information;
setting, by the signal processing device, multiple clusters according to the metal component distribution;
performing, by the signal processing device, first regression component analysis using first training data with respect to spectral data of a metal sample;
calculating, by the signal processing device, a probability that the spectral data will belong to each of the set multiple clusters using the first regress component analysis result;
performing, by the signal processing device, second regression component analysis using second training data which belongs to each cluster with respect to the spectral data of the metal sample; and
discriminating, by the signal processing device, a type of metal sample by a weighted sum of the calculated probability and the second regression component analysis result,
wherein the metal sorting system further comprises a LIBS device outputting the spectral data of the metal sample by irradiating the metal sample with a laser, and a discharge device discharging metals to different collection boxes according to the discriminated metal type.

2. The method of claim 1 , wherein the LIB S library information includes prior information including at least one of a temperature, humidity, fine dust concentration, experimental variables, and existing measured data.

3. The method of claim 1 , wherein the metal component distribution includes information on a correlation between the LIBS library information and the metal.

4. The method of claim 1 , wherein the setting of the multiple clusters includes
classifying the clusters by considering a main component, and
classifying the clusters by considering a linear relationship between a component having a first concentration and the main component, and
the first concentration is lower than a concentration of the main component.

5. The method of claim 4 , wherein the setting of the multiple clusters further includes classifying the clusters by considering the linear relationship between a component having the second concentration and the main component, and
the second concentration is lower than the first concentration.

6. The method of claim 1 , wherein the calculating of the probability further includes performing soft sorting for the metal sample according to the calculated probability.

7. The method of claim 1 , wherein the calculating of the probability includes calculating a probability that the spectral data will belong to each cluster using a Bayesian rule.

8. The method of claim 1 , wherein the performing of the first regression component analysis includes estimating an element concentration of unknown metal data by training the spectral data using the first training data, and
the first training data is all training data.

9. The method of claim 8 , wherein the performing of the second regression component analysis includes estimating the element concentration of the unknown metal data by training the spectral data using the second training data.

10. The method of claim 9 , wherein the discriminating the type of metal sample includes
calculating a final regression analysis result by the weighted sum of the second regression component analysis result and the calculated probability,
estimating at least one metal concentration value depending on the final regression analysis result, and
discriminating the type of metal sample using the LIBS library information and the estimated concentration value.

11. A metal sorting system using laser induced breakdown spectroscopy (LIBS), the system comprising:
an LIBS device outputting spectral data of a metal sample by irradiating the metal sample with a laser;
a signal processing device discriminating a type of metal sample using the spectral data and LIBS library information; and
a discharge device discharging metals to different collection boxes according to the discriminated metal type,
wherein the signal processing device sets multiple clusters according to the LIBS library information and a metal component distribution, calculates a probability that the spectral data will belong to each of the multiple clusters using first regress component analysis using first training data for the spectral data, performs second regression component analysis using second training data which belongs to each cluster with respect to the spectral data, and discriminates the type of metal sample using the calculated probability and the second regression component analysis result.

12. The system of claim 11 , wherein the first regression component analysis is full-scale regression component analysis for the metal sample.

13. The system of claim 12 , wherein the signal processing device estimates an element concentration which is a final regression component analysis result by a weighted sum of the calculated probability and the second regression component analysis result and discriminates the type of metal sample using the LIBS database and the estimated element concentration.

Assignees

• Gwangju Inst Science & Tech

Inventors

• Hwang Eui Seok
• Kim E Den
• Jeong Sung Ho
• Shin Sung Ho

Classifications

• G01N2201/1053

  System of scan mirrors for composite motion of beam · CPC title

• G01N33/2028

  Metallic constituents · CPC title

• G01J3/443

  Emission spectrometry · CPC title

• G01N21/718 Primary

  Laser microanalysis, i.e. with formation of sample plasma · CPC title

• B07C5/34 Primary

  Sorting according to other particular properties {(material testing per se G01N; quality control G07C3/14)} · CPC title