An automatic identification method, device and equipment for amber producing areas and a storage medium

Abstract

The application discloses an automatic identification method, device and equipment of amber origin and a storage medium; the method forms an infrared spectrum data set by acquiring infrared spectrum data corresponding to amber samples of multiple amber origins; the infrared spectrum data of a to-be-tested amber sample is spliced into the infrared spectrum data set to obtain a spliced amber data set, data processing is performed on the spliced amber data set, and central values of the infrared spectrum data corresponding to the multiple amber origins are obtained; a first predicted origin of the to-be-tested amber sample is acquired by calculating and based on a first type distance corresponding to each infrared spectrum data central value, a second type distance between the to-be-tested infrared spectrum data and infrared spectrum data of each sample is calculated, and origin attribute classification data of the to-be-tested amber sample is obtained; and the origin of the amber is obtained according to the first predicted origin and the origin attribute classification data. The application can automatically identify the amber samples of different amber origins according to infrared spectrum characteristics, and the identification result is high in accuracy.

Description

Technical Field

[0001] This invention relates to the technical field of amber detection, and in particular to an automatic method, apparatus, equipment, and storage medium for identifying the origin of amber. Background Technology

[0002] Amber has numerous origins, with over a hundred known worldwide. Currently, the most common amber varieties on the market mainly come from three major producing regions: Europe, the Americas, and Asia. The five most famous producing areas are the Baltic coast, the Dominican Republic, Mexico, Myanmar, and Fushun, China. The origin of amber is a crucial factor influencing its value in the market. Accurately identifying and tracing the origin of amber using scientific methods is vital for maintaining consumer confidence and ensuring the orderly development of the amber market.

[0003] Currently, the market primarily determines the origin of amber based on its appearance, such as body color, transparency, and fluorescence. This method is highly subjective and heavily influenced by experience. In the laboratory, amber origin identification mainly relies on methods like spectral analysis and chemical composition analysis to differentiate the characteristics of amber from different regions. Existing spectral identification methods are mostly designed to distinguish amber from the three major producing areas, judging by the shape and position of absorption peaks in specific wavelengths. However, these methods struggle to differentiate amber from the six major producing areas, and even more so to distinguish amber from different sub-regions within the same producing area.

[0004] Existing methods for differentiating amber origins using ultraviolet-visible spectroscopy, Raman spectroscopy, and fluorescence spectroscopy, followed by human judgment, have some merit. However, human observation is highly subjective and difficult to quantify, thus failing to meet the current needs of jewelry testing laboratories for information technology and digitalization. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an automatic identification method, device, equipment and storage medium for amber origin, which identifies amber samples from different amber origins based on infrared spectral characteristics, thereby improving the accuracy of identification.

[0006] To address the aforementioned technical problems, this invention provides an automatic identification method for the origin of amber, comprising:

[0007] Amber samples from multiple amber-producing regions were selected to generate a standard sample set, and infrared spectral data of each amber sample in the standard sample set were collected.

[0008] Based on the multiple amber origins, corresponding data tags were set for the infrared spectral data of each sample, and the infrared spectral data of all samples were integrated to obtain an infrared spectral dataset.

[0009] The infrared spectral data of the amber sample to be tested are obtained, and the infrared spectral data is stitched together into the infrared spectral dataset to obtain the stitched amber dataset.

[0010] The spliced ​​amber dataset is processed to obtain a standard dataset, and the standard dataset is dimensionality reduced to obtain a dimensionality reduction matrix. Based on the data labels, the dimensionality reduction matrix is ​​classified to obtain the infrared spectral data center values ​​corresponding to the multiple amber origins.

[0011] Based on the dimensionality reduction matrix and the infrared spectral data center value, calculate the first type distance corresponding to each infrared spectral data center value, obtain the minimum value of the first type distance, and take the amber origin corresponding to the minimum value as the first predicted origin of the amber sample to be tested.

[0012] Calculate the second type distance between the infrared spectral data to be tested and the infrared spectral data of each sample in the standard sample set, and based on the second type distance, combine the KNN nonparametric classification supervision algorithm to obtain the origin attribute data corresponding to each amber sample in the standard sample set, and obtain the origin attribute classification data of the amber sample to be tested.

[0013] Based on the first predicted origin and the origin attribute classification data, the origin of the amber sample to be tested is determined, and the origin of the amber sample to be tested is obtained.

[0014] In one possible implementation, the spliced ​​amber dataset is processed to obtain a standard dataset, specifically including:

[0015] Based on a preset formula for calculating average spectral data, the average spectral data of infrared spectral data corresponding to all amber origins in the spliced ​​amber dataset is calculated, and the average spectral data is used as the standard spectrum of amber.

[0016] Linear regression processing is performed on the standard amber spectrum and the spliced ​​amber dataset to obtain the intercept and linear regression coefficients, and the offset of each amber infrared spectral data is obtained based on the intercept;

[0017] Based on the linear regression coefficients and the offset, multivariate scattering correction is performed on each infrared spectral data in the spliced ​​amber dataset to obtain corrected infrared spectral data.

[0018] Differential operations are performed on the corrected infrared spectral data to obtain standard infrared spectral data. All the standard infrared spectral data are then integrated to obtain a standard dataset.

[0019] In one possible implementation, the standard dataset is subjected to dimensionality reduction processing to obtain a dimensionality reduction matrix, specifically including:

[0020] Obtain the set of infrared spectral data wavenumber intensity corresponding to each amber origin in the standard dataset, and calculate the average wavenumber intensity of each set of infrared spectral data wavenumber intensity.

[0021] Subtracting the average wavenumber intensity from each wavenumber intensity value in the infrared spectral data wavenumber intensity set yields the first infrared spectral data wavenumber intensity set.

[0022] Simultaneously, the first covariance matrix of the first infrared spectral data wavenumber intensity set is obtained, and the eigenvalues ​​and eigenvectors of the first covariance matrix are calculated.

[0023] Based on the eigenvalues ​​and eigenvectors, the first infrared spectral data wavenumber intensity set is subjected to dimensionality reduction processing to obtain a dimensionality reduction matrix.

[0024] In one possible implementation, the first type of distance corresponding to each infrared spectral data center value is calculated based on the dimensionality reduction matrix and the infrared spectral data center value, specifically including:

[0025] Obtain the second covariance matrix of the reduced-dimensional matrix, and calculate the inverse of the second covariance matrix;

[0026] Based on the dimensionality reduction matrix, the infrared spectral data center value, and the inverse matrix, calculate the first type distance corresponding to each infrared spectral data center value.

[0027] In one possible implementation, based on the second type of distance, the origin attribute data corresponding to each amber sample in the standard sample set is obtained by combining the KNN nonparametric classification supervision algorithm, thus obtaining the origin attribute classification data of the amber sample to be tested, specifically including:

[0028] Infrared spectral data from different amber origins are divided into bands to obtain multiple band segments, and the second type distances corresponding to the multiple band segments are calculated.

[0029] Obtain all second-class distances, sort all second-class distances from largest to smallest, and select the multiple second-class distances with the smallest distances;

[0030] Obtain the origin attribute data corresponding to the multiple second-class distances, count the first origin attribute data that appears most frequently, and use the first origin attribute data as the first origin attribute classification data of the amber sample to be tested.

[0031] Based on the KNN nonparametric classification supervision algorithm, the first origin attribute classification data and the standard origin attribute classification data corresponding to each amber origin are compared in different band segments to calculate the nearest neighbor attribute value, thereby obtaining the second origin attribute classification data of the amber sample to be tested in each band segment. The second origin attribute data that appears most frequently is counted to obtain the origin attribute classification data of the amber sample to be tested.

[0032] In one possible implementation, the amber sample to be tested is identified based on the first predicted origin and the origin attribute classification data to obtain the amber origin of the amber sample to be tested, specifically including:

[0033] A first weight value is set for the first predicted origin, and a second weight value is set for the origin attribute classification data. It is determined whether the first weight value and the second weight value are the same. If they are, the first predicted origin and the origin attribute classification data are set as the amber origin of the amber sample to be tested. If not, the first weight value and the second weight value are adjusted according to the first predicted origin and the origin attribute classification data. Based on the adjusted weights, the origin of the amber sample to be tested is re-determined to obtain the amber origin of the amber sample to be tested.

[0034] In one possible implementation, the infrared spectral data of each amber sample in the standard sample set are collected, specifically including:

[0035] Using a diamond pressure cell attachment of a micro-infrared spectrometer, infrared spectral data of each amber sample in the standard sample set were acquired. The acquisition conditions were a low-temperature detector and a spectral range of 600 cm⁻¹. -1 -4000cm -1 4cm resolution -1 128 scans.

[0036] The present invention also provides an automatic identification device for amber origin, comprising: an infrared spectral data acquisition module, a data tag setting module, an infrared spectral data splicing module, a spliced ​​amber data processing module, a first predicted origin acquisition module, an origin attribute acquisition module, and an amber origin acquisition module;

[0037] The infrared spectral data acquisition module is used to select amber samples from multiple amber producing areas, generate a standard sample set, and acquire the infrared spectral data of each amber sample in the standard sample set.

[0038] The data tag setting module is used to set corresponding data tags for the infrared spectral data of each sample according to the multiple amber origins, and integrate the infrared spectral data of all samples to obtain an infrared spectral dataset;

[0039] The infrared spectral data stitching module is used to acquire the infrared spectral data of the amber sample to be tested, and stitch the infrared spectral data to be tested into the infrared spectral dataset to obtain the stitched amber dataset.

[0040] The spliced ​​amber data processing module is used to process the spliced ​​amber dataset to obtain a standard dataset, perform dimensionality reduction processing on the standard dataset to obtain a dimensionality reduction matrix, and classify the dimensionality reduction matrix based on the data labels to obtain the infrared spectral data center values ​​corresponding to the multiple amber origins.

[0041] The first predicted origin acquisition module is used to calculate the first type distance corresponding to each infrared spectral data center value based on the dimensionality reduction matrix and the infrared spectral data center value, obtain the minimum value of the first type distance, and take the amber origin corresponding to the minimum value as the first predicted origin of the amber sample to be tested.

[0042] The origin attribute acquisition module is used to calculate the second type distance between the infrared spectral data to be tested and the infrared spectral data of each sample in the standard sample set, and based on the second type distance, combine the KNN nonparametric classification supervision algorithm to obtain the origin attribute data corresponding to each amber sample in the standard sample set, so as to obtain the origin attribute classification data of the amber sample to be tested.

[0043] The amber origin acquisition module is used to determine the origin of the amber sample to be tested based on the first predicted origin and the origin attribute classification data, thereby obtaining the amber origin of the amber sample to be tested.

[0044] In one possible implementation, the spliced ​​amber data processing module is used to process the spliced ​​amber dataset to obtain a standard dataset, specifically including:

[0045] Based on a preset formula for calculating average spectral data, the average spectral data of infrared spectral data corresponding to all amber origins in the spliced ​​amber dataset is calculated, and the average spectral data is used as the standard spectrum of amber.

[0046] Linear regression processing is performed on the standard amber spectrum and the spliced ​​amber dataset to obtain the intercept and linear regression coefficients, and the offset of each amber infrared spectral data is obtained based on the intercept;

[0047] Based on the linear regression coefficients and the offset, multivariate scattering correction is performed on each infrared spectral data in the spliced ​​amber dataset to obtain corrected infrared spectral data.

[0048] Differential operations are performed on the corrected infrared spectral data to obtain standard infrared spectral data. All the standard infrared spectral data are then integrated to obtain a standard dataset.

[0049] In one possible implementation, the spliced ​​amber data processing module is used to perform dimensionality reduction processing on the standard dataset to obtain a dimensionality reduction matrix, specifically including:

[0050] Obtain the set of infrared spectral data wavenumber intensity corresponding to each amber origin in the standard dataset, and calculate the average wavenumber intensity of each set of infrared spectral data wavenumber intensity.

[0051] Subtracting the average wavenumber intensity from each wavenumber intensity value in the infrared spectral data wavenumber intensity set yields the first infrared spectral data wavenumber intensity set.

[0052] Simultaneously, the first covariance matrix of the first infrared spectral data wavenumber intensity set is obtained, and the eigenvalues ​​and eigenvectors of the first covariance matrix are calculated.

[0053] Based on the eigenvalues ​​and eigenvectors, the first infrared spectral data wavenumber intensity set is subjected to dimensionality reduction processing to obtain a dimensionality reduction matrix.

[0054] In one possible implementation, the first predicted origin acquisition module is configured to calculate a first type of distance corresponding to each infrared spectral data center value based on the dimensionality reduction matrix and the infrared spectral data center value, specifically including:

[0055] Obtain the second covariance matrix of the reduced-dimensional matrix, and calculate the inverse of the second covariance matrix;

[0056] Based on the dimensionality reduction matrix, the infrared spectral data center value, and the inverse matrix, calculate the first type distance corresponding to each infrared spectral data center value.

[0057] In one possible implementation, the origin attribute acquisition module is used to acquire the origin attribute data corresponding to each amber sample in the standard sample set based on the second type of distance and combined with the KNN nonparametric classification supervision algorithm, to obtain the origin attribute classification data of the amber sample to be tested, specifically including:

[0058] Infrared spectral data from different amber origins are divided into bands to obtain multiple band segments, and the second type distance between the infrared spectral data to be tested corresponding to the multiple band segments and the infrared spectral data of each sample in the standard sample set is calculated.

[0059] Obtain all second-type distances, sort all Euclidean distances from largest to smallest, and select the multiple second-type distances with the smallest distances;

[0060] Obtain the origin attribute data corresponding to the multiple second-class minimum distance values, count the first origin attribute data that appears most frequently, and use the first origin attribute data as the first origin attribute classification data of the amber sample to be tested.

[0061] Based on the KNN nonparametric classification supervision algorithm, the first origin attribute classification data and the standard origin attribute classification data corresponding to each amber origin are compared in different band segments to calculate the nearest neighbor attribute value, thereby obtaining the second origin attribute classification data of the amber sample to be tested in each band segment. The second origin attribute data that appears most frequently is counted to obtain the origin attribute classification data of the amber sample to be tested.

[0062] In one possible implementation, the amber origin acquisition module is used to determine the origin of the amber sample to be tested based on the first predicted origin and the origin attribute classification data, thereby obtaining the amber origin of the amber sample to be tested, specifically including:

[0063] A first weight value is set for the first predicted origin, and a second weight value is set for the origin attribute classification data. It is determined whether the first weight value and the second weight value are the same. If they are, the first predicted origin and the origin attribute classification data are set as the amber origin of the amber sample to be tested. If not, the first weight value and the second weight value are adjusted according to the first predicted origin and the origin attribute classification data. Based on the adjusted weights, the origin of the amber sample to be tested is re-determined to obtain the amber origin of the amber sample to be tested.

[0064] In one possible implementation, the infrared spectral data acquisition module is used to acquire the infrared spectral data of each amber sample in the standard sample set, specifically including:

[0065] Using a diamond pressure cell attachment of a micro-infrared spectrometer, infrared spectral data of each amber sample in the standard sample set were acquired. The acquisition conditions were a low-temperature detector and a spectral range of 600 cm⁻¹. -1 -4000cm -1 4cm resolution -1 128 scans.

[0066] The present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the automatic identification method for amber origin as described in any of the preceding claims.

[0067] The present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the automatic identification method for amber origin as described in any of the preceding claims.

[0068] This invention provides an automatic identification method, apparatus, device, and storage medium for amber origin, which, compared with existing technologies, has the following advantages:

[0069] By acquiring infrared spectral data of amber samples from multiple amber origins, an infrared spectral dataset is formed. The infrared spectral data of the amber sample to be tested are then stitched together into this dataset. Data processing is performed on the stitched dataset to obtain data center values ​​for the infrared spectral data corresponding to multiple amber origins. Based on the first type of distance corresponding to each infrared spectral data center value, the first predicted origin of the amber sample to be tested is obtained. Then, the second type of distance between the infrared spectral data to be tested and the infrared spectral data of each sample is calculated to obtain the origin attribute classification data of the amber sample to be tested. Based on the first predicted origin and the origin attribute classification data, the origin of the amber sample to be tested is determined, thus identifying the amber origin. Compared with existing technologies, this invention can automatically identify amber samples from different amber origins based on infrared spectral characteristics with high accuracy. Furthermore, it can not only accurately identify amber samples from six major producing areas but also accurately identify amber from different sub-originating areas within the same major producing area. Attached Figure Description

[0070] Figure 1 This is a flowchart illustrating an embodiment of an automatic identification method for the origin of amber provided by the present invention;

[0071] Figure 2 This is a schematic diagram of an embodiment of an automatic amber origin identification device provided by the present invention. Detailed Implementation

[0072] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0073] Example 1

[0074] See Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of an automatic identification method for the origin of amber provided by the present invention. Figure 1As shown, the method includes steps 101-107, as detailed below:

[0075] Step 101: Select amber samples from multiple amber producing areas to generate a standard sample set, and collect the infrared spectral data of each amber sample in the standard sample set.

[0076] In one embodiment, amber samples were selected from six major producing areas: Russia, Ukraine, Dominican Republic, Mexico, Myanmar, and Fushun, China.

[0077] Preferably, the amber sample has been approved by senior experts in the industry for determining the origin of the amber or has proof of origin. If the amber sample is raw material, it needs to be cut and polished.

[0078] Preferably, 292 amber samples from each amber producing region are selected as the standard sample set.

[0079] In one embodiment, before acquiring the infrared spectral data corresponding to each amber sample, the sample surface of each amber sample is effectively cleaned to avoid the introduction of surface contaminants, which could cause significant errors in the subsequently acquired infrared spectral data.

[0080] In one embodiment, a diamond cell attachment of a micro-infrared spectrometer is used to collect infrared spectral data for each amber sample in the standard sample set. The acquisition conditions are: low-temperature detector, spectral range 600 cm⁻¹. -1 -4000cm -1 4cm resolution -1 128 scans.

[0081] Preferably, the infrared transmission spectrum of amber samples is acquired using the micro-infrared diamond pressure cell method. Compared with the existing reflection method and potassium bromide powder transmission method, the obvious advantage of this method is that it is not limited by the appearance of the amber sample, and it is nearly non-destructive. It can acquire stable sample spectra with a high performance-to-noise ratio while keeping the appearance of the amber sample unchanged. It is suitable for the origin identification of most amber jewelry and handicrafts in the circulation field.

[0082] Step 102: Based on the multiple amber origins, set corresponding data tags for the infrared spectral data of each sample, integrate the infrared spectral data of all samples, and obtain an infrared spectral dataset.

[0083] In one embodiment, the infrared spectral data of each amber sample is processed. Specifically, Thermo Fisher Scientific’s OMNIC software is used to set the infrared data points in the sample infrared spectral data to maintain the same horizontal spacing. The data format is (x0, y0), where x0 represents the wavenumber and y0 represents the intensity of the wavenumber.

[0084] In one embodiment, multiple corresponding data tags are set for multiple amber producing areas; preferably, data tags 0, 1, 2, 3, 4, and 5 are set for the six major amber producing areas of Russia, Ukraine, Dominican Republic, Mexico, Myanmar, and Fushun, China, respectively.

[0085] In one embodiment, after generating the infrared spectral dataset, the number of amber samples corresponding to each amber origin is used as the sample wavenumber, and this wavenumber is set as row x of the infrared spectral dataset; the wavenumber intensity corresponding to different sample wavenumbers is set as column y of the infrared spectral dataset.

[0086] Preferably, since the number of amber samples obtained from each amber origin is the same, classifier bias caused by data imbalance can be avoided in the future.

[0087] In one embodiment, the infrared spectral dataset can be set to Excel format and compiled into Tdata.csv; alternatively, the infrared spectral dataset can be set to text format and compiled into Tdata.txt.

[0088] Step 103: Obtain the infrared spectral data of the amber sample to be tested, and stitch the infrared spectral data to the infrared spectral dataset to obtain the stitched amber dataset.

[0089] In one embodiment, based on the data processing method of the sample infrared spectral data corresponding to each amber sample, the test infrared spectral data of the amber sample to be tested is processed to make the test infrared spectral data and the sample infrared spectral data have the same data format.

[0090] In one embodiment, after processing the infrared spectral data to be measured, it is directly input into an infrared spectral dataset to obtain a spliced ​​amber dataset.

[0091] Step 104: Process the spliced ​​amber dataset to obtain a standard dataset, perform dimensionality reduction on the standard dataset to obtain a dimensionality reduction matrix, and classify the dimensionality reduction matrix based on the data labels to obtain the infrared spectral data center values ​​corresponding to the multiple amber origins.

[0092] In one embodiment, when processing the spliced ​​amber dataset, the average spectral data of the infrared spectral data corresponding to all amber origins in the spliced ​​amber dataset is calculated based on a preset average spectral data calculation formula. The average spectral data is used as the standard spectrum for amber; preferably, the infrared spectral data corresponding to all amber origins in the spliced ​​amber dataset are set as... The preset formula for calculating the average spectral data is as follows:

[0093]

[0094] In the formula, For average spectral data, F i To stitch together a single infrared spectral data point from the amber dataset, where n is the total number of infrared spectral data points from all the amber datasets to be stitched together.

[0095] In one embodiment, linear regression processing is performed on the amber standard spectrum and the spliced ​​amber dataset to obtain the intercept, and based on the linear regression coefficients, the intercept is used as the offset for each infrared spectral data; specifically, the... and F i Performing a univariate linear regression on x and y yields the intercept b. i and linear regression coefficient k i The linear regression equation is shown below:

[0096]

[0097] In one embodiment, based on the linear regression coefficient and the offset, multivariate scattering correction is performed on the infrared spectral data of each sample in the pieced amber dataset to obtain corrected infrared spectral data. The multivariate scattering correction involves subtracting the offset from a single infrared spectral data in the pieced amber dataset and then dividing by the linear regression coefficient to obtain the corrected infrared spectral data. The process of performing multivariate scattering correction on each infrared spectral data is as follows:

[0098]

[0099] In the formula, F i(Amber MSC) To correct the infrared spectral data.

[0100] In one embodiment, the corrected infrared spectral data is subjected to a differential operation to obtain standard infrared spectral data, making the spectral differences more obvious; at the same time, all the standard infrared spectral data are integrated to obtain a standard dataset, wherein the differential operation process is as follows:

[0101] Y n0 amber =Y (n+1) -Y n ;

[0102] Y n1 amber =Y (n+1) -Y n ;

[0103] Y n2 amber =Y(n+1) -Y n ;

[0104] Y n3 amber =Y (n+1) -Y n ;

[0105] Y n4 amber =Y (n+1) -Y n ;

[0106] Y n5 amber =Y (n+1) -Y n ;

[0107]

[0108] In one embodiment, when performing dimensionality reduction on the standard dataset, the wavenumber intensity set X of the infrared spectral data corresponding to each amber origin in the standard dataset is obtained. n (n = 1, 2, 3, 4, 5, 6, 7, 8, 9, ..., 1000, ..., 2000, ...), calculate the average wavenumber intensity for each set of infrared spectral data wavenumber intensity. Subtract the average wavenumber intensity from each wavenumber intensity value in the infrared spectral data wavenumber intensity set, i.e. The first infrared spectral data wavenumber intensity set is obtained by averaging the data y in each column of the standard dataset, and then subtracting the average value from the data in each column to obtain a new dataset for each column, thus generating the first infrared spectral data wavenumber intensity set. The first infrared spectral data wavenumber intensity set is an m×n matrix, where m is the sum of the number of infrared spectral data of all samples in the standard dataset and the number of infrared spectral data to be measured, and n is the dimension, with the total number of columns in the standard dataset as the dimension value.

[0109] In one embodiment, the first covariance matrix of the first infrared spectral data wavenumber intensity set is obtained simultaneously, and the eigenvalues ​​and eigenvectors of the first covariance matrix are calculated; preferably, after obtaining the eigenvalues ​​and eigenvectors, the eigenvalues ​​and eigenvectors are arranged in descending order of correlation, and the top 10 eigenvalues ​​and corresponding eigenvectors n×10 that best represent the characteristics of amber from different origins are selected.

[0110] In one embodiment, the first infrared spectral data wavenumber intensity set is dimensionality-reduced based on the eigenvalues ​​and the eigenvectors to obtain a dimensionality-reduced matrix. Specifically, the first infrared spectral data wavenumber intensity set is multiplied by the eigenvectors to obtain the dimensionality-reduced matrix, i.e., Y. n0 amber(n=1, 2, 3, 4, 5, 6, 7, 8, 9, 10).

[0111] Step 105: Based on the dimensionality reduction matrix and the infrared spectral data center value, calculate the first type distance corresponding to each infrared spectral data center value, obtain the minimum value of the first type distance, and take the amber origin corresponding to the minimum value as the first predicted origin of the amber sample to be tested.

[0112] In one embodiment, the infrared spectral data center value is obtained; specifically, based on the data labels set for different amber origins, the infrared spectral data corresponding to different amber origins in the dimensionality reduction matrix are obtained, such as 0-292 for data label 0, 293-585 for data label 1, 586-877 for data label 2, 878-1169 for data label 3, 1170-1461 for data label 4, 1462-1753 for data label 5, and 1754 for the amber sample data to be tested; the infrared spectral data center value C corresponding to the infrared spectral data of different amber origins is taken. n According to different amber origins, amber can be divided into C0, C1, C2, C3, C4, and C5.

[0113] In one embodiment, the first type of distance corresponding to each infrared spectral data center value is calculated. Specifically, the second covariance matrix Σ10×10 of the dimensionality reduction matrix is ​​obtained, and the inverse matrix Σ of the second covariance matrix is ​​calculated. -1 .

[0114] In one embodiment, a first type of distance corresponding to each infrared spectral data center value is calculated based on the dimensionality reduction matrix, the infrared spectral data center value, and the inverse matrix. Specifically, each infrared spectral data point in the dimensionality reduction matrix is ​​subtracted from its corresponding infrared spectral data center value to obtain a first difference matrix, and the diagonal matrix value corresponding to the first difference matrix is ​​calculated. The first difference matrix, the inverse matrix, and the diagonal matrix value are multiplied together to obtain the first type of distance corresponding to each infrared spectral data center value. The formula for calculating the first type of distance is as follows:

[0115]

[0116] In the formula, the first difference matrix is ​​f1-μF1, f2-μF2, ..., f m -μF m In this embodiment, the subscript m represents the number of selected amber origins, and the diagonal matrix value is... The inverse matrix is

[0117] In one embodiment, the objects to be tested are amber samples from six major producing areas: Russia, Ukraine, Dominican Republic, Mexico, Myanmar, and Fushun, China. Therefore, m = 6. The first-type distance corresponding to the infrared spectral data center value of each amber producing area is denoted as D. M0 D M1 D M2 D M3 D M4 D M5 .

[0118] The formula for calculating the first type of distance is as follows:

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125] Step 106: Calculate the second type distance between the infrared spectral data to be tested and the infrared spectral data of each sample in the standard sample set, and based on the second type distance, use the KNN nonparametric classification supervision algorithm to obtain the origin attribute data corresponding to each amber sample in the standard sample set, and obtain the origin attribute classification data of the amber sample to be tested.

[0126] In one embodiment, infrared spectral data from different amber origins are divided into bands to obtain multiple band segments, and the second type distance between the infrared spectral data to be tested corresponding to the multiple band segments and the infrared spectral data of each sample in the standard sample set is calculated.

[0127] Specifically, the infrared spectrum, which represents the regional characteristics of amber, is divided into bands to form multiple band segments, and the second-type distances corresponding to different band segments are calculated; preferably, it is divided into a full-band 400-6000 cm. -1 W1, 600-2000cm -1 W2, 2000-3000cm -1 W3, 3000-4000cm -1 It is W4.

[0128] In one embodiment, all second-type distances are obtained, and the second-type distances are sorted from largest to smallest. The multiple second-type distances with the smallest distances are selected, wherein the number of the multiple second-type distances with the smallest distances is a user-preset number.

[0129] In one embodiment, the origin attribute data of the sample infrared spectral data corresponding to the multiple second-class minimum distance values ​​are obtained, the origin attribute data that appears most frequently is counted and set as the first origin attribute data, and the first origin attribute data is used as the first origin attribute classification data of the amber sample to be tested.

[0130] In one embodiment, based on the KNN nonparametric classification supervision algorithm, the first origin attribute classification data and the standard origin attribute classification data corresponding to each amber origin are compared in different band segments to calculate the nearest neighbor attribute value, thereby obtaining the second origin attribute classification data of the amber sample to be tested in each band segment; the second origin attribute data that appears most frequently is counted to obtain the origin attribute classification data of the amber sample to be tested.

[0131] Specifically, the attribute value of amber from which its origin is to be distinguished is assumed to be O. x W1 range (400-6000cm) -1 Depending on the origin (O1, O2, O3, O4, O5, ...), there will be products similar to O. x The neighboring attribute values ​​are calculated, and the known origin data label with the most points will be used as the second origin attribute classification data 1 to distinguish amber.

[0132] Specifically, the attribute value of amber from which its origin is to be distinguished is assumed to be O. x W2 range (600-2000cm) -1 Depending on the origin (O1, O2, O3, O4, O5, ...), there will be products similar to O. x The neighboring attribute values ​​are calculated, and the known origin data label with the most points will be used as the second origin attribute classification data 2 to distinguish amber.

[0133] Specifically, the attribute value of amber from which its origin is to be distinguished is assumed to be O. x Within the W3 range (2000-3000 cm⁻¹), depending on the origin (O1, O2, O3, O4, O5, ...), there will be [various types] similar to O. x The neighboring attribute values ​​are calculated, and the known origin data label with the most points will be used as the second origin attribute classification data for distinguishing amber.

[0134] Specifically, the attribute value of amber from which its origin is to be distinguished is assumed to be O. xWithin the W4 range (3000-4000 cm⁻¹), depending on the origin (O1, O2, O3, O4, O5, ...), there will be [various types] similar to O. x The neighboring attribute values ​​are calculated, and the known origin data label with the most points will be used as the second origin attribute classification data to distinguish amber.

[0135] Specifically, by combining the second origin attribute classification data 1-4, the second origin attribute data that appears most frequently is statistically analyzed to obtain the final origin attribute classification data of the amber sample to be tested.

[0136] Step 107: Based on the first predicted origin and the origin attribute classification data, determine the origin of the amber sample to be tested to obtain the amber origin of the amber sample to be tested.

[0137] In one embodiment, a first weight value is set for the first predicted origin, a second weight value is set for the origin attribute classification data, and the first weight value and the second weight value are iterated based on different weight ratios to obtain the weight ratio with the most accurate discrimination result.

[0138] Preferably, both the first weight value and the second weight value are set to 0.5.

[0139] In one embodiment, it is determined whether the first weight value and the second weight value are the same. If they are, the first predicted origin and the origin attribute classification data are set as the amber origin of the amber sample to be tested. If not, the weights of the first predicted origin and the origin attribute are adjusted based on the higher accuracy of the reverse inference from the results. The two weights are then weighted to obtain the final amber origin conclusion. That is, the first weight value and the second weight value are adjusted according to the first predicted origin and the origin attribute classification data, and the origin of the amber sample to be tested is re-determined based on the adjusted weights to obtain the amber origin of the amber sample to be tested.

[0140] In one embodiment, after differentiating amber samples from major producing areas using the above steps, such as after differentiating amber samples from six major producing areas, the steps of the above automatic identification method are repeated for amber samples from each major producing area to automatically identify the origin of amber from smaller producing areas within a single major producing area. This yields the identification results of the origin of amber from smaller producing areas within a single major producing area, thereby achieving accurate differentiation of amber from different sub-originating areas within the same major producing area. For example, for Baltic amber, if further differentiation of smaller producing areas is desired, the large Baltic producing area is identified, and then the above steps are repeated for producing areas such as Ukra, Russia, and Lithuania, greatly improving the accuracy.

[0141] Example 2

[0142] See Figure 2 , Figure 2 This is a schematic diagram of one embodiment of an automatic amber origin identification device provided by the present invention, as shown below. Figure 2 As shown, the device includes an infrared spectral data acquisition module 201, a data tag setting module 202, an infrared spectral data stitching module 203, a stitched amber data processing module 204, a first predicted origin acquisition module 205, an origin attribute acquisition module 206, and an amber origin acquisition module 207, as detailed below:

[0143] The infrared spectral data acquisition module 201 is used to select amber samples from multiple amber producing areas, generate a standard sample set, and acquire the sample infrared spectral data corresponding to each amber sample in the standard sample set.

[0144] The data tag setting module 202 is used to set corresponding data tags for the infrared spectral data of each sample according to the multiple amber origins, and integrate the infrared spectral data of all samples to obtain an infrared spectral dataset.

[0145] The infrared spectral data stitching module 203 is used to acquire the infrared spectral data of the amber sample to be tested, and stitch the infrared spectral data to be tested into the infrared spectral dataset to obtain the stitched amber dataset.

[0146] The spliced ​​amber data processing module 204 is used to process the spliced ​​amber dataset to obtain a standard dataset, perform dimensionality reduction processing on the standard dataset to obtain a dimensionality reduction matrix, and classify the dimensionality reduction matrix based on the data labels to obtain the infrared spectral data center values ​​corresponding to the multiple amber origins.

[0147] The first predicted origin acquisition module 205 is used to calculate the first type distance corresponding to each infrared spectral data center value based on the dimensionality reduction matrix and the infrared spectral data center value, obtain the minimum value of the first type distance, and use the amber origin corresponding to the minimum value as the first predicted origin of the amber sample to be tested.

[0148] The origin attribute acquisition module 206 is used to calculate the second type distance between the infrared spectral data to be tested and the infrared spectral data of each sample in the standard sample set, and based on the second type distance, combine the KNN nonparametric classification supervision algorithm to obtain the origin attribute data corresponding to each amber sample in the standard sample set, so as to obtain the origin attribute classification data of the amber sample to be tested.

[0149] The amber origin acquisition module 207 is used to determine the origin of the amber sample to be tested based on the first predicted origin and the origin attribute classification data, thereby obtaining the amber origin of the amber sample to be tested.

[0150] In one embodiment, the spliced ​​amber data processing module 204 is used to process the spliced ​​amber dataset to obtain a standard dataset, specifically including: calculating the average spectral data of all infrared spectral data in the spliced ​​amber dataset based on a preset average spectral data calculation formula, and using the average spectral data as the amber standard spectrum; performing linear regression processing on the amber standard spectrum and the spliced ​​amber dataset to obtain the intercept and linear regression coefficients, and obtaining the offset of the amber infrared spectral data based on the intercept; performing multivariate scattering correction processing on each infrared spectral data in the spliced ​​amber dataset according to the linear regression coefficients and the offset to obtain corrected infrared spectral data; performing first-order difference operation on the corrected infrared spectral data to obtain standard infrared spectral data; and integrating all the standard infrared spectral data to obtain a standard dataset.

[0151] In one embodiment, the spliced ​​amber data processing module 204 is used to perform dimensionality reduction processing on the standard dataset to obtain a dimensionality reduction matrix. Specifically, it includes: obtaining the infrared spectral data wavenumber intensity set corresponding to each amber origin in the standard dataset, and calculating the average wavenumber intensity of each infrared spectral data wavenumber intensity set; subtracting the average wavenumber intensity from each infrared spectral data wavenumber intensity value in the infrared spectral data wavenumber intensity set to obtain a first infrared spectral data wavenumber intensity set; simultaneously obtaining the first covariance matrix of the first infrared spectral data wavenumber intensity set, and calculating the eigenvalues ​​and eigenvectors of the first covariance matrix; and performing dimensionality reduction processing on the first infrared spectral data wavenumber intensity set based on the eigenvalues ​​and the eigenvectors to obtain a dimensionality reduction matrix.

[0152] In one embodiment, the first predicted origin acquisition module 205 is used to calculate a first type distance corresponding to each infrared spectral data center value based on the dimensionality reduction matrix and the infrared spectral data center value. Specifically, it includes: obtaining the second covariance matrix of the dimensionality reduction matrix and calculating the inverse matrix of the second covariance matrix; and calculating the first type distance corresponding to each infrared spectral data center value based on the dimensionality reduction matrix, the infrared spectral data center value and the inverse matrix.

[0153] In one embodiment, the origin attribute acquisition module 206 is used to acquire origin attribute data corresponding to each amber sample in the standard sample set based on the second type distance and combined with the KNN nonparametric classification supervision algorithm, to obtain the origin attribute classification data of the amber sample to be tested. Specifically, this includes: dividing the infrared spectral data of different amber origins into bands to obtain multiple band segments, and calculating the second type distance between the infrared spectral data to be tested corresponding to the multiple band segments and the infrared spectral data of each sample in the standard sample set; acquiring all the second type distances, sorting all the second type distances from largest to smallest, and selecting the multiple second type distances with the smallest distances. The system calculates the origin attribute data corresponding to the minimum value of the multiple second-class distances, identifies the most frequently occurring first-origin attribute data, and uses this first-origin attribute data as the first-origin attribute classification data for the amber sample to be tested. Based on the KNN nonparametric classification supervision algorithm, the system calculates the nearest neighbor attribute value between the first-origin attribute classification data and the standard origin attribute classification data corresponding to each amber origin in different spectral segments to obtain the second-origin attribute classification data for the amber sample to be tested in each spectral segment. The system then identifies the most frequently occurring second-origin attribute data to obtain the origin attribute classification data for the amber sample to be tested.

[0154] In one embodiment, the amber origin acquisition module 207 is used to determine the origin of the amber sample to be tested based on the first predicted origin and the origin attribute classification data, and obtain the amber origin of the amber sample to be tested. Specifically, this includes: setting a first weight value for the first predicted origin, setting a second weight value for the origin attribute classification data, determining whether the first weight value and the second weight value are the same; if so, setting the first predicted origin and the origin attribute classification data as the amber origin of the amber sample to be tested; if not, adjusting the weights of the first weight value and the second weight value based on the first predicted origin and the origin attribute classification data, and re-determining the origin of the amber sample to be tested based on the adjusted weights, and obtaining the amber origin of the amber sample to be tested.

[0155] Since existing infrared spectroscopy identification methods are mostly reflectance and potassium bromide powder transmission methods, the spectral quality of infrared spectra acquired using the reflectance method is significantly affected by the appearance of the sample, and some intricately carved samples cannot obtain effective analytical spectra using the reflectance method. Furthermore, the potassium bromide powder transmission method can damage the appearance of the sample, which does not meet the non-destructive to near-non-destructive testing requirements in jewelry and gemstone identification. Based on these problems, in one embodiment, the infrared spectral data acquisition module 201 is used to acquire the infrared spectral data of each amber sample in the standard sample set. Specifically, it includes: acquiring the infrared spectral data of each amber sample in the standard sample set based on the diamond pressure cell attachment of a micro-infrared spectrometer, wherein the acquisition conditions are a low-temperature detector and a spectral range of 600 cm⁻¹. -1 -4000cm -1 4cm resolution -1 128 scans.

[0156] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0157] It should be noted that the above-described embodiment of the automatic amber origin identification device is merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0158] Based on the above-described embodiments of the automatic identification method for amber origin, another embodiment of the present invention provides an automatic identification terminal device for amber origin. This automatic identification terminal device for amber origin includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the automatic identification method for amber origin according to any embodiment of the present invention.

[0159] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the automatic identification terminal device at the amber origin.

[0160] The automatic identification terminal device for amber origin can be a desktop computer, laptop, handheld computer, or cloud server, etc. The automatic identification terminal device for amber origin may include, but is not limited to, processors and memory.

[0161] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the automatic identification terminal equipment at the amber origin, connecting various parts of the automatic identification terminal equipment throughout the amber origin using various interfaces and lines.

[0162] The memory can be used to store the computer programs and / or modules. The processor, by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory, realizes various functions of the automatic identification terminal device for the amber origin. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0163] Based on the above-described embodiments of the automatic identification method for amber origin, another embodiment of the present invention provides a storage medium comprising a stored computer program, wherein, when the computer program is executed, the device containing the storage medium is controlled to execute the automatic identification method for amber origin according to any embodiment of the present invention.

[0164] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0165] In summary, the present invention provides an automatic identification method, apparatus, device, and storage medium for amber origin. It acquires infrared spectral data of amber samples from multiple origins to form an infrared spectral dataset. The infrared spectral data of the amber sample to be tested are then stitched together into this dataset. Data processing is performed on the stitched dataset to obtain data center values ​​for the infrared spectral data corresponding to multiple amber origins. A first predicted origin for the amber sample to be tested is obtained based on a first type of distance corresponding to each infrared spectral data center value. A second type of distance is then calculated between the infrared spectral data to be tested and the infrared spectral data of each sample to obtain origin attribute classification data for the amber sample to be tested. Based on the first predicted origin and the origin attribute classification data, the origin of the amber sample to be tested is determined, thus identifying the amber origin. Compared with existing technologies, the present invention can identify amber samples based on their origin, improving the accuracy of identification.

[0166] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for automatically identifying the origin of an amber, characterized in that, The method comprises the following steps: selecting a plurality of amber samples from different amber producing areas to form a standard sample set, and collecting sample infrared spectrum data corresponding to each amber sample in the standard sample set; labeling each sample infrared spectrum data according to the producing area of the sample, integrating all sample infrared spectrum data to obtain an infrared spectrum data set; obtaining the infrared spectrum data of the to-be-tested amber sample, and splicing the infrared spectrum data into the infrared spectrum data set to obtain a spliced amber data set; performing data processing on the spliced amber data set to obtain a standard data set, performing dimensionality reduction processing on the standard data set to obtain a dimensionality reduction matrix, and classifying the dimensionality reduction matrix based on the data label to obtain the infrared spectrum data center value corresponding to each amber producing area; calculating the first type distance corresponding to each infrared spectrum data center value based on the dimensionality reduction matrix and the infrared spectrum data center value, obtaining the minimum value of the first type distance, and taking the amber producing area corresponding to the minimum value as the first predicted producing area of the to-be-tested amber sample; calculating the second type distance between the to-be-tested infrared spectrum data and each sample infrared spectrum data in the standard sample set, and obtaining the producing area attribute data corresponding to each amber sample in the standard sample set based on the second type distance and the KNN non-parametric classification supervision algorithm to obtain the producing area attribute classification data of the to-be-tested amber sample; determining the producing area of the to-be-tested amber sample based on the first predicted producing area and the producing area attribute classification data to obtain the amber producing area of the to-be-tested amber sample; wherein the data processing on the spliced amber data set to obtain the standard data set comprises: calculating the average spectrum data of the infrared spectrum data corresponding to each amber producing area in the spliced amber data set based on a preset average spectrum data calculation formula, and taking the average spectrum data as the amber standard spectrum; performing linear regression processing on the amber standard spectrum and the spliced amber data set to obtain the intercept and linear regression coefficient, and obtaining the offset of each amber infrared spectrum data based on the intercept; performing multivariate scatter correction processing on each infrared spectrum data in the spliced amber data set based on the linear regression coefficient and the offset to obtain corrected infrared spectrum data; performing difference operation on the corrected infrared spectrum data to obtain standard infrared spectrum data, and integrating all standard infrared spectrum data to obtain a standard data set.

2. The method of claim 1, wherein the step of automatically identifying the amber origin is performed by a computer program. The dimensionality reduction processing on the standard data set to obtain a dimensionality reduction matrix comprises: obtaining the infrared spectrum data wave number intensity set corresponding to each amber producing area in the standard data set, and calculating the wave number intensity average value of each infrared spectrum data wave number intensity set; subtracting the wave number intensity average value from each infrared spectrum data wave number intensity value in the infrared spectrum data wave number intensity set to obtain a first infrared spectrum data wave number intensity set; simultaneously obtaining a first covariance matrix of the first infrared spectrum data wave number intensity set, and calculating the eigenvalue and eigenvector of the first covariance matrix; According to the characteristic value and the characteristic vector, the first infrared spectrum data wavenumber intensity set is processed by dimension reduction to obtain a dimension reduction matrix.

3. The method of claim 1, wherein the step of automatically identifying the amber origin is performed by a computer program. According to the dimension reduction matrix and the infrared spectrum data center value, a first type distance corresponding to each infrared spectrum data center value is calculated, specifically including: A second covariance matrix of the dimension reduction matrix is obtained, and an inverse matrix of the second covariance matrix is calculated; According to the dimension reduction matrix, the infrared spectrum data center value and the inverse matrix, a first type distance corresponding to each infrared spectrum data center value is calculated.

4. The method of claim 1, wherein the step of automatically identifying the amber origin is performed by a computer program. Based on the second type distance, the origin attribute data corresponding to each amber sample in the standard sample set is obtained by combining the KNN non-parametric classification supervision algorithm to obtain the origin attribute classification data of the test amber sample, specifically including: The infrared spectrum data of different amber origins is divided into wave bands to obtain a plurality of wave band segments, and a second type distance between the test infrared spectrum data and the infrared spectrum data of each sample in the standard sample set corresponding to the plurality of wave band segments is calculated; All second type distances are obtained, and the all second type distances are sorted from large to small, and a plurality of second type distances with the smallest distance are selected; The origin attribute data corresponding to the plurality of second type distances is obtained, the first origin attribute data with the most occurrences is counted, and the first origin attribute data is taken as the first origin attribute classification data of the test amber sample; Based on the KNN non-parametric classification supervision algorithm, the first origin attribute classification data and the standard origin attribute classification data corresponding to each amber origin are calculated in different wave band segments to obtain the second origin attribute classification data corresponding to each wave band segment of the test amber sample, the second origin attribute data with the most occurrences is counted, and the origin attribute classification data of the test amber sample is obtained.

5. The method of claim 1, wherein the step of automatically identifying the amber origin is performed by a computer program. According to the first predicted origin and the origin attribute classification data, the origin of the test amber sample is judged to obtain the amber origin of the test amber sample, specifically including: The first weight value is set for the first predicted origin, the second weight value is set for the origin attribute classification data, it is judged whether the first weight value and the second weight value are the same, if yes, the first predicted origin and the origin attribute classification data are set as the amber origin of the test amber sample, if not, the first weight value and the second weight value are adjusted according to the first predicted origin and the origin attribute classification data, and the amber origin of the test amber sample is judged again based on the adjusted weight to obtain the amber origin of the test amber sample.

6. The method of claim 1, wherein the step of automatically identifying the amber origin is performed by a computer. The sample infrared spectrum data corresponding to each amber sample in the standard sample set is collected, specifically including: Based on the diamond anvil cell accessory of the microscopic infrared spectrometer, the sample infrared spectrum data corresponding to each amber sample in the standard sample set is collected, wherein the collection conditions are low temperature detector, spectrum range 600cm -1 -4000cm -1 , resolution 4cm -1 , scanning times 128 times.

7. An automatic identification device of amber origin, characterized in that, including: An infrared spectrum data acquisition module, a data label setting module, an infrared spectrum data splicing module, a spliced amber data processing module, a first predicted origin acquisition module, an origin attribute acquisition module and an amber origin acquisition module; The infrared spectrum data acquisition module is configured to select amber samples from multiple amber producing areas, generate a standard sample set, and acquire sample infrared spectrum data corresponding to each amber sample in the standard sample set; The data label setting module is configured to set a corresponding data label for each sample infrared spectrum data according to the multiple amber producing areas, integrate all sample infrared spectrum data, and obtain an infrared spectrum data set; The infrared spectrum data splicing module is configured to acquire to-be-tested infrared spectrum data of a to-be-tested amber sample, splice the to-be-tested infrared spectrum data into the infrared spectrum data set, and obtain a spliced amber data set; The spliced amber data processing module is configured to perform data processing on the spliced amber data set to obtain a standard data set, perform dimension reduction processing on the standard data set to obtain a dimension reduction matrix, perform classification on the dimension reduction matrix based on the data label, and obtain central values of infrared spectrum data corresponding to the multiple amber producing areas; The data processing on the spliced amber data set to obtain the standard data set includes: calculating average spectrum data of infrared spectrum data corresponding to all amber producing areas in the spliced amber data set based on a preset average spectrum data calculation formula, and taking the average spectrum data as an amber standard spectrum; performing linear regression processing on the amber standard spectrum and the spliced amber data set to obtain an intercept and a linear regression coefficient, and obtaining an offset of each amber infrared spectrum data based on the intercept; performing multivariate scatter correction processing on each infrared spectrum data in the spliced amber data set based on the linear regression coefficient and the offset to obtain corrected infrared spectrum data; performing difference operation on the corrected infrared spectrum data to obtain standard infrared spectrum data, and integrating all the standard infrared spectrum data to obtain the standard data set; The first predicted producing area acquisition module is configured to calculate a first type distance corresponding to each infrared spectrum data central value based on the dimension reduction matrix and the infrared spectrum data central value, acquire a minimum value of the first type distance, and take an amber producing area corresponding to the minimum value as a first predicted producing area of the to-be-tested amber sample; The producing area attribute acquisition module is configured to calculate a second type distance between the to-be-tested infrared spectrum data and each sample infrared spectrum data in the standard sample set, and acquire producing area attribute data corresponding to each amber sample in the standard sample set based on the second type distance and in combination with a KNN non-parametric classification supervision algorithm to obtain producing area attribute classification data of the to-be-tested amber sample; The amber producing area acquisition module is configured to acquire an amber producing area of the to-be-tested amber sample based on the first predicted producing area and the producing area attribute classification data.

8. A terminal device, comprising: A computer program product includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the automatic identification method of the amber producing area according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the automatic identification method of amber producing areas as claimed in any one of claims 1 to 6 when the computer program is running.

Images