Ceramic artwork traceability verification intelligent management system

Abstract

The invention relates to the technical field of ceramic artworks, in particular to an intelligent management system for traceability verification of ceramic artworks, which is used for solving the problems that accurate matching of images cannot be realized under various transformation conditions and accurate and reliable image comparison standards cannot be provided for traceability of ceramic artworks in the prior art. Through the image comparison module, SIFT feature extraction and an RANSAC algorithm are utilized, the image quality is improved through preprocessing, accurate matching under various transformation conditions is ensured, wrong matching, the proportion of comprehensive matching feature points, the number of inner points and the number of total matching points are effectively removed, a comprehensive similarity score is calculated, and the matching accuracy is improved. An accurate and reliable quantitative standard is provided for ceramic artwork traceability, the accuracy and robustness of image comparison are improved, and fine traceability of artworks is powerfully supported.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ceramic art, more particularly to a ceramic art traceability verification intelligent management system. BACKGROUND

[0002] As an important carrier of traditional culture, the authenticity identification, quality traceability and intellectual property protection of ceramic art have always been a problem in the industry. Traditional traceability methods have problems such as lack of transparency, easy tampering, low traceability efficiency, and are difficult to meet the demand for high quality and high transparency of ceramic art in the modern market.

[0003] The patent application with publication number CN114819995A discloses a traceability anti-counterfeiting method and system for ceramic purple sand teapot art, which includes the following steps: inputting and verifying the identity information of the ceramic purple sand teapot art issuing party; after verification, obtaining the ceramic purple sand teapot art information of the corresponding issuing party; generating a product sales code segment; generating a product traceability code for each ceramic purple sand teapot art; obtaining and establishing a person-goods binding relationship according to the identity information of the holder of the corresponding ceramic purple sand teapot art and the product traceability code of the corresponding ceramic purple sand teapot art, generating identity binding information; obtaining the circulation information of the corresponding ceramic purple sand teapot art according to the product traceability code and identity binding information of the ceramic purple sand teapot art and managing the circulation of goods; the present application can ensure the uniqueness of each product, effectively prevent counterfeiting, and trace each circulation link to effectively trace the product;

[0004] However, the above-mentioned reference patent generates a unique and tamper-proof one-to-one code for each ceramic purple sand teapot art, realizes effective traceability and anti-counterfeiting, and guarantees the authenticity and market integrity of the art, but cannot realize accurate matching of images under various transformation conditions, cannot provide accurate and reliable image comparison standards for ceramic art traceability, cannot focus on the spectral intensity integral of a specific wavelength range, cannot improve the recognition accuracy of spectral comparison, and cannot quantitatively compare key process parameters from multiple dimensions, cannot realize fast and accurate ceramic art traceability, and reduces work efficiency and accuracy.

[0005] Therefore, we propose a ceramic art traceability verification intelligent management system to solve the above problems. SUMMARY

[0006] The purpose of the present application is to provide a ceramic art product traceability verification intelligent management system, which solves the problem that the prior art cannot realize accurate matching of images under various transformation conditions, cannot provide accurate and reliable image comparison standards for ceramic art product traceability, cannot focus on the spectral intensity integral of a specific wavelength range, cannot improve the recognition accuracy of spectral comparison, and cannot quantitatively compare key process parameters from multiple dimensions, thereby realizing fast and accurate ceramic art product traceability and reducing work efficiency and accuracy.

[0007] The purpose of the present application is realized by the following technical solutions:

[0008] A ceramic art product traceability verification intelligent management system comprises a data acquisition module, an image comparison module, a spectral analysis module, a process information comparison module, and a traceability identification module.

[0009] The data acquisition module acquires sample information of various ancient ceramics, stores the sample information in a sample database for comparison and reference, and acquires image data, spectral data, and process information of the ceramic art product to be traced.

[0010] The image comparison module compares the acquired image of the ceramic art product to be traced with the reference image in the sample database, and calculates the similarity between the image of the art product and the reference image.

[0011] The spectral analysis module compares the acquired spectral data of the ceramic art product to be traced with the reference spectral data in the sample database, and calculates the similarity between the spectral data of the art product and the reference spectral data.

[0012] The process information comparison module compares the acquired process information of the ceramic art product to be traced with the reference process information in the sample database, and calculates the similarity between the process information of the art product and the reference process information.

[0013] The traceability identification module integrates the results of image comparison, spectral analysis, and process comparison, and comprehensively analyzes and judges the ceramic art product to be traced.

[0014] As a preferred embodiment of the present application, the specific process of the image comparison module comparing the acquired image of the ceramic art product to be traced with the reference image in the sample database is as follows:

[0015] The image of the ceramic art product to be traced and the reference image in the sample database are acquired, and pretreatment operations are performed on the image of the ceramic art product to be traced and the reference image in the sample database.

[0016] Feature vectors are extracted from the pretreated image of the art product and the reference image, and the specific steps are as follows:

[0017] The scale space is constructed using a Difference of Gaussians (DoG) pyramid, which is built by first applying a Gaussian filter to the input image to generate a Gaussian pyramid, where the standard deviation of the Gaussian kernel, sigma, increases with the pyramid level, and then subtracting the Gaussian-blurred images of adjacent scale layers within the same group to obtain a DoG pyramid, which can be expressed as follows:

[0018] D(x, y, sigma) = G(x, y, sigma2) * I(x, y) - G(x, y, sigma1) * I(x, y);

[0019] where I(x, y) is the input image, G(x, y, sigma) is the Gaussian kernel function, sigma1 and sigma2 are the standard deviations of two adjacent scales, and * represents convolution operation.

[0020] As a preferred embodiment of the present application, local extreme points are sought in the DoG pyramid as candidate key points, and each pixel point is compared with its 8 neighboring pixel points and the corresponding 9 pixel points of the upper and lower layers. If the pixel point is a local extreme point, it is retained as a candidate key point.

[0021] Sub-pixel level precise positioning is performed on the candidate key points to remove key points with low contrast and edge response.

[0022] The gradient direction histogram is calculated within the neighborhood of the key points, and a principal direction is assigned to each key point. The calculation formula of the gradient direction is as follows:

[0023]

[0024] where I x (x, y) and I y (x, y) are the horizontal and vertical gradients of the pixel point (x, y), respectively.

[0025] Within the key point field, the gradient direction histogram is calculated according to the principal direction to generate a 128-dimensional SIFT descriptor vector.

[0026] As a preferred embodiment of the present application, the Euclidean distance is used to calculate the similarity of SIFT feature vectors in the artwork image and the reference image. For two feature vectors v1 and v2, the calculation formula of the Euclidean distance is as follows:

[0027]

[0028] For each feature vector in the artwork image, the nearest feature vector in the reference image is found.

[0029] The two nearest neighbor feature vectors are found, and if the ratio of the nearest neighbor distance to the second nearest neighbor distance is less than a predetermined threshold, it is considered to be a reliable match.

[0030] The RANSAC algorithm is used to remove the wrong matching, the RANSAC algorithm estimates the geometric transformation parameters between images by randomly sampling matching point pairs, calculates the number of inliers, and selects the geometric transformation parameters with the largest number of inliers as the final result;

[0031] The ratio of the number of matched feature points to the number of feature points of the collected image is calculated and denoted as PB, the number of inliers calculated by the RANSAC algorithm is denoted as NS, and the total number of matching points is denoted as ZPS, and the comprehensive similarity score ZXP between the artwork image and the parameter image is calculated by the following formula:

[0032]

[0033] Where w1 and w2 are weight coefficients, and w1 + w2 = 1;

[0034] The maximum score is selected from the comprehensive similarity scores ZXP between the traced ceramic artwork image and each reference image as the similarity score between the traced ceramic artwork image and the reference image, and is marked as TXP.

[0035] As a preferred embodiment of the present application, the specific process of the spectrum analysis module comparing the collected spectrum data of the traced ceramic artwork with the reference spectrum data in the sample database is as follows:

[0036] Obtain the spectrum data of the traced ceramic artwork and the reference spectrum data in the sample database, and perform preprocessing operations on the spectrum data of the traced ceramic artwork and the reference spectrum data in the sample database;

[0037] Suppose the obtained spectrum data of the traced ceramic artwork is X sy , and the reference spectrum data in the sample database is X ck ={X ck，1 , X ck，2 , … X ck，n}, where n is the number of reference samples, and each spectrum data is a vector representing the spectrum intensity at different wavelengths;

[0038] Features are extracted from the preprocessed artwork spectrum data and reference spectrum data, and the integral of the spectrum intensity in a specific wavelength range in the spectrum is selected as the feature, assuming that the selected wavelength range is [λ a ,λ b ], then the calculation formula of the integral feature is as follows:

[0039] Where I(λ) is the spectrum intensity at wavelength λ.

[0040] As a preferred embodiment of the present application, in actual calculation, since the spectral data is discrete, the integral is approximated by summation:

[0041]

[0042] where λ i is a discrete wavelength point, and Δλ is a wavelength interval;

[0043] The similarity between the spectral data of the artwork and the reference spectral data is calculated by the following formula:

[0044]

[0045] where F sy represents the integral feature of the traced ceramic artwork, F ck,i represents the integral feature of the i-th reference sample, · represents the dot product, ‖·‖ represents the norm of the vector, the similarity value is between [-1, 1], and the larger the value, the higher the similarity;

[0046] From the similarity between the spectral data of the traced ceramic artwork and each reference spectral data, a value with the largest similarity is selected as the similarity between the spectral data of the traced ceramic artwork and the reference spectral data, and is marked as GPX.

[0047] As a preferred embodiment of the present application, the specific process of the process information comparison module calculating the similarity between the artwork process information and the reference process information is as follows:

[0048] Obtain the process information of the traced ceramic artwork and the reference process information in the sample database, the process information including firing temperature, temperature rising rate, glaze layer thickness, and body thickness, and perform preprocessing operation on the process information of the traced ceramic artwork and the reference process information in the sample database;

[0049] Generate a collection period, divide the collection period into multiple collection time periods, obtain the relative difference between the firing temperature of the traced ceramic artwork and the firing temperature of the reference sample, and take the arithmetic mean of the obtained relative difference between the firing temperature of the multiple traced ceramic artworks and the firing temperature of the reference sample, and mark the arithmetic mean of the relative difference between the firing temperature of the multiple traced ceramic artworks and the firing temperature of the reference sample as the average firing temperature difference PSW.

[0050] As a preferred embodiment of the present application, the relative difference between the temperature rising rate of the traced ceramic artwork and the temperature rising rate of the reference sample is obtained, and the arithmetic mean of the obtained relative differences between the temperature rising rates of the plurality of traced ceramic artworks and the temperature rising rate of the reference sample is calculated, and the arithmetic mean of the relative differences between the temperature rising rates of the plurality of traced ceramic artworks and the temperature rising rate of the reference sample is recorded as the average temperature rising rate difference PSS;

[0051] The relative difference between the glaze layer thickness of the traced ceramic artwork and the glaze layer thickness of the reference sample is obtained, and the arithmetic mean of the obtained relative differences between the glaze layer thicknesses of the plurality of traced ceramic artworks and the glaze layer thickness of the reference sample is calculated, and the arithmetic mean of the relative differences between the glaze layer thicknesses of the plurality of traced ceramic artworks and the glaze layer thickness of the reference sample is recorded as the average glaze layer thickness difference PYH;

[0052] The relative difference between the body thickness of the traced ceramic artwork and the body thickness of the reference sample is obtained, and the arithmetic mean of the obtained relative differences between the body thicknesses of the plurality of traced ceramic artworks and the body thickness of the reference sample is calculated, and the arithmetic mean of the relative differences between the body thicknesses of the plurality of traced ceramic artworks and the body thickness of the reference sample is recorded as the average body thickness difference PPH.

[0053] As a preferred embodiment of the present application, the average firing temperature difference PSW, the average temperature rising rate difference PSS, the average glaze layer thickness difference PYH and the average body thickness difference PPH are obtained, and the process similarity deviation index GXP between the process information of the traced ceramic artwork and the reference process information in the sample database is calculated by the following formula:

[0054]

[0055] Wherein f1, f2, f3 and f4 are preset proportional factor coefficients, f4>f3>f2>f1>0, and the larger the value of GXP is, the lower the similarity between the process information of the traced ceramic artwork and the process information in the sample database is;

[0056] From the process similarity deviation indexes between the process information of the traced ceramic artwork and each reference process information, the one with the smallest value is selected as the similarity between the process information of the traced ceramic artwork and the reference process information, and is marked as GXX.

[0057] As a preferred embodiment of the present application, the specific process of the trace identification module for comprehensive analysis and judgment of the traced ceramic artwork is as follows:

[0058] The similarity score TXP between the traced artwork image and the reference image, the similarity GPX between the traced artwork spectrum data and the reference spectrum data, and the similarity GXX between the traced artwork process information and the reference process information are obtained, and a traceability identification coefficient SYJ is calculated by the following formula:

[0059]

[0060] wherein g1, g2 and g3 are preset proportional factor coefficients, g3>g2>g1>0, the traceability identification coefficient SYJ is compared with a preset traceability identification coefficient threshold value:

[0061] If the traceability identification coefficient SYJ is less than the preset traceability identification coefficient threshold value, it indicates that the traced ceramic artwork is a fake;

[0062] If the traceability identification coefficient SYJ is greater than or equal to the preset traceability identification coefficient threshold value, it indicates that the traced ceramic artwork is a genuine product, and the traceability result of the traced ceramic artwork is given.

[0063] Compared with the prior art, the present application has the following advantages:

[0064] (1) In the present application, the image comparison module uses SIFT feature extraction and RANSAC algorithm to improve image quality through preprocessing, ensure accurate matching under various transformation conditions, and effectively remove false matches. The comprehensive similarity score is calculated by comprehensively matching the feature point proportion, the number of inliers and the total number of matching points, which provides an accurate and reliable quantitative standard for ceramic artwork tracing, improves the accuracy and robustness of image comparison, and effectively supports the fine tracing of artwork.

[0065] (2) In the present application, the spectrum analysis module first performs preprocessing operation on the spectrum data to ensure data accuracy, then focuses on the spectrum intensity integration of specific wavelength range to improve recognition accuracy, uses dot product and norm to calculate similarity, provides a general and sensitive comparison standard, and finally quantitatively evaluates by calculating the maximum similarity value GPX to realize rapid screening and accurate tracing, which improves the efficiency and accuracy of ceramic artwork tracing.

[0066] (3) In the present application, the process information comparison module accurately processes the data, comprehensively analyzes the key process parameters such as firing temperature, heating rate, glaze thickness and body thickness, uses average difference to quantify the difference and calculates the process similarity deviation index GXP, adopts adjustable proportional factor to enhance flexibility, and finally takes the minimum deviation index GXX as the score to realize rapid and accurate ceramic artwork tracing, which effectively improves the work efficiency and accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1System block diagram for the system of embodiment one in the present application;

[0068] Figure 2 System block diagram for the system of embodiment two in the present application;

[0069] Figure 3 Logic flow diagram for embodiment two in the present application. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application; obviously, the described embodiments are only a part of the embodiments of the present application, and not all the embodiments; based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0071] Embodiment one: as shown in the figure, the present application proposes a ceramic artwork traceability verification intelligent management system, which comprises a data acquisition module, an image comparison module, a spectrum analysis module, a process information comparison module and a traceability identification module; Figure 1 The data acquisition module is in one-way communication connection with the image comparison module, the spectrum analysis module and the process information comparison module, and the traceability identification module is in one-way communication connection with the image comparison module, the spectrum analysis module and the process information comparison module.

[0072] The data acquisition module acquires sample information of various ancient ceramics, and the sample information includes image data, spectrum data and process information; the sample information is stored in a sample database for comparison and reference, and image data, spectrum data and process information of the ceramic artwork to be traced are acquired.

[0073] The data acquisition module provides a solid foundation for the traceability verification of ceramic artworks by comprehensively acquiring image, spectrum and process information; the multi-dimensional data acquisition method significantly improves the accuracy and reliability of identification; systematic data collection supports historical and cultural research, promotes interdisciplinary cooperation, and optimizes management efficiency through automation process; in general, the module improves the comprehensiveness, accuracy and reliability of traceability verification, and effectively supports the healthy development of cultural products and fair competition in the market.

[0074] The image comparison module compares the collected image of the ceramic artwork to be traced with the reference image in the sample database, and calculates the similarity between the artwork image and the reference image.

[0075] The specific process of comparing the collected image of the ceramic artwork to be traced with the reference image in the sample database by the image comparison module is as follows:

[0076]

[0077] ​An image of a ceramic artwork to be traced is acquired and reference images in a sample database are preprocessed, the preprocessing operations including but not limited to image format conversion, size adjustment, denoising, contrast enhancement, color space conversion and geometric correction;

[0078] Feature vectors are extracted from the preprocessed artwork image and reference images, the specific steps being as follows:

[0079] A Gaussian difference pyramid is used to construct a scale space, the input image is first subjected to Gaussian filtering to generate a Gaussian pyramid, the standard deviation σ of the Gaussian kernel increasing with the increase of the pyramid level, the Gaussian blurred images of adjacent scale layers in the same group being subtracted to obtain a DoG pyramid, the expression formula being as follows:

[0080] D(x, y, σ) = G(x, y, σ2) * I(x, y) - G(x, y, σ1) * I(x, y);

[0081] where I(x,y) is the input image, representing the pixel value of the input image at (x,y), G(x,y,σ) is the Gaussian kernel function, σ1 and σ2 are the standard deviations of two adjacent scales, and * represents convolution operation;

[0082] Local extreme points in the DoG pyramid are searched for as candidate key points, each pixel point being compared with its 8 neighboring pixel points and the corresponding 9 pixel points of the upper and lower two layers, if the pixel point is a local extreme point, it is reserved as a candidate key point;

[0083] The candidate key points are subjected to sub-pixel level accurate positioning, and the key points with low contrast and edge response are removed;

[0084] The gradient direction histogram is calculated in the neighborhood of the key points, and a main direction is assigned to each key point, the calculation formula of the gradient direction being as follows:

[0085]

[0086] where I x (x,y) and I y (x,y) are the horizontal and vertical gradients of the pixel point (x,y), respectively;

[0087] In the key point field, according to the main direction, the gradient direction histogram is calculated to generate a 128-dimensional SIFT descriptor vector, once the key point and the main direction are determined, the neighborhood is normalized based on the direction (rotated to a unified angle), and then the neighborhood (usually 16x16 pixels) is divided into 4x4 sub-regions, and a direction histogram (usually 8 bins) is generated for each sub-region, so that each sub-region contributes 8 values, and a total of 4x4=16 sub-regions, thus forming a 128-dimensional (16*8=128) descriptor vector;

[0088] The vector describes the gradient direction and amplitude information of the pixels around the key point and is normalized;

[0089] The similarity of the SIFT feature vectors in the artwork image and the reference image is calculated using the Euclidean distance, and for two feature vectors v1 and v2, the calculation formula of the Euclidean distance is as follows:

[0090]

[0091] For each feature vector in the artwork image, find the nearest feature vector in the reference image;

[0092] Find the two nearest neighbor feature vectors, if the ratio of the nearest neighbor distance to the second nearest neighbor distance is less than a predetermined threshold, it is considered to be a reliable match;

[0093] Use the RANSAC algorithm to remove false matches, the RANSAC algorithm estimates the geometric transformation parameters between images by randomly sampling matching point pairs, and calculates the number of inliers, and selects the geometric transformation parameters with the most inliers as the final result, the RANSAC algorithm is an existing algorithm, and the calculation process will not be described in detail here;

[0094] Calculate the proportion of the number of matched feature points to the number of feature points in the collected image, and record it as PB, the number of inliers calculated by the RANSAC algorithm is recorded as NS, and the total number of matching points is recorded as ZPS, and the comprehensive similarity score ZXP between the artwork image and the parameter image is calculated by the following formula:

[0095]

[0096] Where w1 and w2 are weight coefficients, and w1+w2=1;

[0097] By analyzing historical data, the influence degree of the weight coefficients w1 and w2 on the similarity between the artwork image and the parameter image is understood, and according to the historical data, the specific values of w1 and w2 are determined to ensure that the comprehensive similarity score ZXP calculated can accurately reflect the similarity between the artwork image and the parameter image;

[0098] The maximum one is selected from the comprehensive similarity scores ZXP between the traced ceramic artwork image and each reference image as the similarity score between the traced ceramic artwork image and the reference image, and is marked as TXP;

[0099] The SIFT feature extraction and RANSAC algorithm are used by the image comparison module to improve image quality through preprocessing, ensure accurate matching under various transformation conditions, and effectively remove false matches. The comprehensive similarity score is calculated by comprehensively considering the matching feature point proportion, the number of inliers, and the total number of matching points, providing an accurate and reliable quantitative standard for ceramic artwork tracing, improving the accuracy and robustness of image comparison, and providing strong support for fine tracing of artworks.

[0100] The spectral analysis module compares the collected spectral data of the traced ceramic artwork with the reference spectral data in the sample database, and calculates the similarity between the artwork spectral data and the reference spectral data;

[0101] The specific process of the spectral analysis module comparing the collected spectral data of the traced ceramic artwork with the reference spectral data in the sample database is as follows:

[0102] The spectral data of the traced ceramic artwork and the reference spectral data in the sample database are obtained, and preprocessing operations are performed on the spectral data of the traced ceramic artwork and the reference spectral data in the sample database. The preprocessing operations include but are not limited to wavelength correction, smoothing processing, baseline correction, and standardization processing;

[0103] Suppose the obtained spectral data of the traced ceramic artwork is X sy , and the reference spectral data in the sample database is X ck ={X ck，1 , X ck，2 , … X ck，n}, where n is the number of reference samples, and each spectral data is a vector representing the spectral intensity at different wavelengths;

[0104] Features are extracted from the preprocessed artwork spectral data and reference spectral data, and the spectral intensity integral in a specific wavelength range in the spectrum is selected as the feature. Assuming that the selected wavelength range is [λ a ,λ b ], the calculation formula of the integral feature is as follows:

[0105] Where I(λ) is the spectral intensity at wavelength λ;

[0106] In actual calculation, since the spectral data is discrete, the integral is approximated by summation:

[0107]

[0108] where λ i is a discrete wavelength point, Δλ is a wavelength interval, and the selection of a suitable integral region needs to be determined according to the spectral characteristics of the to-be-detected substance, and the region where the spectral absorption peak or other significant features are located is usually selected;

[0109] The similarity between the spectral data of the artwork and the reference spectral data is calculated, and is represented by the following formula:

[0110]

[0111] where F sy represents the integral feature of the traced ceramic artwork, F ck,i represents the integral feature of the i-th reference sample, · represents the dot product, ‖·‖ represents the norm of the vector, and the similarity value is between [-1, 1], and the greater the value, the higher the similarity;

[0112] From the similarity between the spectral data of the traced ceramic artwork and each reference spectral data, a value of the greatest similarity is selected as the similarity between the spectral data of the traced ceramic artwork and the reference spectral data, and is marked as GPX;

[0113] The spectral analysis module first performs a preprocessing operation on the spectral data to ensure data accuracy, and then focuses on the spectral intensity integration of a specific wavelength range to improve recognition accuracy. The similarity is calculated using the dot product and the norm, providing a general and sensitive comparison standard. Finally, the maximum similarity value GPX is calculated for quantitative evaluation, realizing fast screening and accurate tracing, and improving the tracing efficiency and accuracy of ceramic artworks.

[0114] The process information comparison module compares the collected process information of the traced ceramic artwork with the reference process information in the sample database, and calculates the similarity between the process information of the artwork and the reference process information.

[0115] The specific process of the process information comparison module calculating the similarity between the process information of the artwork and the reference process information is as follows:

[0116] The process information of the traced ceramic artwork and the reference process information in the sample database are obtained, and the process information includes firing temperature, heating rate, glaze thickness, and body thickness. The process information of the traced ceramic artwork and the reference process information in the sample database are preprocessed, and the preprocessing operation includes but is not limited to data cleaning, filtering processing, and normalization processing.

[0117] A collection cycle is generated, the collection cycle is divided into a plurality of collection time periods, a relative difference between a firing temperature of the traced ceramic artwork and a firing temperature of the reference sample is obtained, and an arithmetic mean of the obtained relative differences between the firing temperatures of the plurality of traced ceramic artworks and the reference sample is calculated. The arithmetic mean of the relative differences between the firing temperatures of the plurality of traced ceramic artworks and the reference sample is denoted as an average firing temperature difference PSW;

[0118] A relative difference between a temperature rising rate of the traced ceramic artwork and a temperature rising rate of the reference sample is obtained, and an arithmetic mean of the obtained relative differences between the temperature rising rates of the plurality of traced ceramic artworks and the reference sample is calculated. The arithmetic mean of the relative differences between the temperature rising rates of the plurality of traced ceramic artworks and the reference sample is denoted as an average temperature rising rate difference PSS;

[0119] A relative difference between a glaze layer thickness of the traced ceramic artwork and a glaze layer thickness of the reference sample is obtained, and an arithmetic mean of the obtained relative differences between the glaze layer thicknesses of the plurality of traced ceramic artworks and the reference sample is calculated. The arithmetic mean of the relative differences between the glaze layer thicknesses of the plurality of traced ceramic artworks and the reference sample is denoted as an average glaze layer thickness difference PYH;

[0120] A relative difference between a body thickness of the traced ceramic artwork and a body thickness of the reference sample is obtained, and an arithmetic mean of the obtained relative differences between the body thicknesses of the plurality of traced ceramic artworks and the reference sample is calculated. The arithmetic mean of the relative differences between the body thicknesses of the plurality of traced ceramic artworks and the reference sample is denoted as an average body thickness difference PPH;

[0121] The average firing temperature difference PSW, the average temperature rising rate difference PSS, the average glaze layer thickness difference PYH, and the average body thickness difference PPH are obtained, and a process similarity deviation index GXP between the process information of the traced ceramic artwork and the reference process information in the sample database is calculated by the following formula:

[0122]

[0123] wherein f1, f2, f3, and f4 are preset proportional factor coefficients, f4>f3>f2>f1>0, and the greater the value of GXP, the lower the similarity between the process information of the traced ceramic artwork and the process information in the sample database;

[0124] One of the process similarity deviation indexes between the process information of the traced ceramic artwork and each reference process information with the smallest value is selected as the similarity between the process information of the traced ceramic artwork and the reference process information, and is denoted as GXX.

[0125] The data is accurately processed by the process information comparison module, key process parameters such as firing temperature, temperature rising rate, glaze thickness and body thickness are comprehensively analyzed, the average difference is used to quantify the difference and calculate the process similarity deviation index GXP, the adjustable proportion factor is used to enhance flexibility, finally the minimum deviation index GXX is used as the score, the rapid and accurate ceramic artwork traceability is realized, and the work efficiency and accuracy are effectively improved.

[0126] Embodiment two: the technical scheme of the embodiment of the application is different from that of embodiment one in that:

[0127] As shown in Figure 2 and Figure 3 , the traceability identification module integrates the results of image comparison, spectrum analysis and process comparison, and comprehensively analyzes and judges the traced ceramic artwork;

[0128] The specific process of the traceability identification module for comprehensively analyzing and judging the traced ceramic artwork is as follows:

[0129] After the similarity score TXP between the traced artwork image and the reference image, the similarity GPX between the traced artwork spectrum data and the reference spectrum data, and the similarity GXX between the traced artwork process information and the reference process information are obtained, the traceability identification coefficient SYJ is calculated by the following formula:

[0130]

[0131] Wherein g1, g2 and g3 are all preset proportion factor coefficients, g3>g2>g1>0, the traceability identification coefficient SYJ is compared with the preset traceability identification coefficient threshold value:

[0132] If the traceability identification coefficient SYJ is less than the preset traceability identification coefficient threshold value, it indicates that the traced ceramic artwork is a fake;

[0133] If the traceability identification coefficient SYJ is greater than or equal to the preset traceability identification coefficient threshold value, it indicates that the traced ceramic artwork is a genuine product, and the traceability result of the traced ceramic artwork is given, the traceability result includes place of origin, age, process technique, material source, historical circulation, author (if any) and repair condition (if any);

[0134] The provenance identification module integrates the results of image comparison, spectral analysis and process information comparison, evaluates the authenticity of ceramic artworks from multiple dimensions, calculates the provenance identification coefficient SYJ by scoring image similarity, spectral similarity and process information similarity, and flexibly allocating weights by using a preset proportion factor coefficient, ensures the accuracy and adaptability of the evaluation, compares the provenance identification coefficient SYJ with a preset threshold, provides clear true and false judgment standards, and outputs detailed provenance information including origin, age, process technique and the like for artworks confirmed as genuine, improves the historical value and academic research value thereof; the module has significant advantages in improving identification efficiency and accuracy due to comprehensive data processing capability, flexible weight mechanism and detailed provenance information, and is an important tool for cultural relic protection and authenticity identification.

[0135] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change according to the technical scheme and improvement concept of the present application within the technical range disclosed by the present application, which should be covered in the protection scope of the present application.

Claims

1. A smart management system for tracing and verifying the provenance of ceramic artworks, characterized in that, It includes a data acquisition module, an image comparison module, a spectral analysis module, a process information comparison module, and a traceability and identification module; The data acquisition module collects sample information of various ancient ceramics, stores the sample information in the sample database for comparison and reference, and collects image data, spectral data and process information of the traceable ceramic artworks; The image comparison module compares the collected images of traceable ceramic artworks with reference images in the sample database and calculates the similarity between the artwork images and the reference images. The spectral analysis module compares the collected spectral data of the traceable ceramic artworks with the reference spectral data in the sample database and calculates the similarity between the artwork spectral data and the reference spectral data. The process information comparison module compares the collected process information of the traceable ceramic artwork with the reference process information in the sample database and calculates the similarity between the artwork's process information and the reference process information. The traceability and authentication module integrates the results of image comparison, spectral analysis, and process comparison to conduct a comprehensive analysis and judgment on the traceable ceramic artwork.

2. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 1, characterized in that, The specific process by which the image comparison module compares the acquired images of the traceable ceramic artworks with reference images in the sample database is as follows: Obtain the images of the traceable ceramic artworks and reference images from the sample database, and perform preprocessing operations on the images of the traceable ceramic artworks and reference images from the sample database; The feature vectors are extracted from the preprocessed artwork image and the reference image. The specific steps are as follows: The DoG pyramid is constructed using the difference of Gaussian (DoG) pyramid. First, the input image is Gaussian filtered to generate the pyramid. The standard deviation σ of the Gaussian kernel increases with the pyramid level. The Gaussian blurred images of adjacent scale levels within the same group are subtracted to obtain the DoG pyramid, expressed by the following formula: D(x,y,σ)=G(x,y,σ2)*I(x,y)-G(x,y,σ1)*I(x,y); Where I(x,y) is the input image, G(x,y,σ) is the Gaussian kernel function, σ1 and σ2 are the standard deviations of two adjacent scales, and * indicates convolution operation.

3. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 2, characterized in that, In the DoG pyramid, local extrema are searched as candidate keypoints. Each pixel is compared with its 8 neighboring pixels and the 9 corresponding pixels in the two layers above and below. If the pixel is a local extrema, it is retained as a candidate keypoint. Perform subpixel-level precise localization of candidate keypoints and remove keypoints with low contrast and edge response; Calculate the gradient direction histogram within the neighborhood of each keypoint, and assign a principal direction to each keypoint. The formula for calculating the gradient direction is as follows: Where I x (x,y) and I y (x,y) are the horizontal and vertical gradients of the pixel (x,y), respectively; Within the keypoint domain, a gradient direction histogram is calculated based on the principal direction, generating a 128-dimensional SIFT descriptor vector.

4. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 3, characterized in that, The similarity between SIFT feature vectors in an artwork image and a reference image is calculated using Euclidean distance. For two feature vectors v1 and v2, the formula for calculating the Euclidean distance is as follows: For each feature vector in the artwork image, find the feature vector in the reference image that is closest to it; Find two nearest neighbor feature vectors. If the ratio of the nearest neighbor distance to the second nearest neighbor distance is less than a preset threshold, then it is considered a reliable match. The RANSAC algorithm is used to remove erroneous matches. The RANSAC algorithm estimates the geometric transformation parameters between images by randomly sampling matching point pairs, calculates the number of inliers, and selects the geometric transformation parameter with the most inliers as the final result. The proportion of matched feature points to the total number of feature points in the acquired image is calculated and denoted as PB. The number of interior points calculated using the RANSAC algorithm is denoted as NS, and the total number of matched points is denoted as ZPS. The comprehensive similarity score ZXP between the artwork image and the parametric image is calculated using the following formula: Where w1 and w2 are weighting coefficients, and w1 + w2 = 1; The highest score (ZXP) among the traced ceramic artwork images and each reference image is selected as the similarity score between the traced ceramic artwork image and the reference images, and it is marked as TXP.

5. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 1, characterized in that, The specific process by which the spectral analysis module compares the collected spectral data of the traceable ceramic artwork with the reference spectral data in the sample database is as follows: Obtain the spectral data of the traceable ceramic artwork and the reference spectral data in the sample database, and perform preprocessing operations on the spectral data of the traceable ceramic artwork and the reference spectral data in the sample database; Assuming the obtained spectral data of the traceable ceramic artwork is X sy The reference spectral data in the sample database is X. ck ={X ck，1 X ck，2 , ...X ck，n }, where n is the number of reference samples, and each spectral data is a vector representing the spectral intensity at different wavelengths; Features are extracted from the preprocessed spectral data of the artwork and the reference spectral data. The spectral intensity integral within a specific wavelength range is selected as the feature. Assume the selected wavelength range is [λ]. a ,λ b The formula for calculating the integral characteristic is as follows: Where I(λ) is the spectral intensity at wavelength λ.

6. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 5, characterized in that, In practical calculations, since spectral data is discrete, the integral is approximated by summation: Where λ i These are discrete wavelength points, and Δλ is the wavelength interval. The similarity between the artwork's spectral data and the reference spectral data is calculated using the following formula: Where F sy F represents the integral characteristic of the traced ceramic artwork. ck,i Let represent the integral feature of the i-th reference sample, · represent the dot product, and ||·| represent the norm of the vector. The similarity value is between [-1, 1], and the larger the value, the higher the similarity. Select the highest similarity score from the similarity scores between the traced ceramic artwork spectral data and each reference spectral data, and label it as GPX.

7. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 1, characterized in that, The specific process by which the process information comparison module calculates the similarity between the artwork's process information and the reference process information is as follows: Obtain the process information of the traceable ceramic artwork and the reference process information in the sample database. The process information includes firing temperature, heating rate, glaze thickness and body thickness. Perform preprocessing operations on the process information of the traceable ceramic artwork and the reference process information in the sample database. A collection period is generated, which is divided into multiple collection time periods. The relative difference between the firing temperature of the traceable ceramic artwork and the firing temperature of the reference sample is obtained. The arithmetic mean of the relative differences between the firing temperatures of the multiple traceable ceramic artworks and the firing temperatures of the reference sample is calculated. The arithmetic mean of the relative differences between the firing temperatures of the multiple traceable ceramic artworks and the firing temperatures of the reference sample is recorded as the average firing temperature difference (PSW).

8. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 7, characterized in that, The relative difference between the heating rate of the traceable ceramic artwork and the heating rate of the reference sample is obtained, and the arithmetic mean of the relative differences between the heating rates of multiple traceable ceramic artworks and the heating rates of the reference sample is calculated. The arithmetic mean of the relative differences between the heating rates of multiple traceable ceramic artworks and the heating rates of the reference sample is recorded as the average heating rate difference (PSS). Obtain the relative difference between the glaze thickness of the traceable ceramic artwork and the glaze thickness of the reference sample, and calculate the arithmetic mean of the relative differences between the glaze thickness of multiple traceable ceramic artworks and the glaze thickness of the reference sample. The arithmetic mean of the relative differences between the glaze thickness of multiple traceable ceramic artworks and the glaze thickness of the reference sample is recorded as the average glaze thickness difference PYH. Obtain the relative difference between the body thickness of the traceable ceramic artwork and the body thickness of the reference sample, and calculate the arithmetic mean of the relative differences between the body thickness of multiple traceable ceramic artworks and the body thickness of the reference sample. The arithmetic mean of the relative differences between the body thickness of multiple traceable ceramic artworks and the body thickness of the reference sample is recorded as the average body thickness difference PPH.

9. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 8, characterized in that, The average firing temperature difference (PSW), average heating rate difference (PSS), average glaze thickness difference (PYH), and average body thickness difference (PPH) are obtained. The process similarity deviation index (GXP) between the process information of the traceable ceramic artwork and the reference process information in the sample database is calculated using the following formula: Among them, f1, f2, f3 and f4 are all preset scaling factor coefficients, f4>f3>f2>f1>0. The larger the GXP value, the lower the similarity between the traceable ceramic artwork's craft information and the craft information in the sample database. Select the smallest value from the process similarity deviation index between the traced ceramic artwork's process information and each reference process information as the similarity between the traced ceramic artwork's process information and the reference process information, and label it as GXX.

10. The intelligent management system for tracing and verifying the provenance of ceramic artworks according to claim 1, characterized in that, The specific process by which the provenance identification module conducts a comprehensive analysis and judgment on the traceable ceramic artwork is as follows: The similarity score TXP between the traced artwork image and the reference image, the similarity score GPX between the traced artwork spectral data and the reference spectral data, and the similarity score GXX between the traced artwork's craftsmanship information and the reference craftsmanship information are obtained. The traceability identification coefficient SYJ is then calculated using the following formula: Where g1, g2, and g3 are all preset scaling factor coefficients, g3 > g2 > g1 > 0. The traceability identification coefficient SYJ is compared with the preset traceability identification coefficient threshold: If the traceability identification coefficient SYJ is less than the preset traceability identification coefficient threshold, it indicates that the traced ceramic artwork is a fake. If the traceability identification coefficient SYJ is greater than or equal to the preset traceability identification coefficient threshold, it indicates that the traceable ceramic artwork is genuine, and the traceability result of the traceable ceramic artwork is given.

Images