Mining and metallurgy product intelligent identification method and system based on dynamic knowledge base

Abstract

The invention provides a mining and metallurgy product intelligent identification method and system based on a dynamic knowledge base, and relates to the technical field of mining and metallurgy product customs supervision. Aiming at the problem of low mineral identification accuracy and efficiency in the prior art, the method comprises the following steps: acquiring mineral attribute information, classification standard rules, risk characteristics and multi-source public data, updating the multi-source public data in real time, and carrying out information association to form a multi-dimensional dynamic knowledge base; all-element spectral data and crystal structure diffraction data of the to-be-detected mining and metallurgy product are collected on site, and element information and phase information are obtained; performing multi-modal data fusion on the element information, the phase information and the mineral industrial grade or element abundance, and determining a final mineral type name; and according to the quality standard, the classification rule, the forbidden and limited list and the solid waste characteristics in the knowledge base, performing compliance checking, classification judgment and risk screening, and outputting a mineral identification result, a risk label and a compliance suggestion. And an efficient and accurate identification method is provided for customs on-site supervision scenes.

Description

Technical Field

[0001] This invention relates to the field of customs supervision technology for mining and metallurgical products, and in particular to an intelligent identification method and system for mining and metallurgical products based on a dynamic knowledge base. Background Technology

[0002] Mineral and metallurgical product identification is a crucial and fundamental issue in fields such as geology, resource exploration, geotechnical investigation, and metallurgical engineering, and it is also an important issue in customs mineral supervision. Improving the speed and intelligence of mineral analysis and identification is an important technological research direction in customs supervision. It is of great significance for improving the efficiency of identifying unknown goods, reducing knowledge barriers for practitioners, quickly screening cargo risks, improving customs clearance efficiency, and enhancing the level of smart customs.

[0003] Customs supervision primarily deals with resource-based minerals that have undergone processing such as crushing and flotation. These minerals have complex compositions, and due to the non-correlation between mineral names and elemental contents, and the comprehensive correlation between mineral type names and their elemental composition, phase composition, elemental value, and classification rules, traditional image recognition and other technologies cannot directly determine mineral names and types. In the field of customs supervision, current mineral identification, determination, and supervision heavily rely on comprehensive judgment by laboratory technicians. This process is lengthy, costly, and unable to correlate mineral type names with prohibited or restricted categories, solid waste standards, classification rules, and product specifications in real time, resulting in poor identification accuracy and low efficiency.

[0004] Therefore, developing a method and system for intelligent identification and risk warning of mineral and metallurgical products based on a dynamic knowledge base is of great significance for reducing the reliance on manual labor and improving the efficiency and intelligence of mineral identification and risk warning in customs supervision. Summary of the Invention

[0005] To address the problems of high reliance on manual identification, low identification efficiency, and low level of intelligence in existing technologies, this invention proposes an intelligent identification method for mining and metallurgical products based on a dynamic knowledge base, specifically including the following steps:

[0006] S1. Obtain mineral attribute information, classification standard rules, risk characteristics and multi-source public data, and update the multi-source public data in real time. Link the obtained information with the preset mineral main elements as the core to form a multi-dimensional dynamic knowledge base.

[0007] S2. By collecting full-element spectral data and / or crystal structure diffraction data of the mineral and metallurgical products to be tested at the regulatory site, the elemental information and / or phase information of the mineral and metallurgical products to be tested are obtained, and the mineral type is determined in a multi-dimensional dynamic knowledge base based on the elemental information and phase information.

[0008] S3. Based on a multidimensional dynamic knowledge base, retrieve the corresponding mineral industrial grade or element abundance according to the mineral type, and perform weight allocation and multimodal data fusion analysis on the element information, phase information and mineral industrial grade or element abundance to determine the final mineral type name of the mineral and metallurgical product to be tested.

[0009] S4. Based on the final mineral type name, element information, and phase information, and in conjunction with the quality standards, classification rules, prohibited and restricted lists, and solid waste characteristics in the multi-dimensional dynamic knowledge base, conduct compliance verification, classification judgment, risk screening, and early warning, and output mineral identification results, risk labels, and compliance recommendations.

[0010] Furthermore, in S1, the multi-source publicly available data includes data from international organizations, national official data, academic resource data, and market intelligence data. Real-time updates to the multi-source publicly available data include: monitoring data from international organizations, national official data, academic resource data, and market intelligence data through web crawlers; and combining customs node detection data from federated learning and sharing of national official data to iterate the multi-source publicly available data in real time.

[0011] Furthermore, in step S2, the elemental information and / or phase information of the mineral product to be tested are obtained by collecting full-element spectral data and / or crystal structure diffraction data of the product at the regulatory site. This includes: collecting full-element spectral data of the product using a first portable device, standardizing the full-element spectral data, matching the standardized data with a mineral database in a multi-dimensional dynamic knowledge base to obtain elemental information, which includes elemental composition types and elemental content; collecting crystal structure diffraction data of the product using a second portable device, extracting features from the crystal structure diffraction data, importing the feature extraction results into a phase combination card library in a multi-dimensional dynamic knowledge base corresponding to the predicted mineral type, performing full-spectrum fitting and refinement to obtain phase content and goodness-of-fit information.

[0012] Furthermore, the first portable device includes a portable X-ray fluorescence spectrometer or a laser-induced breakdown spectrometer; the second portable device includes a portable X-ray diffraction spectrometer.

[0013] Furthermore, in step S3, weight allocation and multimodal data fusion analysis are performed on the elemental information, phase information, and mineral industrial grade or elemental abundance. This includes: extracting prominent factors, ranking results, and elemental matching degrees from the elemental information; extracting the phase fitting goodness from the phase information; and extracting the coupling matching degree between elemental information and phase information. In the calculation of prominent factors, for conventionally mined minerals, the mineral industrial grade is used as the key weighting term, and for rare minerals, elemental abundance is used as the key weighting term. The priority order of elemental information matching is determined based on the ranking results of prominent factors. The priority order of phase information fitting is determined based on the ranking results of prominent factors. Multimodal data fusion calculation is performed by integrating elemental matching degree, phase fitting goodness, coupling matching degree, ranking results of prominent factors, and key weighting terms.

[0014] Furthermore, the multimodal data fusion calculation includes: matching the elemental information of the mineral product to be tested with the elemental information of the top three main element minerals in the mineral database to obtain the elemental information matching degree, and converting the matching degree to obtain the first contribution value of the elemental matching dimension; converting the phase fitting goodness to obtain the second contribution value of the phase matching dimension; matching the elemental information of the mineral product to be tested with the elemental information obtained by converting the phase content, and converting the matching result to obtain the third contribution value of the coupling matching of elemental information and phase information; weighted summing the first contribution value, the second contribution value and the third contribution value to obtain the comprehensive score of the mineral product to be tested and each candidate mineral; and selecting the candidate mineral with the highest comprehensive score as the final mineral type name.

[0015] Furthermore, in S4, the compliance verification includes: associating the final mineral type name, element information, and phase information with the quality standards in a multi-dimensional dynamic knowledge base, and verifying whether the element content, mineral industrial grade, and toxic element limits meet the requirements.

[0016] Furthermore, the risk screening includes: matching the final mineral type name, element information, and phase information with the solid waste characteristics, classification rules, and prohibited and restricted lists in the multi-dimensional dynamic knowledge base; analyzing the consistency between the final mineral type name, element information, and phase information and the data in the multi-dimensional dynamic knowledge base; and forming risk labels and compliance recommendations.

[0017] Furthermore, after outputting mineral identification results, risk labels, and compliance recommendations, the process also includes: verifying the output mineral identification results; supplementing the verified accurate data into the multidimensional dynamic knowledge base; marking correction information based on the verified deviation results and optimizing the parameters in the multimodal data fusion analysis process; and regularly incorporating elemental and phase information of new mining and metallurgical products and solid waste to dynamically update the multidimensional dynamic knowledge base.

[0018] This invention also provides an intelligent identification system for mining and metallurgical products based on a dynamic knowledge base. The system is used in any of the above-described intelligent identification methods for mining and metallurgical products based on a dynamic knowledge base, and the system includes:

[0019] The knowledge base construction module is used to acquire mineral attribute information, classification standard rules, risk characteristics and multi-source public data, and to update the multi-source public data in real time. The acquired information is associated with preset mineral main elements as the core to form a multi-dimensional dynamic knowledge base.

[0020] The data acquisition module is used to collect full-element spectral data and / or crystal structure diffraction data of the mineral and metallurgical products to be tested at the regulatory site, obtain elemental information and / or phase information of the mineral and metallurgical products to be tested, and determine the mineral type in a multi-dimensional dynamic knowledge base based on the elemental information and phase information.

[0021] The analysis module is used to retrieve the corresponding mineral industrial grade or element abundance based on the mineral type according to the multidimensional dynamic knowledge base, and to perform weight allocation and multimodal data fusion analysis on the element information, phase information and mineral industrial grade or element abundance to determine the final mineral type name of the mineral and metallurgical product to be tested.

[0022] The output module is used to perform compliance verification, classification judgment, risk screening and early warning based on the final mineral type name, element information and phase information, and to associate with quality standards, classification rules, prohibited and restricted lists and solid waste characteristics in a multi-dimensional dynamic knowledge base, and output mineral identification results, risk labels and compliance suggestions.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0024] By updating publicly available data from multiple sources in real time, the timeliness and coverage of the knowledge base are ensured. Elemental and phase information of the mineral and metallurgical products to be tested is integrated, along with the industrial grade or elemental abundance of minerals suitable for the mineral type in a multi-dimensional dynamic knowledge base. Targeted weight allocation and multi-modal data fusion analysis are then performed. The weight allocation design dynamically focuses on core judgment criteria based on mineral characteristics, and multi-modal fusion achieves complementary verification of multi-dimensional data, significantly improving the accuracy of final mineral type identification. This also replaces the traditional model relying on the comprehensive judgment of professional personnel's knowledge and experience, reducing human error and the professional threshold. These technical effects contribute to improved identification accuracy.

[0025] Furthermore, by collecting samples on-site at the regulatory site, there is no need to send samples to a laboratory for testing, solving the problems of long testing processes and low customs clearance efficiency in traditional methods and significantly shortening the testing cycle. Based on the identification results and their association with quality standards, prohibited and restricted lists, compliance verification and classification are automatically completed, which can avoid regulatory risks caused by difficulties in detecting concealment or misreporting, and improve identification efficiency. Attached Figure Description

[0026] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a flowchart of an intelligent identification method for mining and metallurgical products based on a dynamic knowledge base, provided by an embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of the structure of an intelligent identification system for mining and metallurgical products based on a dynamic knowledge base, provided in an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0030] The specific embodiments of the present invention will be described below.

[0031] To address the issues of poor accuracy and low efficiency in existing technologies, this invention acquires mineral attribute information, classification standards and rules, risk characteristics, and multi-source publicly available data, and updates this data in real time. The acquired information is then linked around a pre-defined mineral principal element to form a multi-dimensional dynamic knowledge base. By collecting full-element spectral data and crystal structure diffraction data of the mineral product to be tested at the regulatory site, elemental and phase information of the product is obtained. Weighting and multi-modal data fusion analysis are performed on the elemental information, phase information, and mineral industrial grade or elemental abundance in the multi-dimensional dynamic knowledge base to determine the final mineral type name of the product. Based on the final mineral type name, elemental information, and phase information, the invention links quality standards, classification rules, prohibited and restricted lists, and solid waste characteristics in the multi-dimensional dynamic knowledge base to perform compliance verification, classification judgment, and risk screening, outputting mineral identification results, risk labels, and compliance recommendations. The identification method provided by this invention has high accuracy and efficiency.

[0032] This invention provides a method for intelligent identification of mining and metallurgical products based on a dynamic knowledge base. Figure 1This is a flowchart of an intelligent identification method for mining and metallurgical products based on a dynamic knowledge base, provided by an embodiment of the present invention. Figure 1 As shown, the specific steps include the following:

[0033] S1. Obtain mineral attribute information, classification standard rules, risk characteristics, and multi-source public data, and update the multi-source public data in real time. Link the obtained information with the preset mineral main elements as the core to form a multi-dimensional dynamic knowledge base.

[0034] Mineral attribute information refers to the basic characteristics and physicochemical properties of minerals, including mineral category, common mineral names, chemical formulas, crystal structures, CAS numbers, common associated minerals, origin, elemental composition and common content ranges, phase composition and common content ranges, atomic weights, and elemental abundances of each mineral element. Classification standards and rules refer to the regulatory system related to mineral classification, quality assessment, and supervision, covering mineral category division, industrial grade, boundary grade standards, customs tariff classification rules, and HS code correspondences. Mineral categories include non-ferrous metals, ferrous metals, and rare earth minerals. Risk characteristics refer to the regulatory risk indicators associated with minerals, including solid waste characteristics (e.g., ferroalloy phase), smuggling-prone attributes (e.g., rare earth oxides), misdeclaration characteristics (e.g., magnesite misdeclared as dolomite), toxic element thresholds (e.g., As > 0.5%), and limits on dual-use elements. Preset mineral principal elements refer to the core elements in a mineral with high smelting value that determine its mineral category, serving as the core link in information association. The multidimensional dynamic knowledge base is a comprehensive database that integrates mineral attributes, classification rules, risk characteristics, and publicly available data with the main element as its core. It has the ability to update data in real time and supports intelligent identification and regulatory decision-making. It includes, but is not limited to, information such as mineral attributes, risk goods attributes, mineral standards and regulations, and solid waste cases. Among them, risk goods attributes include prohibited and restricted goods involving environmental elements, dual-use items, and solid waste. Mineral standards and regulations include mineral quality standards, classification rules, and regulatory requirements.

[0035] Multi-source publicly available data includes data from international organizations, national official data, academic resource data, and market intelligence data. Real-time updates to multi-source publicly available data include: monitoring data from international organizations, national official data, academic resource data, and market intelligence data through web crawlers; and combining customs node detection data from federated learning and sharing of national official data to iterate the multi-source publicly available data in real time.

[0036] The system comprehensively acquires four core information categories: mineral attribute information, classification standard rules, risk characteristics, and publicly available data from multiple sources. It utilizes web crawlers to automatically monitor the dynamics of publicly available data and combines this with federated learning to share customs node detection data, enabling real-time iteration of the multi-source publicly available data and avoiding data lag. The acquired information is then correlated around preset mineral principal elements; for example, the principal element Zn is linked to typical elemental content ranges, phase combinations, solid waste risk characteristics, and HS codes of zinc ore. All correlated information is integrated to form a multi-dimensional dynamic knowledge base encompassing data query, rule matching, and risk assessment functions, supporting subsequent comparative analysis of on-site detection data.

[0037] S2. By collecting full-element spectral data and / or crystal structure diffraction data of the mineral and metallurgical products to be tested at the regulatory site, the elemental information and / or phase information of the mineral and metallurgical products to be tested are obtained. Based on the elemental information and phase information, the mineral type is determined in a multi-dimensional dynamic knowledge base.

[0038] Collecting only full-element spectral data allows for rapid acquisition of elemental information, suitable for preliminary screening and quick on-site verification of product names, such as determining whether prohibited or restricted elements are present, or confirming the presence of target elements. Collecting only crystal structure diffraction data is suitable for scenarios where the elemental composition is known but phase verification is required, such as distinguishing between natural minerals and industrial solid waste. Combining both types of data is the core acquisition method, ensuring the accuracy of subsequent identification by obtaining both elemental and phase data. Using elemental and phase information as dual core data inputs, and relying on pre-stored mineral-element-phase association rules in a multi-dimensional dynamic knowledge base, the corresponding mineral type is selected.

[0039] Specifically, the process includes: acquiring full-element spectral data of the mineral product to be tested using a first portable device, standardizing the full-element spectral data, matching the standardized data with a mineral database in a multi-dimensional dynamic knowledge base to obtain elemental information, including elemental composition and elemental content; acquiring crystal structure diffraction data of the mineral product to be tested using a second portable device, extracting features from the crystal structure diffraction data, importing the feature extraction results into a phase combination card library in the multi-dimensional dynamic knowledge base corresponding to the predicted mineral type, performing full-spectrum fitting and refinement to obtain phase content and goodness-of-fit information. The first portable device includes a portable X-ray fluorescence spectrometer or a laser-induced breakdown spectrometer; the second portable device includes a portable X-ray diffraction spectrometer.

[0040] The first portable device is an on-site testing device for collecting mineral elemental spectral data, including a portable X-ray fluorescence spectrometer (XRF) or laser-induced breakdown spectrometer (LIBS). Its core function is to rapidly acquire full elemental composition information of the sample. The second portable device is an on-site testing device for collecting mineral crystal structure data, namely a portable X-ray diffraction spectrometer (XRD). Its core function is to acquire diffraction characteristic data of the sample's phase structure. Full elemental spectral data, obtained through scanning with the first portable device, reflects the raw data of all elements contained in the tested mineral product and their spectral characteristics, forming the basis for qualitative and quantitative elemental analysis. The mineral database is a sub-base of a multidimensional dynamic knowledge base, storing data such as typical elemental composition, content distribution range, and main element characteristics of various minerals for elemental data matching and comparison. Elemental information is the core result parsed from the full elemental spectral data, including the elemental types in the tested product, the mass percentage of each element, and the predicted mineral type obtained through matching. Crystal structure diffraction data, obtained through scanning with the second portable device, is raw spectral data reflecting the diffraction peak positions, peak intensities, and full width at half maximum (FWHM) characteristics of the mineral crystals in the tested product. The phase combination card library is a sub-library of the multidimensional dynamic knowledge base. It is a collection of phase combinations established for different mineral types. Each combination card contains structural information of the main mineral and common associated minerals of the corresponding mineral type, as well as elemental composition correlation data. Phase information refers to the final result of phase analysis, including the types of phases in the product to be tested, the content of each phase, and the reliability based on goodness-of-fit transformation.

[0041] S3. Based on a multidimensional dynamic knowledge base, retrieve the corresponding mineral industrial grade or element abundance according to the mineral type, and perform weight allocation and multimodal data fusion analysis on element information, phase information, and mineral industrial grade or element abundance to determine the final mineral type name of the mineral and metallurgical product to be tested.

[0042] First, three types of core data were collected: elemental information acquired and standardized by a first portable device; phase information collected by a second portable device and refined through full-spectrum fitting; and corresponding data retrieved from a multi-dimensional dynamic knowledge base based on mineral type, ensuring consistent data units and no missing values. Then, dynamic weight allocation was performed, clarifying the importance and priority of each data dimension based on data reliability and mineral type. Next, multi-modal data fusion analysis was conducted using intelligent algorithms to verify the consistency between elemental and phase information, calculate the overall confidence level, and eliminate contradictory data. Finally, the result with the highest overall confidence level and consistent with the essential mineral properties was selected as the final mineral type name for the tested mineral product.

[0043] By fusing multi-dimensional data from elements, phases, and databases, and combining dynamic weights to adapt to different mineral types, the accuracy of mineral identification has been significantly improved. Simultaneously, intelligent algorithms replace manual comprehensive judgment, reducing reliance on the professional knowledge and experience of practitioners. This approach is applicable to both common industrial minerals and rare mineral identification scenarios, providing customs on-site supervision with a rapid and accurate basis for mineral name determination, avoiding risks such as concealment and false declarations, and laying a solid foundation for subsequent classification, compliance assessment, and risk warning.

[0044] Specifically, weighting and multimodal data fusion analysis are performed on elemental information, phase information, and mineral industrial grade or elemental abundance. This includes: extracting prominent factors, ranking results, and elemental matching degrees from elemental information; extracting phase fit goodness from phase information; and extracting the coupling matching degree between elemental and phase information. In the calculation of prominent factors, for conventionally mined minerals, mineral industrial grade is used as the key weighting term, and for rare minerals, elemental abundance is used as the key weighting term. The priority order of elemental information matching is determined based on the ranking results of prominent factors. The priority order of phase information fitting is determined based on the ranking results of prominent factors. Multimodal data fusion calculation is performed by integrating elemental matching degree, phase fit goodness, coupling matching degree, ranking results of prominent factors, and key weighting terms.

[0045] The salient factor refers to the result of dividing the element content value by the corresponding mineral element's industrial grade or abundance. It is used to determine the core contribution of an element to mineral naming. The ranking result is the order of elements according to the salient factor from largest to smallest. The elemental composition of the tested mineral product is matched with the principal element mineral data ranked first by the salient factor in the database to obtain the elemental matching degree. The elemental matching degree is represented by a confidence level in the [0,100%] interval, derived from the distance between the matrix data of the mineral to be matched and the matched mineral, reflecting the degree of fit at the elemental level. Phase fit goodness is a core quantitative indicator measuring the degree of agreement between the crystal structure diffraction data of the tested mineral product and the phase combination cards of the target mineral in the multidimensional dynamic knowledge base. It is a key input for the phase dimension in multimodal data fusion analysis, directly determining the reliability of phase information and the accuracy of mineral identification. Mineral industrial grade refers to the minimum elemental content standard for industrial mining of a mineral, and is the core judgment benchmark for the smelting value of conventional industrial minerals. Elemental abundance refers to the average abundance of an element in the Earth (i.e., the Clarke value), used to determine the value and category of rare minerals without a clearly defined industrial grade. Conventional industrially mined minerals are those with clearly defined industrial grade standards and capable of large-scale mining and smelting, such as copper ore, iron ore, and copper concentrate. Rare minerals refer to minerals with high scarcity, such as gold and silver ore, and their value is judged based on elemental abundance.

[0046] The multimodal data fusion calculation includes: matching the elemental information of the mineral product to be tested with the elemental information of the top three principal elements of the minerals in the mineral database to obtain the elemental information matching degree, and converting the matching degree to obtain the first contribution value of the elemental matching dimension; converting the phase fit goodness to obtain the second contribution value of the phase matching dimension; matching the elemental information of the mineral product to be tested with the elemental information obtained by converting the phase content, and converting the matching result to obtain the third contribution value of the coupling matching of elemental information and phase information; weighted summing the first contribution value, the second contribution value, and the third contribution value to obtain the comprehensive score of the mineral product to be tested and each candidate mineral; and selecting the candidate mineral with the highest comprehensive score as the final mineral type name.

[0047] The first contribution value refers to the quantitative index formed by matching the measured elemental information of the tested mineral product with the typical elemental information of the candidate minerals corresponding to the top three principal elements of the prominent factors, obtaining the elemental information matching degree, and then standardizing and converting it. For example, the percentage matching degree is uniformly converted into a contribution value of 0-100 points. The second contribution value refers to the quantitative index formed by standardizing and converting the goodness of fit obtained from phase analysis. For example, the conversion logic is to convert the goodness of fit into a confidence level of [0,100] through a piecewise linear function. The third contribution value refers to the quantitative index formed by matching the consistency between the measured elemental information of the tested mineral and the theoretical elemental information converted from phase content. The phase content conversion elemental information includes the content of each phase obtained through phase analysis. Combined with the chemistry of each phase, the theoretical elemental content can be calculated. The matching result of the measured elemental content (elemental information of the tested mineral product) and the theoretical elemental content (elemental information obtained from phase content conversion) is converted into a quantitative value of 0-100 points.

[0048] Based on the priority weighting of matching elements and substances according to the ranking of prominent factors, the top 3 main minerals with prominent factors are selected for matching, which significantly reduces the matching range and improves calculation and matching efficiency. Furthermore, high-precision prediction is achieved through multi-dimensional data fusion, including elemental, phase, and element-phase coupling matching information.

[0049] S4. Based on the final mineral type name, element information, and phase information, and in conjunction with the quality standards, classification rules, prohibited and restricted lists, and solid waste characteristics in the multi-dimensional dynamic knowledge base, conduct compliance verification, classification judgment, risk screening, and early warning, and output mineral identification results, risk labels, and compliance recommendations.

[0050] Quality standards are a core sub-item in the multidimensional dynamic knowledge base, including mandatory or industry-specific standards such as mineral industrial grade, boundary grade, concentrate quality requirements, and limits on toxic and hazardous elements. Classification rules are a regulatory sub-item in the multidimensional dynamic knowledge base, including HS code correspondence rules, classification priorities, and purity-related standards. Prohibited and restricted lists are a risk management sub-item in the multidimensional dynamic knowledge base, including CITES Appendix species, international sanctions lists, dual-use item control catalogs, and environmentally prohibited and restricted goods lists. Solid waste characteristics refer to the criteria for solid waste identification in the multidimensional dynamic knowledge base, including sensitive phases (such as cement phase, metallic phase, and fritillary olivine), abnormal phase combinations (such as the presence of metallurgical slag characteristic phases in natural minerals), and inconsistencies in element-matter matching. Compliance verification is the process of verifying whether the mineral under test meets regulatory requirements such as quality standards and prohibited and restricted lists. Classification determination is the process of determining the corresponding HS code, tax rate, and classification priority of a mineral based on classification rules. Risk labels are intuitive identifiers generated based on risk screening results, such as solid waste risk, return risk, dual-use item control, and underreporting risk. Compliance recommendations are specific action guidelines provided based on compliance checks and risk screening results.

[0051] Specifically, compliance verification includes: linking the final mineral type name, element information, and phase information to quality standards in a multi-dimensional dynamic knowledge base, and verifying whether the element content, mineral industrial grade, and limits for toxic elements meet the requirements.

[0052] First, three types of core data are collected: the final mineral type name determined through multimodal data fusion; elemental information including standardized content of all elements; and phase information including phase composition and reliability. Simultaneously, a specific quality standard bound to the final mineral type name is retrieved from a multidimensional dynamic knowledge base, covering the typical elemental composition range of the mineral, its industrial grade, and mandatory limits for toxic and hazardous elements. Then, using the final mineral type name as the core link, a precise mapping between the testing data and the quality standard is established, and three-dimensional verification is carried out sequentially. This includes comparing the total elemental content in the elemental information with the typical distribution range of the corresponding mineral in the knowledge base to determine if it is within a reasonable range. For minerals whose industrial mining value needs to be differentiated, the core element content is verified to ensure it meets the corresponding industrial grade or concentrate quality requirements. The content of toxic and hazardous elements in the elemental information is checked against mandatory limits, and cross-validation is performed using phase information to identify any instances of elemental compliance with abnormal phase composition. Finally, a comprehensive judgment is made regarding compliance.

[0053] This compliance verification process effectively addresses the pain points of traditional mineral content verification, which relies on laboratory testing, is time-consuming, and suffers from significant errors in manual judgment. By linking on-site testing data with a dynamic knowledge base in real time, it achieves automated and immediate verification, significantly improving customs clearance efficiency. Its multi-dimensional verification logic covers the basic properties, industrial value, and safety baseline of minerals, and avoids regulatory loopholes in single-element testing through cross-phase verification, significantly improving the accuracy of compliance judgments. It has already accurately identified risks in real-world cases such as excessive arsenic in copper concentrate. Furthermore, standardized verification rules replace reliance on the knowledge reserves of professional personnel, lowering the enforcement threshold. Clearly defined violation markers provide a clear basis for subsequent handling measures such as return shipments and further inspection, strongly supporting intelligent customs supervision and risk control.

[0054] Risk screening includes: matching the final mineral type name, element information, and phase information with the solid waste characteristics, classification rules, and prohibited and restricted lists in a multi-dimensional dynamic knowledge base; analyzing the consistency between the final mineral type name, element information, and phase information and the data in the multi-dimensional dynamic knowledge base; and forming risk labels and compliance recommendations.

[0055] The system collects the final mineral species names, elemental information including full element content and matching degree, and phase information including phase composition and reliability, determined through multimodal data fusion. Simultaneously, it retrieves corresponding solid waste characteristics and prohibited / restricted lists from a multidimensional dynamic knowledge base. The core detection information is then precisely matched with the retrieved knowledge base content, analyzing the consistency between the final mineral species names, elemental content, phase composition, and solid waste characteristics and prohibited / restricted list data. If sensitive phases or elemental contents of solid waste are detected that meet prohibited / restricted standards, or if the mineral name belongs to a prohibited / restricted category, data inconsistency is identified, and risk points are identified. If all information matches the compliant data in the knowledge base, consistency is determined. Finally, based on the matching results, clear risk labels such as solid waste risk and dual-use item control are generated, along with targeted compliance recommendations such as laboratory solid waste identification and export license verification.

[0056] The risk screening process effectively addresses the pain points of slow risk identification and high reliance on manual labor in traditional supervision. By real-time correlation and matching of on-site detection data with a dynamic knowledge base, it achieves rapid and accurate identification of risks such as disguised solid waste and undeclared prohibited or restricted materials. It has successfully supported the investigation of multiple cases, including the seizure of copper slag solid waste disguised as copper concentrate. It not only replaces manual experience-based judgment with standardized algorithms, reducing knowledge barriers and human error in law enforcement, but also provides on-site customs officers with direct and actionable handling guidance through clear risk labels and compliance suggestions. This significantly improves the efficiency of risk cargo screening, prevents the inflow or outflow of prohibited or restricted materials, hazardous waste, and other illegal goods, and builds a solid risk prevention and control barrier for intelligent customs supervision.

[0057] For example, this embodiment also provides a working process for copper mine identification and risk warning, specifically including:

[0058] Elemental analysis: A mineral was scanned using an on-site X-ray fluorescence spectrometer. The main elemental composition was found to be Si 26.1%, Fe 12.0%, Cu 8.52%, S 6.54%, Al 2.35%, Ca 2.31%, and Mg 1.72%. Based on elemental prominence factor calculations, the top three elements were Cu, Si, and Fe. The elemental content was compared with various copper ores and copper-bearing materials, silica minerals and silicon-bearing materials, and iron ores and iron-bearing materials databases in a multidimensional dynamic knowledge base. Copper ore ranked highest with a matching degree of 96.7%.

[0059] Phase analysis: Card combinations related to copper ore were retrieved and automatically fitted. The highest fitting degree was found for chalcopyrite (27.1%), quartz (46.3%), and chlorite (26.6%), with a confidence level of 91.5%.

[0060] Intelligent analysis: The quality requirements for copper ore were checked, and the sample met the requirements; the phase composition was examined, and it was found to be consistent with the common phase composition of copper ore, with no characteristic phases such as metallurgical slag found; the regulatory policies were checked, and it was found that the gold value of copper ore could be exempted from part of the value-added tax by declaring it separately.

[0061] Output results: The mineral name is copper ore, with a matching degree of 96.7%; the mineral elemental composition is Si 26.1%, Fe 12.0%, Cu 8.52%, S 6.54%, Al 2.35%, Ca 2.31%, Mg 1.72%; the mineral phase composition is chalcopyrite 27.1%, quartz 46.3%, chlorite 26.6%, with a confidence level of 91.5%; it is recommended to declare the gold content of the mineral separately to enjoy the exemption of value-added tax on the gold value.

[0062] Based on the above embodiments, after outputting mineral identification results, risk labels, and compliance recommendations, the method further includes: verifying the output mineral identification results; supplementing the verified accurate data into the multidimensional dynamic knowledge base; marking correction information according to the verified deviation results and optimizing the parameters in the multimodal data fusion analysis process; and periodically incorporating elemental and phase information of new mining and metallurgical products and solid waste to dynamically update the multidimensional dynamic knowledge base.

[0063] After outputting mineral identification results, risk labels, and compliance recommendations, the first step is to verify the accuracy of the output mineral identification results. Verified data is then added to a multi-dimensional dynamic knowledge base to enrich the basic information such as the elemental composition and phase characteristics of the minerals. For identification deviations discovered during the verification process, corresponding correction information is annotated, and based on this, relevant parameters such as weight allocation and matching thresholds in the multi-modal data fusion analysis process are optimized. Simultaneously, elemental and phase information of new mining and metallurgical products and various solid wastes is regularly collected and incorporated into the multi-dimensional dynamic knowledge base to achieve continuous dynamic updates and ensure that data coverage is synchronized with regulatory needs.

[0064] This embodiment constructs a closed-loop optimization mechanism of output, verification, optimization, and updating. On the one hand, by supplementing accurate data and incorporating information on new materials, it continuously improves the comprehensiveness, accuracy, and timeliness of the multi-dimensional dynamic knowledge base, addressing the pain point that traditional knowledge bases are difficult to adapt to the supervision of new mining and metallurgical products and solid waste. On the other hand, by optimizing the multi-modal data fusion analysis parameters based on deviation results, it can continuously improve the accuracy of mineral identification and the precision of risk screening, avoiding misjudgment problems caused by fixed model parameters. Ultimately, this enables the intelligent identification and supervision system to form a self-iterative and upgrading capability, adapting to the supervision and analysis needs of various scenarios such as customs, mines, and trading companies in the long term, and continuously consolidating the technical support for intelligent supervision.

[0065] This invention also provides an intelligent identification system for mining and metallurgical products based on a dynamic knowledge base. Figure 2 This is a schematic diagram of the structure of an intelligent identification system for mining and metallurgical products based on a dynamic knowledge base, as provided in an embodiment of the present invention. Figure 2 As shown, the system includes:

[0066] The knowledge base construction module is used to acquire mineral attribute information, classification standard rules, risk characteristics and multi-source public data, and to update the multi-source public data in real time. The acquired information is associated with preset mineral main elements as the core to form a multi-dimensional dynamic knowledge base.

[0067] The data acquisition module is used to collect full-element spectral data and / or crystal structure diffraction data of the mineral and metallurgical products to be tested at the regulatory site, obtain elemental information and / or phase information of the mineral and metallurgical products to be tested, and determine the mineral type in a multi-dimensional dynamic knowledge base based on the elemental information and phase information.

[0068] The analysis module is used to retrieve the corresponding mineral industrial grade or element abundance based on the mineral type according to the multidimensional dynamic knowledge base, and to perform weight allocation and multimodal data fusion analysis on the element information, phase information and mineral industrial grade or element abundance to determine the final mineral type name of the mineral and metallurgical product to be tested.

[0069] The output module is used to perform compliance verification, classification judgment, risk screening and early warning based on the final mineral type name, element information and phase information, and to associate with quality standards, classification rules, prohibited and restricted lists and solid waste characteristics in a multi-dimensional dynamic knowledge base, and output mineral identification results, risk labels and compliance suggestions.

[0070] The intelligent identification system for mining and metallurgical products based on a dynamic knowledge base provided in this embodiment is used to execute the intelligent identification method for mining and metallurgical products based on a dynamic knowledge base provided in any of the above embodiments, and has the beneficial effects of any of the above embodiments, which will not be repeated here.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent identification of mining and metallurgical products based on a dynamic knowledge base, characterized in that, The method comprises the following steps: S1, acquiring mineral attribute information, classification standard rules, risk characteristics and multi-source public data, and updating the multi-source public data in real time, correlating the acquired information with a preset main element of the mineral to form a multi-dimensional dynamic knowledge base; S2, collecting full-element spectrum data and / or crystal structure diffraction data of the to-be-tested mineral products on the supervision site to obtain element information and / or phase information of the to-be-tested mineral products, and determining the mineral type in the multi-dimensional dynamic knowledge base according to the element information and the phase information; S3, based on the multi-dimensional dynamic knowledge base, the corresponding mineral industrial grade or element abundance is called according to the mineral type, and the element information, the phase information and the mineral industrial grade or the element abundance are subjected to weight distribution and multi-modal data fusion analysis to determine the final mineral type name of the to-be-tested mineral products; S4, based on the final mineral type name, the element information and the phase information, the quality standard, the classification rule, the prohibited and restricted list and the solid waste characteristics in the multi-dimensional dynamic knowledge base are correlated to perform compliance checking, classification determination, risk screening and early warning, and the mineral identification result, the risk label and the compliance suggestion are output.

2. The method according to claim 1, wherein, In S1, the multi-source public data includes international organization data, national official data, academic resource data and market intelligence data, and the multi-source public data is updated in real time, which comprises: Monitoring the international organization data, the national official data, the academic resource data and the market intelligence data through a network crawler; and sharing the customs node detection data in the national official data through federated learning to iteratively update the multi-source public data in real time. 3.The method of claim 1, wherein, In S2, the element information and / or the phase information of the to-be-tested mineral products are obtained by collecting full-element spectrum data and / or crystal structure diffraction data of the to-be-tested mineral products on the supervision site, which comprises: Collecting full-element spectrum data of the to-be-tested mineral products through a first portable device, and performing standardization processing on the full-element spectrum data, matching the data after the standardization processing with a mineral database in the multi-dimensional dynamic knowledge base to obtain element information, wherein the element information includes element composition types and element contents; Collecting crystal structure diffraction data of the to-be-tested mineral products through a second portable device, performing feature extraction on the crystal structure diffraction data, importing the feature extraction results into a phase combination card library in the multi-dimensional dynamic knowledge base corresponding to the predicted mineral type, performing full-spectrum fitting and refining to obtain phase content and fitting goodness information.

4. The method according to claim 3, wherein, The first portable device comprises a portable X-ray fluorescence spectrometer or a laser-induced breakdown spectrometer; and the second portable device comprises a portable X-ray diffraction spectrometer.

5. The method according to claim 3, wherein, In S3, the element information, the phase information and the mineral industrial grade or the element abundance are subjected to weight distribution and multi-modal data fusion analysis, which comprises: Extracting prominent factors, sorting results and element matching degrees in the element information, extracting phase fitting goodness in the phase information, and extracting coupling matching degrees of the element information and the phase information; In the prominent factor calculation, for a conventional industrial mining mineral, the mineral industrial grade is used as a key weighting item, and for a rare mineral, the element abundance is used as a key weighting item; determine the priority order of the element information matching according to the ranking result of the highlight factors; determine the priority order of the phase information fitting according to the ranking result of the highlight factors; integrate the element matching degree, the phase fitting degree, the coupling matching degree, the ranking result of the highlight factors, and the key weighting items to perform the multi-modal data fusion calculation.

6. The method according to claim 5, wherein, The multi-modal data fusion calculation includes: matching the element information of the to-be-tested mining and metallurgical product with the element information of the main element mineral species in the top three highlight factors in the mineral database to obtain an element information matching degree, and converting the element information matching degree to obtain a first contribution value in the element matching dimension; converting the phase fitting degree to obtain a second contribution value in the phase matching dimension; matching the element information of the to-be-tested mining and metallurgical product with the element information obtained by converting the phase content, and converting the matching result to obtain a third contribution value of the coupling matching of the element information and the phase information; performing weighted summation on the first contribution value, the second contribution value, and the third contribution value to obtain a comprehensive score of the to-be-tested mining and metallurgical product and each candidate mineral species; selecting the candidate mineral species with the highest comprehensive score as the final mineral species name.

7. The method according to claim 1, wherein, In S4, the compliance check includes: associating the final mineral species name, the element information, and the phase information with the quality standards in the multi-dimensional dynamic knowledge base to verify whether the element content, the mineral industrial grade, and the toxic element limit meet the requirements. 8.The method of claim 1, wherein, The risk screening includes: matching the final mineral species name, the element information, and the phase information with the solid waste characteristics, the classification rules, and the prohibited and limited list in the multi-dimensional dynamic knowledge base to analyze the consistency of the final mineral species name, the element information, and the phase information with the data in the multi-dimensional dynamic knowledge base, and form a risk label and a compliance suggestion. 9.The method of claim 1, wherein, After outputting the mineral identification result, the risk label, and the compliance suggestion, the system further includes: verifying the output mineral identification result; supplementing the verified accurate data to the multi-dimensional dynamic knowledge base; annotating correction information according to the deviation result of the verification and optimizing the parameters in the multi-modal data fusion analysis process; periodically incorporating the element information and the phase information of new mining and metallurgical products and solid waste to dynamically update the multi-dimensional dynamic knowledge base.

10. A dynamic knowledge base based intelligent identification system for mining and metallurgical products, characterized in that, The system is used to perform the intelligent identification method of mining and metallurgical products based on the dynamic knowledge base according to any one of claims 1-9, and the system includes: a knowledge base construction module configured to acquire mineral attribute information, classification standard rules, risk characteristics, and multi-source public data, update the multi-source public data in real time, associate the acquired information with a preset main element of a mineral as a core, and construct a multi-dimensional dynamic knowledge base; a collection module configured to collect full-element spectrum data and / or crystal structure diffraction data of a to-be-tested mining and metallurgical product on a supervision site to obtain element information and / or phase information of the to-be-tested mining and metallurgical product, and determine a mineral species type in the multi-dimensional dynamic knowledge base according to the element information and the phase information; an analysis module configured to determine a final mineral species name of the to-be-tested mining and metallurgical product by performing weight distribution and multi-modal data fusion analysis on the element information, the phase information, and a mineral industrial grade or element abundance corresponding to the mineral species type based on the multi-dimensional dynamic knowledge base. An output module is configured to associate the quality standards, classification rules, prohibited and restricted lists, and solid waste characteristics in the multi-dimensional dynamic knowledge base based on the final mineral species name, element information, and phase information, perform compliance checking, classification determination, risk screening, and early warning, and output mineral identification results, risk labels, and compliance suggestions.

Images