Special investigation method and system for key natural resource assets

Abstract

The application relates to the technical field of data processing, and discloses a special investigation method and system for key natural resource assets. The method comprises the following steps: performing three-dimensional identification on mineral data through a resource classification matrix to generate a key resource list; recognizing different format coordinates by using a regular expression to form a standardized coordinate set; performing batch coordinate conversion by using a convolutional neural network; performing neighborhood interpolation correction based on coordinate offsets; performing coordinate precision verification and abnormality repair; and outputting an investigation data set containing quality scores. The application solves the technical problems of lacking intelligent identification of key resources, automatic conversion of specific format coordinates and intelligent control of coordinate quality in the existing natural resource asset investigation method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a special investigation method and system for key natural resource assets. BACKGROUND

[0002] The existing natural resource asset investigation technology mainly adopts a traditional investigation method based on single mineral classification, and completes asset investigation work through manual price system input and manual coordinate data processing. The natural resource asset investigation system in the prior art can perform basic classification processing on mineral data in the reserve database, supports conversion between standard coordinate systems such as 2000 national geodetic coordinate system and 1984 Xi'an coordinate system, and provides a mineral resource asset price system input function. The investigation process includes basic steps such as investigation parameter configuration, price system data input, and investigation data production, and can generate an investigation report containing resource distribution information and value evaluation results.

[0003] However, the prior art has significant deficiencies, mainly manifested in the lack of intelligent identification capability of key resources, and only simple classification according to preset mineral types, which cannot dynamically determine the investigation focus according to the strategic importance, reserve size and market value of resources. The prior art is limited to standard coordinate system conversion in terms of coordinate data processing, and lacks automatic identification and batch conversion capability when facing specific format coordinates such as historical legacy non-standard coordinate format, degree-minute-second format, and local self-defined coordinate system. In addition, the prior art lacks sufficient control of coordinate precision, and when there are problems such as missing geometric coordinates and center point coordinate offset in the reserve database, only error logs can be generated for manual processing, and lacks intelligent coordinate quality control and automatic repair mechanism.

[0004] Based on the limitation analysis of the prior art, the special investigation method for key natural resource assets needs to solve a series of progressive technical problems: how to establish a multi-dimensional intelligent identification system for key resources, automatically select natural resources with strategic value and investigation priority; how to realize automatic identification and batch conversion of specific format coordinates, process historical legacy non-standard coordinate data; how to use deep learning technology to improve the precision and automation level of coordinate conversion; how to establish an intelligent detection and correction mechanism for coordinate offset, automatically identify and repair abnormal coordinate data; how to build a full-process coordinate quality verification and abnormal repair system to ensure the accuracy and integrity of the investigation data. SUMMARY

[0005] The present application provides a special investigation method and system for key natural resource assets, which solves the technical problems of lack of intelligent identification of key resources, automatic conversion of specific format coordinates, and intelligent control of coordinate quality in the existing natural resource asset investigation method.

[0006] In a first aspect, the application provides a special investigation method for key natural resource assets, which comprises: performing three-dimensional identification processing on mineral data in a reserve database through a resource classification matrix to obtain a key resource list containing type identification codes, hierarchical codes and priority scores; identifying different format coordinates through regular expression matching according to coordinate data in the key resource list to obtain a standardized coordinate set containing original coordinates, format identification and conversion parameters; performing batch coordinate conversion processing in a unified coordinate system on the standardized coordinate set through a convolutional neural network to obtain target coordinate data; performing neighborhood interpolation correction processing on center point coordinates according to coordinate offsets in the target coordinate data to obtain a corrected coordinate data set; and performing coordinate precision verification and automatic repair processing on abnormal coordinates on the corrected coordinate data set to obtain an investigation data set containing quality scores, repair records and coordinate error analysis.

[0007] In a second aspect, the application provides a special investigation system for key natural resource assets, which comprises:

[0008] An identification module is configured to perform three-dimensional identification processing on mineral data in a reserve database through a resource classification matrix to obtain a key resource list containing type identification codes, hierarchical codes and priority scores;

[0009] An identification module is configured to identify different format coordinates through regular expression matching according to coordinate data in the key resource list to obtain a standardized coordinate set containing original coordinates, format identification and conversion parameters;

[0010] A conversion module is configured to perform batch coordinate conversion processing in a unified coordinate system on the standardized coordinate set through a convolutional neural network to obtain target coordinate data;

[0011] A correction module is configured to perform neighborhood interpolation correction processing on center point coordinates according to coordinate offsets in the target coordinate data to obtain a corrected coordinate data set;

[0012] A repair module is configured to perform coordinate precision verification and automatic repair processing on abnormal coordinates on the corrected coordinate data set to obtain an investigation data set containing quality scores, repair records and coordinate error analysis.

[0013] In a third aspect, a special investigation device for key natural resource assets is provided, which comprises a memory and at least one processor, the memory storing instructions; the at least one processor invokes the instructions in the memory to enable the special investigation device for key natural resource assets to perform the special investigation method for key natural resource assets described above.

[0014] In a fourth aspect, a computer readable storage medium is provided, in which instructions are stored, when executed on a computer, cause the computer to perform the special investigation method for key natural resource assets.

[0015] In the technical scheme provided in the present application, the mineral data in the reserve database is processed by three-dimensional identification through the resource classification matrix, breaking through the limitation of the traditional investigation method according to single mineral classification, establishing a three-dimensional classification system based on resource type, management level and priority score, and dynamically identifying key investigation resources according to reserve scale, market price fluctuation and management authority, significantly improving the pertinence and efficiency of the investigation work. Through regular expression matching, different format coordinates are identified and processed, solving the technical problem that the existing technology cannot process historical non-standard coordinate formats, realizing the automatic identification and unified processing of multiple coordinate expression methods such as degree-minute-second format, decimal degree format, projection coordinates, and greatly reducing the need for manual intervention. The application of convolutional neural network in batch coordinate conversion processing in the unified coordinate system realizes intelligent learning and accurate conversion of complex coordinate transformation relationship through two-dimensional matrix coding, convolution layer feature extraction, pooling layer dimension reduction and full connection layer mapping calculation, which has higher conversion accuracy and stronger adaptability than traditional parameterized conversion methods. The neighborhood interpolation correction processing calculates the coordinate offset and intelligently corrects the adjacent mining right coordinates, effectively solving the problem of center point coordinate deviation from the geometric boundary of the mining right, and improving the spatial rationality of the coordinate data. The coordinate precision verification and abnormal coordinate automatic repair processing establish a complete data quality control system, through precision grading, anomaly detection, missing data filling and error replacement, etc., to ensure the accuracy and integrity of the investigation data, and provide a reliable data foundation for natural resource asset management.

[0016] In the specific application field of natural resource asset special investigation, the introduction of convolutional neural network algorithm has important technical contribution. Through local feature extraction and global mapping learning, it can automatically identify the transformation mode and conversion rule between different coordinate formats, avoiding the shortcomings of traditional methods relying on manual setting of conversion parameters. Especially in dealing with non-standard formats such as local coordinate system and enterprise self-defined coordinate system, the adaptive learning ability of neural network is significantly better than that of fixed parameter model. The application of inverse distance weighted interpolation algorithm in coordinate correction links fully considers the spatial continuity characteristics of natural resource distribution, realizes the reasonable correction of abnormal coordinates through the distance weight distribution of neighborhood coordinates, and avoids the spatial relationship distortion problem that may be caused by simple average or nearest neighbor replacement. The regular expression matching algorithm can accurately identify the coordinate data of long measurement time and non-uniform format by constructing a coordinate format identification rule library according to the characteristics of complex and diverse historical data formats in natural resource investigation. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings.

[0018] Figure 1 An embodiment schematic diagram of the special investigation method for the key natural resource assets in the embodiments of the present application;

[0019] Figure 2 An embodiment schematic diagram of the special investigation system for the key natural resource assets in the embodiments of the present application;

[0020] Figure 3 An embodiment schematic diagram of the structure of the special investigation equipment for the key natural resource assets in the embodiments of the present application. DETAILED DESCRIPTION

[0021] The embodiments of the present application provide a special investigation method and system for key natural resource assets. The terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present application and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0022] For the convenience of understanding, the specific process of the embodiments of the present application will be described below. Please refer to Figure 1 An embodiment of the special investigation method for the key natural resource assets in the embodiments of the present application includes:

[0023] Step S101, performing three-dimensional identification processing on the mineral data in the reserve database through the resource classification matrix to obtain a key resource list containing type identification code, hierarchical code and priority score;

[0024] Step S102, according to the coordinate data in the key resource list, identifying and processing different format coordinates through regular expression matching to obtain a standardized coordinate set containing original coordinates, format identification and conversion parameters;

[0025] Step S103, the standardized coordinate set is subjected to coordinate batch conversion processing in a unified coordinate system through a convolutional neural network to obtain target coordinate data;

[0026] Step S104, according to the coordinate offset in the target coordinate data, a neighborhood interpolation correction processing is performed on the center point coordinate to obtain a corrected coordinate data set;

[0027] Step S105, the corrected coordinate data set is subjected to coordinate precision verification and abnormal coordinate automatic repair processing to obtain an inspection data set containing quality scores, repair records and coordinate error analysis.

[0028] It can be understood that the execution subject of the present application can be a special inspection system for key natural resource assets, and can also be a terminal or a server, which is not limited here. The server is taken as an example for illustration in the embodiments of the present application.

[0029] Specifically, the mineral data in the reserve database is subjected to three-dimensional identification processing through a resource classification matrix. The resource classification matrix adopts a three-dimensional cubical classification system constructed by resource type, management level and inspection priority. A mapping relationship of mineral management level is established based on a mineral rights transfer registration permission table. The mineral codes in the reserve database are distributed to the provincial, municipal and county management levels according to the management authority. The reserve scale coefficient of each mineral is calculated according to its reserve scale parameter. The reserve scale coefficient is obtained by dividing the total reserve by the average reserve in the region to obtain a relative scale value. Then, a market price fluctuation parameter is introduced to perform weighted processing on the reserve scale score. The price fluctuation coefficient is calculated based on the recent market price change amplitude. Finally, the reserve scale score and the price weighted score are comprehensively calculated through a priority score algorithm to generate a key resource list containing type identification code, level code and priority score. According to the coordinate data in the key resource list, different format coordinates are identified through regular expression matching. The coordinate format identification first scans the coordinate data for format type, identifies degree-minute-second format, decimal system format and projection coordinate three main format types. The regular expression matching library contains matching patterns for different coordinate formats. The degree, minute, second separators or decimal point separators in the coordinates are extracted through separator identification. The precision bit number analysis determines the significant digits of the coordinate value. The sign identification identifies the positive and negative signs or the east-west-south-north direction identification. After the format analysis is completed, the corresponding ellipsoid parameters, projection parameters and reference point parameters are obtained by querying the historical coordinate system parameter database according to the coordinate system identification. The coordinate format standardization algorithm uniformly converts various formats to decimal system representation. The conversion parameter configuration determines the parameter settings required for subsequent conversion according to the coordinate system parameters to generate a standardized coordinate set containing original coordinates, format identification and conversion parameters.

[0030] The standardized coordinate set is subjected to batch coordinate conversion processing in a unified coordinate system through a convolutional neural network. First, the coordinate data is subjected to two-dimensional matrix coding, and the X coordinate value and the Y coordinate value are subjected to numerical normalization processing. The normalization is achieved by subtracting the minimum value from the coordinate value and dividing by the value range. The format identifier is converted to a numerical form through numerical coding mapping. The national coordinate system, the regional coordinate system and the local coordinate system correspond to different coding values respectively. The normalized coordinate values and the format coding data construct a three-dimensional array. Three channels store the X coordinate, the Y coordinate and the format identifier information respectively. The array is subjected to matrix reshaping according to a fixed size, forming a two-dimensional coordinate matrix structure with a standard row and column specification. A coordinate feature matrix is generated through pixel value filling and zero padding processing. A convolutional layer extracts local features from the coordinate feature matrix, extracts coordinate offset features and format conversion features. A pooling layer reduces the dimension of the local feature map and filters noise. A fully connected layer calculates the coordinate mapping relationship, and outputs the conversion coefficient and the offset parameter. Finally, batch coordinate conversion calculation is performed on the standardized coordinate set, and the target coordinate data in the unified coordinate system is obtained.

[0031] The center point coordinates are corrected by neighborhood interpolation based on the coordinate offset in the target coordinate data. The coordinate offset is determined by calculating the distance between the center point coordinates and the mining right boundary coordinates. The offset includes two components: radial offset distance and angular offset. The radial offset distance represents the straight-line distance of the center point from the geometric center of the mining right, and the angular offset represents the angle between the offset direction and the standard azimuth. Abnormal center point coordinates that exceed the reasonable range are filtered out according to a preset offset threshold. The neighborhood coordinate search algorithm searches for the coordinates of adjacent mining rights around the abnormal coordinate points. The search range is dynamically adjusted according to the size of the mining right area. Different distance weights are assigned to the coordinates of adjacent mining rights according to their distance. The inverse distance weight interpolation algorithm recalculates the reasonable position of the abnormal point using the neighborhood coordinate data. In the interpolation calculation, the closer the point is, the greater the weight. Finally, the corrected X coordinates, Y coordinates, and confidence score are generated. The confidence score is calculated based on the number and distance distribution of the neighborhood points involved in the interpolation. The corrected coordinate dataset is used for coordinate accuracy verification and automatic repair of abnormal coordinates. First, coordinate points are classified into accuracy levels based on confidence scores. High-precision coordinates correspond to coordinate points with higher confidence scores, while medium-precision and low-precision coordinates correspond to points with lower confidence scores, respectively. Neighborhood distance calculation checks the rationality of spatial relationships by measuring the Euclidean distance between adjacent coordinate points. Coordinate points with excessively large or small distances are marked as abnormal coordinates. Abnormal coordinates are classified into two types: missing coordinates and erroneous coordinates. Missing coordinates refer to coordinates with empty values ​​or values ​​that are significantly outside the geographical range, while erroneous coordinates refer to coordinates with unreasonable spatial relationships with surrounding coordinate points. Missing coordinates are filled with values ​​using the average of surrounding valid coordinate points, and erroneous coordinates are replaced with the closest valid coordinate values. During the repair process, the repair operation type and the changes in coordinate values ​​before and after repair are recorded. The quality score is calculated based on the repair success rate and the degree of coordinate consistency. Error statistical analysis includes the mean, variance, and maximum deviation of coordinate deviations before and after repair.

[0032] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0033] Based on the table of mining rights transfer registration authority, the management level of mineral types is mapped to obtain the provincial mineral type list, the municipal mineral type list and the county mineral type list;

[0034] The mineral type codes in the reserves database are input into the hierarchical mapping table for hierarchical identification processing, resulting in hierarchical identification data containing mineral type codes and management level identifiers.

[0035] Based on the reserve scale parameters in the hierarchical identification data, the reserve scale coefficient of each mineral type is calculated to obtain the reserve scale scoring data.

[0036] The reserve scale scoring data is subjected to price fluctuation coefficient weighting processing based on market price fluctuation parameters to obtain price weighting scoring data;

[0037] The price weighting scoring data is subjected to comprehensive scoring calculation processing by priority scoring algorithm to obtain a key resource list containing type identification code, level code and priority score.

[0038] Specifically, the management level mapping processing of the mineral species is carried out based on the mineral rights transfer registration authority division table. The mineral rights transfer registration authority division table is a management authority allocation file formulated by the natural resource management department, which clearly stipulates that the transfer registration authority of different mineral species belongs to the provincial, municipal or county natural resource management department. The division table determines the management level according to factors such as strategic importance, reserve scale, economic value and environmental impact of the mineral species. The level mapping processing establishes the corresponding relationship database of the mineral species and the management level by reading the mineral species name and the corresponding management level identification in the division table. The provincial mineral species list contains national strategic minerals and large deposits, the municipal mineral species list contains regional important minerals and medium-sized deposits, and the county mineral species list contains local minerals and small deposits. The mapping processing result generates three independent mineral species list files, each of which contains three core fields of mineral species code, mineral species name and management level identification. The mineral species code in the reserve database is input into the level mapping table for attribution level identification processing. The reserve database stores the basic information of each mineral area, including mineral species code, reserve data, geographical location and other key information. The attribution level identification processing queries the corresponding management level identification in the level mapping table by taking the mineral species code as the key field. The query process adopts an exact matching method to ensure the consistency of the mineral species code. The query result writes the management level identification into the corresponding record of the reserve database. The level identification data contains the original mineral species code and the newly added management level identification field. The management level identification adopts numerical coding form. The provincial management level code is one, the municipal management level code is two, and the county management level code is three. The level identification data lays a data foundation for subsequent reserve scale analysis and priority score calculation.

[0039] The reserve scale parameter is extracted from the reserve database, including proven reserves, controlled reserves, inferred reserves and other types of reserve values. The reserve scale coefficient is calculated by comparing the total reserves of each mineral with the average reserves of the same type of mineral. The calculation process first calculates the total reserves of the same mineral in different mining areas, and then calculates the average reserve value of the mineral in all mining areas. The reserve scale coefficient is equal to the reserve value of a certain mining area divided by the average reserve value of the mineral. A coefficient greater than one indicates that the reserve scale of the mining area exceeds the average level, and a coefficient less than one indicates that the reserve scale is lower than the average level. The reserve scale score data is obtained by multiplying the reserve scale coefficient by the weight factor. The weight factor is determined according to the strategic importance of the mineral. The weight factor of strategic minerals is higher, and the weight factor of general minerals is lower. The price fluctuation coefficient of the reserve scale score data is weighted based on the market price fluctuation parameter. The market price fluctuation parameter is obtained by analyzing the recent market price trend of mineral products. The price fluctuation coefficient is calculated by the ratio of the current market price to the historical average price. A ratio greater than one indicates an upward price trend, and a ratio less than one indicates a downward price trend. The price fluctuation coefficient weighted processing multiplies the reserve scale score data by the price fluctuation coefficient to obtain the price weighted score data. The weighted processing considers the impact of market price changes on the value of mineral resources. Minerals with rising prices receive higher weighted scores, and minerals with falling prices receive relatively lower weighted scores.

[0040] The price weighted score data is processed by a priority score algorithm to calculate the comprehensive score. The priority score algorithm considers multiple factors such as reserve scale, price fluctuation, management level and environmental constraints. The algorithm calculates the comprehensive score of each mineral using a weighted average method. The comprehensive score is equal to the reserve scale score multiplied by the reserve weight plus the price fluctuation score multiplied by the price weight plus the management level score multiplied by the level weight. The sum of the reserve weight, price weight and level weight is equal to one. The weight distribution is determined according to the inventory target and policy orientation. The reserve weight focuses on the richness of resource reserves, the price weight focuses on the trend of economic value, and the level weight focuses on the importance of management authority. The comprehensive score calculation result generates a key resource list containing type identification code, level code and priority score. The type identification code identifies the classification attribute of the mineral, the level code identifies the management authority, and the priority score identifies the priority level of the inventory.

[0041] In a specific embodiment, the process of performing step S102 can specifically include the following steps:

[0042] The coordinate data in the key resource list is scanned and processed in a format type to obtain coordinate format classification data in degrees, minutes and seconds, decimal degrees and projection coordinates.

[0043] The separator recognition and precision bit number analysis processing is performed on the coordinate format classification data based on a regular expression matching library, to obtain format analysis data containing a coordinate separator, precision bit number and symbol identifier;

[0044] According to the coordinate system identifier in the format analysis data, parameter query matching processing is performed on a historical coordinate system parameter database, to obtain coordinate system parameter data containing ellipsoid parameters, projection parameters and reference point parameters;

[0045] The format analysis data is subjected to unified format conversion processing through a coordinate format standardization algorithm, to obtain standard format coordinate data in a decimal system;

[0046] The standard format coordinate data is subjected to conversion parameter configuration processing based on the coordinate system parameter data, to obtain a standardized coordinate set containing original coordinates, format identifier and conversion parameters.

[0047] Specifically, the coordinate data in the key resource list is subjected to format type scanning processing. Format type scanning is a process of automatically identifying the text format of coordinate data. The scanning algorithm judges the coordinate format type by analyzing the character composition and numerical arrangement mode in the coordinate data. The degree-minute-second format coordinate contains three numerical units of degrees, minutes and seconds, and is usually separated by the symbols "degrees", "minutes", "seconds" or "°", "′", "″". The decimal degree format coordinate represents the angle value in the form of decimal point. The numerical range is between zero and three hundred and sixty degrees. The projection coordinate represents the geographical position in a rectangular coordinate system, usually as six or seven integers. The coordinate format classification data is obtained by traversing each coordinate record in the key resource list, identifying the format characteristics of the coordinate data one by one, and storing the identification results as format classification identifiers. The degree-minute-second format identifier is DMS, the decimal degree format identifier is DD, and the projection coordinate format identifier is PROJ. After the classification processing, a classification data set containing original coordinate data and format classification identifiers is generated.

[0048] The separator identification is determined by matching special characters in the coordinate data to determine the separation mode between the numerical values. The separator of the degree-minute-second format includes the "degree", "minute", "second" character or "°", "′", "″" mathematical symbol. The separator of the decimal degree format is the decimal point. The projection coordinate format usually has no separator or uses a comma to separate the X coordinate and the Y coordinate. The precision bit number analysis determines the coordinate precision by counting the number of digits after the decimal point. The more the number of decimal digits, the higher the coordinate precision. The symbol identification includes the positive and negative sign identification and the direction identification. The positive and negative sign identification indicates the direction of the east and west longitude and the north and south latitude. The direction identification includes the east longitude E, the west longitude W, the north latitude N, the south latitude S, etc. The format analysis data stores the separator type, the precision bit number and the symbol identification information, and establishes the corresponding relationship between the coordinate data and the analysis result.

[0049] The parameter query matching processing is performed on the historical coordinate system parameter database according to the coordinate system identification in the format analysis data. The coordinate system identification is extracted from the additional information or the file header information of the coordinate data. The historical coordinate system parameter database stores the technical parameters of various coordinate systems, including the detailed parameter information of the national geodetic coordinate system, the regional coordinate system and the local coordinate system. The ellipsoid parameter defines the geometric shape of the earth ellipsoid, including the long semi-axis, the short semi-axis and the flattening. The projection parameter defines the mathematical transformation relationship of the map projection, including the central meridian, the projection origin and the scale factor. The reference point parameter defines the initial reference of the coordinate system, including the reference point coordinate, the azimuth angle and the scale. The parameter query matching retrieves the corresponding parameter record in the database by taking the coordinate system identification as the key. After the matching is successful, the complete coordinate system parameter data is extracted.

[0050] The format analysis data is subjected to the unified format conversion processing by the coordinate format standardization algorithm. The coordinate format standardization algorithm converts various coordinate formats into a unified decimal degree representation. The calculation method for converting the degree-minute-second format into the decimal degree format is to add the degree number to the fraction divided by sixty and add the second number divided by three thousand six hundred. In the conversion process, the degree number, the fraction number and the second number in the degree-minute-second format are first extracted, and then the corresponding decimal degree value is calculated according to the conversion formula. The projection coordinate is converted into the geographic coordinate by using the corresponding inverse projection formula. The inverse transformation calculation converts the plane coordinate into the latitude and longitude coordinate according to the projection parameter. The conversion process considers the influence of the ellipsoid parameter and the projection deformation. The standard format coordinate data is represented by the unified decimal degree value, which eliminates the differences between different formats.

[0051] The standard format coordinate data is converted and parameter configured based on the coordinate system parameter data. The conversion parameter configuration is to prepare necessary calculation parameters for subsequent coordinate system conversion. The configuration process determines conversion parameters according to the parameter difference between the source coordinate system and the target coordinate system. The conversion parameters include seven parameters such as translation parameters, rotation parameters and scale parameters. The translation parameters represent the displacement amount of the coordinate origin. The rotation parameters represent the rotation angle of the coordinate axis. The scale parameters represent the scaling ratio of the coordinate scale. The parameter configuration is calculated by comparing the ellipsoid parameters and the reference point parameter difference between the source coordinate system and the target coordinate system. The standardized coordinate set contains original coordinate data, format identification information and configured conversion parameters, and forms a complete coordinate conversion data set.

[0052] In a specific embodiment, the process of performing step S103 can specifically include the following steps:

[0053] The coordinate data in the standardized coordinate set is subjected to two-dimensional matrix encoding processing to obtain a coordinate feature matrix containing X coordinates, Y coordinates and format identification;

[0054] Based on the coordinate feature matrix, local feature extraction processing is performed through a convolution layer to obtain a local feature map containing coordinate offset features and format conversion features;

[0055] The local feature map is input into a pooling layer for feature dimension reduction and noise filtering processing to obtain coordinate conversion feature data after dimension reduction;

[0056] According to the coordinate conversion feature data, coordinate mapping relationship calculation processing is performed through a full connection layer to obtain coordinate transformation parameters containing conversion coefficients and offset parameters;

[0057] Based on the coordinate transformation parameters, batch coordinate conversion calculation processing is performed on the standardized coordinate set to obtain target coordinate data under a unified coordinate system.

[0058] Specifically, the coordinate data in the standardized coordinate set is processed by two-dimensional matrix encoding. The two-dimensional matrix encoding is a data preprocessing process for converting the coordinate data into the input format of the convolutional neural network. The encoding process first normalizes the X coordinate value and the Y coordinate value. The normalization maps the coordinate values to the value interval of zero to one through the maximum and minimum value scaling method. The normalization formula is the original coordinate value minus the minimum coordinate value divided by the coordinate value range. The normalization process eliminates the numerical scale difference between different coordinate systems. The format identifier is converted to a numerical form through numerical encoding mapping. The national coordinate system is encoded as one, the regional coordinate system is encoded as two, and the local coordinate system is encoded as three. The encoding mapping establishes the corresponding relationship between the format type and the numerical value. The three-dimensional array constructs the normalized X coordinate value, Y coordinate value, and format encoding data according to the channel dimension. The first channel stores the X coordinate value, the second channel stores the Y coordinate value, and the third channel stores the format encoding value. The matrix reshaping process converts the one-dimensional coordinate data array into a two-dimensional matrix structure. The number of matrix rows and columns is determined according to the number of coordinate points and the size of the convolution kernel. The pixel value filling supplements the missing positions with zero values. The zero padding process ensures that the matrix size meets the input requirements of the convolution layer. The coordinate feature matrix contains complete coordinate information and format information.

[0059] The coordinate feature matrix is processed by a convolution layer for local feature extraction. The convolution layer is the core component of the convolutional neural network. The feature map is obtained by sliding the convolution kernel on the input matrix. The local feature extraction scans the local area of the coordinate feature matrix through a small size convolution kernel. The convolution kernel contains weight parameters and bias parameters. The weight parameters are obtained through training and learning. The convolution calculation process multiplies the convolution kernel and the matrix elements at the corresponding positions one by one and then sums them up. The convolution result plus the bias parameter obtains the pixel value of the feature map. The coordinate offset feature represents the position relationship and trend between adjacent coordinate points. The format conversion feature represents the conversion rule and mapping relationship between different coordinate formats. The local feature map output by the convolution layer contains multiple feature channels. Each channel corresponds to different types of feature information. The feature map size is determined according to the size of the convolution kernel and the step parameter. The step parameter controls the moving distance of the convolution kernel. The local feature map preserves the spatial structure information and format conversion information of the coordinate data.

[0060] The local feature map is input into a pooling layer for feature dimension reduction and noise filtering processing. The pooling layer reduces the spatial size and parameter quantity of the feature map through a sampling operation. The feature dimension reduction adopts a max pooling or average pooling method. The max pooling selects the maximum value in the pooling window as the output value. The average pooling calculates the average value of all values in the pooling window as the output value. The pooling window size and sliding step determine the dimension reduction degree. The larger the window size, the higher the dimension reduction degree. The noise filtering eliminates high-frequency noise and redundant information in the feature map through the pooling operation. The pooling processing retains important feature information while reducing the computational complexity. The coordinate conversion feature data after dimension reduction contains key feature information after screening and compression. The feature data dimension is significantly reduced compared to the original local feature map. The dimension reduction processing reduces the parameter quantity and computational burden of the subsequent fully connected layer.

[0061] The coordinate mapping relationship calculation processing is performed on the coordinate conversion feature data through a fully connected layer. The fully connected layer is the output layer of the neural network, which is responsible for mapping the feature data into the final prediction result. The coordinate mapping relationship calculation processes the feature data through linear transformation and nonlinear activation function. The linear transformation multiplies the input feature data with a weight matrix and adds a bias vector. The weight matrix and the bias vector are learned through the training data. The nonlinear activation function increases the expression ability of the network. Common activation functions include ReLU function and Sigmoid function. The fully connected layer outputs two types of coordinate transformation parameters, conversion coefficients and offset parameters. The conversion coefficients represent the transformation relationship of coordinate scaling and rotation. The offset parameters represent the transformation relationship of coordinate translation. The coordinate transformation parameters reflect the mathematical transformation relationship between the source coordinate system and the target coordinate system. The parameter values are calculated through the forward propagation of the neural network.

[0062] Batch coordinate conversion is performed on the standardized coordinate set based on the coordinate transformation parameters. The batch coordinate conversion is a calculation process of applying the transformation parameters output by the neural network to all coordinate points. The conversion calculation adopts an affine transformation mathematical model. The affine transformation includes linear transformation and translation transformation. The linear transformation calculates the new coordinate position by multiplying the conversion coefficient matrix with the coordinate vector. The translation transformation adjusts the absolute position of the coordinate by adding the offset parameter vector. The batch processing simultaneously performs the same transformation calculation on all coordinate points in the standardized coordinate set. The transformed coordinate data is uniformly represented in the target coordinate system. The target coordinate data eliminates the differences between different source coordinate systems, forming a consistent coordinate expression in space.

[0063] In a specific embodiment, the process of performing two-dimensional matrix encoding on the coordinate data in the standardized coordinate set can specifically include the following steps:

[0064] The X coordinate value and the Y coordinate value in the standardized coordinate set are subjected to numerical normalization processing to obtain normalized X coordinate values and Y coordinate values.

[0065] According to the format identification, the coordinate system type is numerically coded and mapped to obtain format coding data containing national coordinate system coding, regional coordinate system coding and local coordinate system coding;

[0066] Based on the normalized X coordinate value, Y coordinate value and format coding data, three-dimensional array construction processing is performed to obtain a three-channel coordinate data array;

[0067] The coordinate data array is subjected to matrix reshaping processing according to a fixed size to obtain a standard two-dimensional coordinate matrix structure;

[0068] According to the two-dimensional coordinate matrix structure, pixel value filling and zero padding processing are performed to obtain a coordinate feature matrix containing X coordinate, Y coordinate and format identification.

[0069] Specifically, the X coordinate value and Y coordinate value in the standardized coordinate set are subjected to numerical normalization processing. Numerical normalization is a standardization process of mapping coordinate data in different numerical ranges to a unified numerical interval. The normalization processing adopts the maximum and minimum value scaling method. First, the maximum value and the minimum value of all X coordinate values in the standardized coordinate set are counted to calculate the numerical range of the X coordinate. Then, normalization calculation is performed on each X coordinate value. The normalization formula is the current X coordinate value minus the minimum value of the X coordinate divided by the numerical range of the X coordinate. The calculation result is mapped to the numerical interval of zero to one. The normalization processing of the Y coordinate value adopts the same calculation method. The maximum value and the minimum value of the Y coordinate are counted respectively to calculate the numerical range of the Y coordinate. Normalization transformation is performed on each Y coordinate value. The normalization processing eliminates the numerical scale difference between different coordinate systems to ensure that coordinate data of different sources have the same numerical order of magnitude. The normalized X coordinate value and Y coordinate value are distributed in the standard interval of zero to one. According to the format identification, the coordinate system type is numerically coded and mapped. Numerical coding and mapping is a process of converting the text form of the coordinate system identification into a numerical form of the code. The format identification contains the type information of the coordinate system. The national coordinate system represents the national standard coordinate system. The regional coordinate system represents the provincial and municipal regional coordinate system. The local coordinate system represents the county-level or enterprise-level local coordinate system. The coding mapping establishes the corresponding relationship between the coordinate system type and the numerical code. The national coordinate system coding is numerical one. The regional coordinate system coding is numerical two. The local coordinate system coding is numerical three. The coding process converts the format identification of each coordinate point into the corresponding numerical code by looking up the coding mapping table. The format coding data contains the numerical format identification information of all coordinate points.

[0070] Based on the normalized X coordinate value, Y coordinate value and format coding data, three-dimensional array construction processing is performed. The three-dimensional array construction is a data structuring process of organizing different types of coordinate information according to the channel dimension. The three-dimensional array includes three data channels. The first channel stores the normalized X coordinate value, the second channel stores the normalized Y coordinate value, and the third channel stores the format coding data. The array construction fills the corresponding X coordinate, Y coordinate and format coding into the corresponding channel position according to the arrangement order of the coordinate points. Each coordinate point has the same index position in the three channels. The three-channel coordinate data array maintains the integrity and correlation of the coordinate information. The array structure facilitates the multi-channel data processing of the convolutional neural network. The coordinate data array is subjected to matrix reshaping processing according to a fixed size. The matrix reshaping is a shape transformation process of converting one-dimensional arranged coordinate data into a two-dimensional matrix structure. The fixed size is determined according to the number of coordinate points and the matrix processing requirement. The reshaping processing rearranges the one-dimensional coordinate data of the three channels into a matrix form with a standard row and column specification. The number of rows and columns is set considering the requirements of the convolution kernel size and the calculation efficiency. The matrix reshaping maintains the original values of the data unchanged, and only changes the arrangement mode and access index of the data. The two-dimensional coordinate matrix structure facilitates the subsequent convolution calculation and feature extraction processing.

[0071] According to the two-dimensional coordinate matrix structure, pixel value filling and zero padding processing are performed. The pixel value filling and zero padding processing is a perfecting process of adjusting the matrix size and supplementing the missing data. The pixel value filling fills the coordinate value as the pixel value into the corresponding position of the matrix. The filling process fills the normalized coordinate value and the format coding value into the corresponding matrix elements according to the row and column indexes of the matrix. The zero padding processing fills zero values in the positions where the matrix size is insufficient or the data is missing. The zero padding position usually appears in the edge area of the matrix or the case where the number of coordinate points is insufficient to fill the entire matrix. The zero padding processing ensures that the matrix has a complete standard size, meets the strict requirements of the convolutional neural network on the input data format, and the coordinate feature matrix contains complete X coordinate, Y coordinate and format identification information. The matrix structure standardization facilitates the batch processing and parallel calculation of the neural network.

[0072] In a specific embodiment, the process of performing step S104 can specifically include the following steps:

[0073] Based on the center point coordinates and the mining right boundary coordinates in the target coordinate data, offset distance calculation processing is performed to obtain coordinate offset data including radial offset distance and angular offset amount;

[0074] According to the offset threshold in the coordinate offset data, abnormal coordinate point screening processing is performed to obtain a list of abnormal center point coordinates that exceed the offset threshold;

[0075] The coordinate points in the list of abnormal center point coordinates are searched for adjacent points by a neighborhood coordinate search algorithm to obtain neighborhood coordinate data containing adjacent mining right coordinates and distance weights;

[0076] The coordinate position is recalculated based on the neighborhood coordinate data by an inverse distance weighted interpolation algorithm to obtain interpolation corrected coordinates containing corrected X coordinates, corrected Y coordinates and confidence scores.

[0077] The interpolation corrected coordinates are replaced with the original center point coordinates and data are merged to obtain a corrected coordinate dataset containing corrected coordinates and confidence scores.

[0078] Specifically, the offset distance is calculated based on the center point coordinates and mining right boundary coordinates in the target coordinate data, the center point coordinates represent the geometric center position of the mining right or mining area, the mining right boundary coordinates define the boundary contour of the mining right range, the offset distance calculation calculates the offset degree of the center point coordinates relative to the geometric center of the mining right boundary through a geometric distance formula, the radial offset distance calculates the straight line distance between the center point and the boundary geometric center using the Euclidean distance formula, the calculation process obtains the radial distance by taking the square root of the sum of the square of the X coordinate difference and the square of the Y coordinate difference between the center point coordinates and the boundary geometric center coordinates, the angle offset is calculated by the inverse tangent function to calculate the azimuth deviation of the center point relative to the boundary geometric center, the calculation method is the inverse tangent value of the Y coordinate difference divided by the X coordinate difference, the angle offset reflects the degree of deviation of the center point from the standard orientation, the coordinate offset data contains two components of radial offset distance and angle offset of each mining right, the larger the offset value, the farther the center point coordinate deviates from the mining right geometric center. The abnormal coordinate points are screened based on the offset threshold in the coordinate offset data, the offset threshold is a standard value for judging whether the center point coordinate deviates from a reasonable range, the threshold is set based on the size of the mining right area and the geological exploration accuracy requirement, the offset threshold of large mining right is relatively large, and the offset threshold of small mining right is relatively small, the abnormal coordinate points are screened by comparing the size relationship between the radial offset distance of each mining right and the preset threshold to determine the abnormal coordinate, the center point coordinates whose radial offset distance exceeds the threshold range are marked as abnormal coordinate points, the screening process traverses the coordinate offset data of all mining rights, and judges whether the offset distance exceeds the corresponding threshold standard one by one, the list of abnormal center point coordinates contains the identification of all mining rights and the corresponding center point coordinate information that exceed the offset threshold.

[0079] The coordinate points in the list of abnormal center point coordinates are searched for neighboring points by a neighborhood coordinate search algorithm. The neighborhood coordinate search algorithm is a spatial search method for finding neighboring coordinate points around a target coordinate point. The search algorithm takes the abnormal center point coordinate as the search center, sets a search radius range, and searches for other mineral right center point coordinates within the search range. The search radius is determined according to the distribution density of mineral rights and geographical environmental characteristics. The search process calculates the distance between the abnormal coordinate point and all other mineral right center points. Coordinate points with a distance less than the search radius are identified as neighboring points. Adjacent mineral right coordinates refer to mineral right center point coordinates that are spatially adjacent to the abnormal coordinate point. Distance weights are assigned according to the distance between the neighboring points and the abnormal coordinate point. The closer the distance, the higher the weight value. The farther the distance, the lower the weight value. Neighborhood coordinate data includes the coordinate positions of neighboring points and corresponding distance weight information. Based on the neighborhood coordinate data, the coordinate positions are recalculated by an inverse distance weighted interpolation algorithm. The inverse distance weighted interpolation algorithm is a spatial interpolation method for calculating the reasonable position of a target point based on the coordinate positions and distance weights of neighboring points. The interpolation calculation multiplies the coordinate values of the neighboring points by the corresponding distance weights, sums them up, and then divides by the sum of all distance weights to obtain the interpolated coordinate. The calculation method for the corrected X coordinate is to multiply all the X coordinate values of the neighboring points by the sum of the corresponding distance weights and divide by the sum of the distance weights. The corrected Y coordinate uses the same calculation method. The confidence score is determined based on the number of neighboring points involved in the interpolation calculation and the uniformity of the distance distribution. The more the number of neighboring points and the more uniform the distribution, the higher the confidence score. The interpolated and corrected coordinates include the recalculated X coordinate, Y coordinate, and confidence score information.

[0080] The interpolated and corrected coordinates are replaced by the original center point coordinates and the data is merged. Coordinate replacement is a data updating process in which the abnormal center point coordinates are replaced by the interpolated and corrected coordinates. The replacement operation locates the corresponding original coordinate record based on the mineral right identifier in the list of abnormal center point coordinates, replaces the original X coordinate and Y coordinate values with the X coordinate and Y coordinate values of the interpolated and corrected coordinates, and merges the corrected coordinate data with the normal coordinate data that has not been corrected into a complete data set. The merging process maintains the correspondence between the mineral right identifier, the coordinate value, and the confidence score. The corrected coordinate data set contains the final coordinate information of all mineral rights. The abnormal coordinate point uses the corrected coordinate values, and the normal coordinate point maintains the original coordinate values. Each coordinate record contains the correction status identifier and the confidence score information.

[0081] In a specific embodiment, the process of performing step S105 can specifically include the following steps:

[0082] Based on the confidence score in the corrected coordinate data set, an accuracy level division process is performed to obtain an accuracy classification data containing high-precision coordinates, medium-precision coordinates, and low-precision coordinates.

[0083] The coordinate points in the precision grading data are subjected to neighborhood distance calculation and spatial relationship check processing to obtain verification result data containing verified coordinates and abnormal coordinate markers;

[0084] According to the abnormal coordinate markers in the verification result data, missing coordinate identification and error coordinate classification processing are performed to obtain an abnormal coordinate list containing coordinate missing records and coordinate error types;

[0085] Based on the abnormal coordinate list, numerical filling for missing coordinates and numerical replacement for error coordinates are performed to obtain coordinate repair data containing repaired coordinates and repair confidence;

[0086] The coordinate repair data and the verification result data are subjected to quality score calculation and error statistical analysis processing to obtain an inspection data set containing quality scores, repair records, and coordinate error analysis.

[0087] Specifically, based on the confidence score in the corrected coordinate data set, precision level division processing is performed. The confidence score reflects the reliability and correction quality of the coordinate data. The precision level division divides the coordinate data into different precision categories by setting interval thresholds for the confidence score. High-precision coordinates correspond to coordinate points with confidence scores in the interval of eight to ten, indicating that the coordinate position has been fully verified and the correction effect is good. Medium-precision coordinates correspond to coordinate points with confidence scores in the interval of five to eight, indicating that the coordinate position is basically reliable but has some uncertainty. Low-precision coordinates correspond to coordinate points with confidence scores in the interval of zero to five, indicating that the coordinate position has low reliability and needs further processing. The precision grading processing traverses each coordinate record in the corrected coordinate data set, determines the precision level category according to the confidence score value, and the precision grading data contains coordinate identification, coordinate value, and corresponding precision level identification. The grading result establishes the correspondence between coordinate quality and precision level. The coordinate points in the precision grading data are subjected to neighborhood distance calculation and spatial relationship check processing. Neighborhood distance calculation determines the spatial distribution rationality by measuring the Euclidean distance between the target coordinate point and the surrounding neighboring coordinate points. The calculation process selects neighboring coordinate points within a certain range around the target coordinate point, calculates the straight-line distance between the target point and each neighboring point, and uses the square root of the sum of coordinate differences to calculate the distance. Spatial relationship check determines the logical rationality of the spatial position by analyzing the distance distribution pattern between neighboring coordinate points. The check content includes whether the neighboring distance is too large, whether the neighboring distance is too small, whether the distance distribution is uniform, and other spatial relationship features. Verified coordinates refer to coordinate points whose spatial relationship check results are within a reasonable range. Abnormal coordinates refer to coordinate points found to have problems by spatial relationship check. The verification result data records the check state and abnormal type information of each coordinate point.

[0088] According to the abnormal coordinate mark in the verification result data, missing coordinate identification and error coordinate classification processing are performed. The missing coordinate identification is performed by checking the integrity of the coordinate data to find the case that the coordinate value is empty or obviously exceeds the geographical reasonable range. The missing coordinate includes the record with empty X or Y coordinate value, the record with coordinate value obviously exceeding the measurement area range, and the record with the lowest coordinate accuracy level and the confidence score close to zero. The error coordinate classification is refined according to the specific problem type of the abnormal coordinate mark. The error types include coordinate offset error, coordinate accuracy error, coordinate format error, and other different categories. The coordinate offset error refers to the case that the coordinate position deviates from the reasonable range, but the value format is correct. The coordinate accuracy error refers to the case that the coordinate value accuracy does not meet the requirements. The coordinate format error refers to the case that the coordinate value format does not meet the standard. The abnormal coordinate list records the specific position and missing type of the missing coordinate, and records the specific problem and error classification information of the error coordinate. Based on the abnormal coordinate list, the missing coordinate is filled and the error coordinate is replaced. The missing coordinate value is filled by the statistical method of the surrounding effective coordinate points. The missing X coordinate value is filled by calculating the average value of the X coordinates of the nearest effective coordinate points around the missing coordinate point. The missing Y coordinate value is filled by calculating the average value of the Y coordinates of the nearest effective coordinate points around the missing coordinate point. The error coordinate value is replaced by the correct value verified. The replacement value is obtained by searching the historical coordinate records of the same mining right or adjacent mining right. The replacement process preferentially selects the coordinate value with higher accuracy level and higher confidence score. The repair confidence is evaluated according to the data source quality and processing method reliability of the filling or replacement operation. The coordinate repair data includes the repaired coordinate value, the repair operation type, and the repair confidence score.

[0089] The coordinate repair data and the verification result data are subjected to quality score calculation and error statistical analysis processing. The quality score calculation comprehensively considers the success rate of coordinate repair, the accuracy level of the repaired coordinate, the repair confidence score, and other quality indicators. The calculation method divides the number of successfully repaired coordinates by the total number of coordinates to obtain the repair success rate. The proportion of high-precision coordinates and the proportion of medium-precision coordinates are used as the accuracy distribution indicators. The average value of the repair confidence score is used as the overall reliability indicator. The quality score is obtained by weighted average method to obtain the final data quality score. The error statistical analysis calculates the deviation degree and distribution characteristics of the coordinate position before and after repair. The error calculation includes the mean, variance, maximum deviation, and minimum deviation of the coordinate difference before and after repair. The repair record details the original state, repair operation, repair result, and quality change of each coordinate point. The coordinate error analysis includes error distribution chart, error statistical table, quality improvement report, and other analysis results. The inspection data set integrates the quality score, repair record, and coordinate error analysis to form a complete data quality control file.

[0090] The above describes the special inventory method for key natural resource assets in the embodiments of this application. The following describes the special inventory system for key natural resource assets in the embodiments of this application. Please refer to [link / reference]. Figure 2 One embodiment of the special inventory system for key natural resource assets in this application includes:

[0091] The identification module is used to perform three-dimensional identification processing on mineral data in the reserves database through the resource classification matrix, and obtain a list of key resources including type identification code, level code and priority score;

[0092] The identification module is used to identify coordinates of different formats based on the coordinate data in the key resource list by using regular expression matching, so as to obtain a standardized coordinate set containing the original coordinates, format identifiers and transformation parameters;

[0093] The conversion module is used to perform batch coordinate conversion processing on the standardized coordinate set under a unified coordinate system through a convolutional neural network to obtain target coordinate data.

[0094] The correction module is used to perform neighborhood interpolation correction processing on the center point coordinates based on the coordinate offset in the target coordinate data to obtain the corrected coordinate dataset;

[0095] The repair module is used to perform coordinate accuracy verification and automatic repair of abnormal coordinates on the corrected coordinate dataset, resulting in a cleanup dataset containing quality scores, repair records, and coordinate error analysis.

[0096] above Figure 2 The system for special inventory of key natural resource assets in this embodiment of the invention is described in detail from the perspective of modular functional entities. The special inventory equipment for key natural resource assets in this embodiment of the invention is described in detail from the perspective of hardware processing.

[0097] Reference Figure 3 This invention also provides a special inventory device for key natural resource assets. This device can be a server, and its internal structure can be as follows: Figure 3The special investigation equipment for the key natural resource assets comprises a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. The processor of the computer is used to provide computing and control capabilities. The memory of the special investigation equipment for the key natural resource assets comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium. The database of the special investigation equipment for the key natural resource assets is used to store the corresponding data in the embodiment. The network interface of the special investigation equipment for the key natural resource assets is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement the above method.

[0098] Those skilled in the art can understand that, Figure 3 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the special investigation equipment for the key natural resource assets to which the scheme of the present application is applied.

[0099] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium. The computer readable storage medium stores instructions, and when the instructions are run on a computer, the computer executes the steps of the special investigation method for the key natural resource assets.

[0100] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, system and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0101] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical scheme of the present application or the whole or part of the technical scheme that essentially contributes to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a special investigation equipment for the key natural resource assets (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk and various program code storage media.

[0102] The above examples are only used to illustrate the technical solutions of the present application, but not to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those ordinarily skilled in the art should understand: the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for special investigation of key natural resource assets, characterized in that, The method includes: The mineral data in the reserves database is subjected to three-dimensional identification processing by resource classification matrix to obtain a list of key resources containing type identification code, level code and priority score; Based on the coordinate data in the key resource list, coordinates of different formats are identified and processed by regular expression matching to obtain a standardized coordinate set containing the original coordinates, format identifiers and transformation parameters; The standardized coordinate set is processed by a convolutional neural network to perform batch coordinate transformation under a unified coordinate system to obtain target coordinate data. This process includes: encoding the coordinate data in the standardized coordinate set into a two-dimensional matrix to obtain a coordinate feature matrix containing X-coordinates, Y-coordinates, and format identifiers; extracting local features from the coordinate feature matrix using a convolutional layer to obtain a local feature map containing coordinate offset features and format transformation features; inputting the local feature map into a pooling layer for feature dimensionality reduction and noise filtering to obtain dimensionality-reduced coordinate transformation feature data; calculating coordinate mapping relationships using a fully connected layer based on the coordinate transformation feature data to obtain coordinate transformation parameters containing transformation coefficients and offset parameters; and performing batch coordinate transformation calculations on the standardized coordinate set based on the coordinate transformation parameters to obtain target coordinate data under a unified coordinate system. Based on the coordinate offset in the target coordinate data, the center point coordinates are corrected by neighborhood interpolation to obtain the corrected coordinate dataset. The corrected coordinate dataset is subjected to coordinate accuracy verification and automatic abnormal coordinate repair processing to obtain a cleanup dataset containing quality scores, repair records, and coordinate error analysis.

2. The method of claim 1, wherein, The process involves performing three-dimensional labeling on mineral data in the reserves database using a resource classification matrix to obtain a list of key resources containing type identifiers, hierarchical codes, and priority scores, including: Based on the table of authority for mining rights transfer and registration, the management hierarchy of mineral types is mapped to obtain provincial, municipal and county-level mineral types lists. The mineral type codes in the reserves database are input into the hierarchical mapping table for hierarchical identification processing, resulting in hierarchical identification data containing mineral type codes and management level identifiers; Based on the reserve scale parameters in the hierarchical identification data, the reserve scale coefficient of each mineral type is calculated to obtain the reserve scale scoring data. The reserve scale scoring data is weighted by price fluctuation coefficient based on market price fluctuation parameters to obtain price-weighted scoring data. The price-weighted scoring data is processed by a priority scoring algorithm to obtain a list of key resources that includes type identifiers, level codes, and priority scores.

3. The method of claim 1, wherein, The process involves identifying coordinates of different formats using regular expression matching based on the coordinate data in the key resource list, resulting in a standardized coordinate set containing the original coordinates, format identifiers, and transformation parameters. The coordinate data in the key resource list is scanned for format type to obtain coordinate format classification data in degree-minute-second format, decimal degree format, and projected coordinate. The separator recognition and precision bit analysis processing are performed on the coordinate format classification data based on a regular expression matching library, to obtain format analysis data containing a coordinate separator, precision bits and a symbol identifier; According to the coordinate system identifier in the format analysis data, parameter query matching processing is performed on a historical coordinate system parameter database to obtain coordinate system parameter data containing ellipsoid parameters, projection parameters and reference point parameters; The format analysis data is subjected to unified format conversion processing through a coordinate format standardization algorithm to obtain standard format coordinate data in decimal system representation; Based on the coordinate system parameter data, conversion parameter configuration processing is performed on the standard format coordinate data to obtain a standardized coordinate set containing original coordinates, format identifiers and conversion parameters.

4. The method of claim 1, wherein, The coordinate data in the standardized coordinate set is subjected to two-dimensional matrix encoding processing to obtain a coordinate feature matrix containing X coordinates, Y coordinates and format identifiers, including: The X coordinate values and Y coordinate values in the standardized coordinate set are subjected to numerical normalization processing to obtain normalized X coordinate values and Y coordinate values; According to the format identifier, numerical encoding mapping processing is performed on the coordinate system type to obtain format encoding data containing national coordinate system encoding, regional coordinate system encoding and local coordinate system encoding; Based on the normalized X coordinate values, Y coordinate values and format encoding data, three-dimensional array construction processing is performed to obtain a three-channel coordinate data array; The coordinate data array is subjected to matrix reshaping processing according to a fixed size to obtain a two-dimensional coordinate matrix structure with standard row and column specifications; According to the two-dimensional coordinate matrix structure, pixel value filling and zero padding processing are performed to obtain a coordinate feature matrix containing X coordinates, Y coordinates and format identifiers.

5. The method of claim 1, wherein, According to the coordinate offset in the target coordinate data, neighborhood interpolation correction processing is performed on the center point coordinates to obtain a corrected coordinate data set, including: Based on the center point coordinates and mining right boundary coordinates in the target coordinate data, offset distance calculation processing is performed to obtain coordinate offset data containing radial offset distance and angular offset; According to the offset threshold in the coordinate offset data, abnormal coordinate point screening processing is performed to obtain a list of abnormal center point coordinates that exceed the offset threshold; The coordinate points in the list of abnormal center point coordinates are subjected to adjacent point searching processing through a neighborhood coordinate searching algorithm to obtain neighborhood coordinate data containing adjacent mining right coordinates and distance weights; Based on the neighborhood coordinate data, coordinate position recalculation processing is performed through an inverse distance weighting interpolation algorithm to obtain interpolated correction coordinates containing corrected X coordinates, corrected Y coordinates and confidence scores; The interpolated correction coordinates and the original center point coordinates are subjected to coordinate replacement and data merging processing to obtain a corrected coordinate data set containing corrected coordinates and confidence scores.

6. The method of claim 1, wherein, The corrected coordinate data set is subjected to coordinate precision verification and abnormal coordinate automatic repair processing to obtain an inspection data set containing quality scores, repair records and coordinate error analysis, including: Based on the confidence score in the modified coordinate data set, precision level division processing is performed to obtain precision grading data containing high precision coordinates, medium precision coordinates and low precision coordinates; The coordinate points in the precision grading data are subjected to neighborhood distance calculation and spatial relationship checking processing to obtain verification result data containing verified coordinates and abnormal coordinate markers; According to the abnormal coordinate markers in the verification result data, missing coordinate identification and error coordinate classification processing are performed to obtain an abnormal coordinate list containing coordinate missing records and coordinate error types; Based on the abnormal coordinate list, numerical filling for missing coordinates and numerical replacement for error coordinates are performed to obtain coordinate repair data containing repaired coordinates and repair confidence; The coordinate repair data and the verification result data are subjected to quality score calculation and error statistical analysis processing to obtain an inspection data set containing quality scores, repair records and coordinate error analysis.

7. A system for a special investigation of a key natural resource asset, characterized by, The key natural resource asset-oriented special inspection system for implementing the key natural resource asset-oriented special inspection method as claimed in any one of claims 1 to 6 comprises: An identification module for performing three-dimensional identification processing on the mineral data in the reserve database through a resource classification matrix to obtain a key resource list containing type identification codes, hierarchical codes and priority scores; An identification module for identifying different format coordinates through regular expression matching according to the coordinate data in the key resource list to obtain a standardized coordinate set containing original coordinates, format identifiers and conversion parameters; A conversion module for performing batch coordinate conversion processing of the standardized coordinate set in a unified coordinate system through a convolutional neural network, comprising: performing two-dimensional matrix encoding processing on the coordinate data in the standardized coordinate set to obtain a coordinate feature matrix containing X coordinates, Y coordinates and format identifiers; performing local feature extraction processing on the coordinate feature matrix through a convolutional layer to obtain a local feature map containing coordinate offset features and format conversion features; inputting the local feature map into a pooling layer for feature dimension reduction and noise filtering processing to obtain reduced coordinate conversion feature data; performing coordinate mapping relationship calculation processing on the coordinate conversion feature data through a fully connected layer to obtain coordinate transformation parameters containing conversion coefficients and offset parameters; performing batch coordinate conversion calculation processing on the standardized coordinate set based on the coordinate transformation parameters to obtain target coordinate data in a unified coordinate system; A correction module for performing neighborhood interpolation correction processing on the center point coordinates according to the coordinate offset in the target coordinate data to obtain a modified coordinate data set; A repair module for performing coordinate precision verification and abnormal coordinate automatic repair processing on the modified coordinate data set to obtain an inspection data set containing quality scores, repair records and coordinate error analysis.

8. A specialized inventory device for key natural resource assets, characterized in that, A key natural resource asset-oriented special inspection method comprising a memory and a processor, the memory storing a computer program executable on the processor, and the processor implementing the computer program to realize the key natural resource asset-oriented special inspection method as claimed in any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, causes the processor to perform the special investigation method for key natural resource assets as claimed in any one of claims 1 to 6.

Images