Granularity evaluation method based on SHEPIDED triangulation diagram

Abstract

The invention relates to the field of marine geology, and discloses a particle size evaluation method based on a SHEPIDED triangulation diagram.The method comprises the following steps that S1, a marine geological sample is collected, and particle size data are automatically obtained through a laser particle size analyzer; s2, transmitting the granularity data to a remote server in real time through a wireless communication module; s3, the remote server receives the granularity data and preprocesses the granularity data, wherein the preprocessing comprises detecting and removing outliers; s4, the remote server analyzes the granularity data according to a preset granularity analysis algorithm and feeds back a result to the field device or the user terminal; and S5, analyzing historical granularity data based on a deep learning model. According to the method, missing values or discontinuous sections are intelligently supplemented through a combination mode of K-nearest neighbor interpolation and spline interpolation, so that the data has higher smoothness and consistency while keeping a real change trend, and the effects of continuously supplementing and remarkably enhancing the stability and consistency of the data are achieved.

Description

Technical Field

[0001] This invention relates to the field of marine geology, specifically to a grain size assessment method based on Shepard triangulation. Background Technology

[0002] Grain size analysis is a fundamental task in geological research, especially in marine geology. By studying sediment grain size, we can effectively understand the sedimentary environment, provenance characteristics, and sedimentary processes. Grain size data of marine geological samples typically include three main components: sand, silt, and clay. Shepard ternary diagrams are a common grain size analysis tool because they can visually display the relative abundance of these three components.

[0003] Shepard triangulations, with their unique geometric structure, effectively map grain size data into points in two-dimensional space, helping geologists quickly understand the grain size distribution of samples. This method is widely used in various fields such as marine sediment classification, geological exploration, and environmental monitoring. Through this graphical representation, researchers can analyze the composition of grain size intuitively without complex calculations.

[0004] With the advancement of technology, data acquisition methods are gradually shifting from traditional manual sampling to automated equipment. The large amount of granular data collected on-site requires efficient and standardized processing methods. Among related technologies, interpolation methods, such as linear interpolation, can fill in missing data, but they cannot fully maintain the true trend of data change. The data after filling in the missing data often shows excessive smoothing or unnatural fluctuations, thus affecting the stability of the data. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a granularity evaluation method based on Shepard triangle diagrams, which solves the problem that the true trend of data change cannot be fully preserved, and the data after supplementation often exhibits excessive smoothing or unnatural fluctuations.

[0006] In a first aspect, the present invention provides the following technical solution: a granularity evaluation method based on Shepard triangle diagrams, comprising the following steps:

[0007] S1. Collect marine geological samples and automatically acquire particle size data using a laser particle size analyzer;

[0008] S2. Transmit granular data to a remote server in real time via a wireless communication module;

[0009] S3. The remote server receives granular data and performs preprocessing, including detecting and removing outliers and using an interpolation algorithm to fill data gaps.

[0010] S4. The remote server analyzes the granular data according to the preset granularity analysis algorithm, obtains the granularity distribution results, and feeds the results back to the field equipment or user terminal.

[0011] S5. Based on deep learning models, analyze historical granular data to automatically classify and predict newly collected data, and output the predicted geological sample type and its environmental significance.

[0012] As a further description of the above technical solution: S3 includes the following steps:

[0013] First, determine if there are outliers in the granularity data;

[0014] If outliers are found, an outlier detection procedure is triggered, which uses the Z-Score method to calculate the Z value of each data point and removes outliers according to a preset threshold.

[0015] Based on the data gaps, an interpolation algorithm is triggered to fill the data gaps. The interpolation algorithm is either K-nearest neighbor interpolation or spline interpolation.

[0016] The outlier detection program's judgment logic is as follows: by calculating the Z value of each granularity data point, and then setting a threshold for the Z value, when the Z value is greater than the set threshold, the data is determined to be an outlier and removed.

[0017] The above technical solution automatically triggers K-nearest neighbor interpolation or spline interpolation based on data gaps, so that the missing parts are reasonably filled, avoiding inaccurate granular feature calculation due to breakpoints. The Z-Score is used to calculate the Z value of each granular point, and outliers that deviate too much from the mean are automatically identified and removed, avoiding outliers from skewing the overall granularity distribution and improving the stability of subsequent granularity calculation and triangular diagram mapping.

[0018] As a further description of the above technical solution: S3 also includes the following steps:

[0019] Based on the missing data in the dataset, the interpolation algorithm selected includes:

[0020] If the K-nearest neighbor interpolation algorithm is used, the nearest neighbor sample data is selected for interpolation based on the distance between the samples;

[0021] If spline interpolation is used, a smooth interpolation curve between data points is calculated to fill in the missing data.

[0022] Through the above technical solution, the system can intelligently select a suitable interpolation algorithm based on the missing data in the dataset. K-nearest neighbor interpolation is suitable for sparse data, ensuring that missing data points are filled by weighted average, while spline interpolation can provide a smooth and continuous curve when the data is relatively dense, avoiding abrupt changes during the interpolation process. By selecting an appropriate interpolation algorithm, this technical solution ensures that missing data can be filled quickly and accurately, thereby improving the efficiency and accuracy of the entire data processing flow and providing a solid data foundation for subsequent granularity evaluation.

[0023] As a further description of the above technical solution: S4 includes the following steps:

[0024] By introducing a deep learning model and training it based on historical granular data, it is used to automatically classify and predict the characteristics of different types of geological samples.

[0025] The server collects granular analysis data from multiple regions in the past as a training set.

[0026] Convolutional neural network models are constructed and trained on different granularity distribution features to learn the relationship between sample types and granularity distributions;

[0027] The newly collected granular data is input into the trained model, which outputs the predicted geological sample type and its possible environmental significance.

[0028] The granularity analysis algorithm includes the following steps:

[0029] First, cluster analysis is performed based on the distribution characteristics of the granular data to divide the data into multiple categories;

[0030] Granularity characteristics analysis is performed on each category based on distribution features, and the analysis results are output.

[0031] The above technical solution, by introducing a convolutional neural network model trained on historical granular data, can automatically learn and identify the complex relationships between different types of geological samples and their granular distribution. This deep learning method can efficiently process large-scale data, automatically perform classification and prediction, thereby reducing manual intervention and improving the automation and intelligence of granularity analysis.

[0032] As a further description of the above technical solution: S5 includes the following steps:

[0033] Training on historical granular data based on deep learning models;

[0034] Use the trained model to classify and predict newly acquired granular data;

[0035] Based on the prediction results, output the geological sample types and their environmental impact analysis.

[0036] By using the above technical solutions and training historical granular data with deep learning models, the system can automatically learn the intrinsic relationship between granularity distribution and geological sample type, making the classification and prediction process of new data fully automated and greatly improving analysis efficiency.

[0037] As a further description of the above technical solution: the particle size analysis in S4 also includes:

[0038] Based on the distribution characteristics of the data, a hierarchical clustering algorithm is used to divide granular data into multiple categories;

[0039] Output the granularity characteristics for each category, perform statistical analysis, and provide a granularity distribution map and data summary.

[0040] Through the above technical solutions, the hierarchical clustering algorithm can automatically classify data with similar granularity patterns into the same category based on the distribution characteristics of granular data, making the classification results more detailed and the structure clearer, and significantly improving the accuracy and interpretability of granularity analysis.

[0041] As a further description of the above technical solution: the S5 prediction output includes the following steps:

[0042] Based on the model prediction results, output the geological sample type;

[0043] Output the predicted environmental impact analysis report, including the types of mineral resources present in the sample and their ecological and environmental impacts;

[0044] The geological sample types include sandy and muddy.

[0045] Through the above technical solutions, by combining the granularity and environmental relationship learned by the model, the system can automatically output the types of mineral resources that the sample may be associated with and their potential ecological and environmental impacts, thereby providing timely and reliable auxiliary decision-making basis for geological exploration, mineral development and environmental assessment.

[0046] Secondly, the present invention provides the following technical solution: a particle size assessment system based on Shepard triangle diagram, the system comprising: a data acquisition module for collecting marine geological samples and acquiring particle size data;

[0047] The wireless communication module is used to transmit granular data to a remote server in real time.

[0048] The remote server module is used to receive and process granular data, and perform data preprocessing, granular analysis, and deep learning classification.

[0049] The user terminal module is used to receive analysis results from the remote server module and provide them to the user for decision-making, including on mobile devices or computers.

[0050] Through the above technical solution, the entire particle size assessment system can achieve efficient flow from sample collection to analysis result feedback, improve the processing efficiency and accuracy of marine geological sample data, and provide users with real-time decision support, with a high degree of intelligence and remote operation capabilities.

[0051] Thirdly, the invention provides the following technical solution: a computer device including a memory, a laser particle size analyzer, a wireless communication module, a central processing unit, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned particle size assessment system based on Shepard triangle diagram.

[0052] Fourthly, the present invention provides the following technical solution: a readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-mentioned granularity evaluation system based on Shepard triangle diagram.

[0053] This invention provides a granularity evaluation method based on Shepard triangle diagrams. It has the following beneficial effects:

[0054] 1. In this invention, a combination of K-nearest neighbor interpolation and spline interpolation is used to intelligently fill in missing values ​​or discontinuous segments, so that the data can maintain the true trend of change while having higher smoothness and consistency, achieving continuous filling and significantly enhancing the stability and consistency of the data.

[0055] 2. In this invention, by introducing a Z-Score-based outlier detection mechanism in data preprocessing and combining K-nearest neighbor interpolation and spline interpolation algorithms, dynamic cleaning and continuity completion of granular data are achieved, resulting in a highly stable and consistent granular dataset, thereby effectively improving the reliability of subsequent analysis and model prediction.

[0056] 3. In this invention, by introducing cluster analysis and hierarchical clustering algorithms on the server side, the distribution characteristics of granular data are identified, categories are divided, and granular parameters are statistically analyzed, thereby realizing the structured analysis of samples of different types of granularity and obtaining interpretable granularity categories and characteristic results, thus improving the scientificity and interpretability of geological sample analysis.

[0057] 4. In this invention, by introducing a convolutional neural network model to train historical granular data, automatic classification, feature recognition, and environmental significance prediction of newly collected granular data are achieved, resulting in automated judgment of geological sample types and analysis of their impact on the sedimentary environment, thereby improving the intelligence level and decision support capability of the granularity assessment system. Attached Figure Description

[0058] Figure 1 This is a flowchart of the method of the present invention;

[0059] Figure 2 This is the Shepard triangle diagram of the present invention;

[0060] Figure 3 This is a system architecture diagram of the present invention. Detailed Implementation

[0061] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0062] During a geological survey in a certain sea area, a drone equipped with a laser particle size analyzer was used to immediately transmit the data to a remote server on a marine platform via a 5G network after sampling. The server has a built-in high-precision particle size analysis algorithm that can quickly process and analyze the received particle size data. The particle size distribution results are then pushed to the tablets used by scientists via a cloud platform to facilitate on-site decision-making.

[0063] To better understand the above technical solutions, the technical solutions of the present invention will be clearly and completely described below in conjunction with embodiments.

[0064] Example 1

[0065] Reference Figures 1-2 In the first embodiment of the present invention, a granularity evaluation method based on Shepard triangle diagrams is provided, comprising the following steps:

[0066] Drawing a Shepard triangle involves the following steps:

[0067] First, establish a rectangular coordinate system and draw a base map on the coordinate system. Then, draw the particle size of each sample on the base map according to the particle size data. The specific formula is as follows, where R is the side length of the large equilateral triangle of the Shepard triangle diagram, and L is the particle size data of sand, silt, and clay.

[0068] Shepard Triangle Drawing Method

[0069] Formulas for calculating the coordinates of the vertices of two equilateral triangles, one inside and one outside:

[0070] (0, -3 * R / 2)

[0071] (-1 / 2 * R, 0)

[0072] (1 / 2 * R, 0)

[0073] (0, -3 * R * 3 / 10)

[0074] (-1 / 5 * R, -3 * R / 10)

[0075] (-1 / 5 * R, -3 * R / 10)

[0076] Formula for calculating the coordinates of the endpoints of line segment 25-75:

[0077] (1 / 8 * R, -3 * 3 * R / 8)

[0078] (-1 / 8*R, -3*33*R / 8)

[0079] (-3 / 8 * R, -3 * R / 8)

[0080] (-1 / 4 * R, 0) (1 / 4 * R, 0)

[0081] (3 / 8 * R, -3 * R / 8)

[0082] Formula for calculating the coordinates of the endpoints of line segment 50:

[0083] (0, -3 * 3 * R / 8)

[0084] (-1 / 4 * R, -3 * R / 4)

[0085] (-1 / 10 * R, -3 * R / 5)

[0086] (-5 / 16*R, -1*3*R / 16)

[0087] (0, 0)

[0088] (0, -3 * R / 10)

[0089] (5 / 16*R, -1*33*R / 16)

[0090] (1 / 4 * R, -3 * R / 4)

[0091] (1 / 10 * R, -3 * R / 5)

[0092] Scale coordinate calculation formula

[0093] (1 / 8 * R, -3 * 3 * R / 8)

[0094] (-5 / 32*R, -3*3*R / 8)

[0095] (-9 / 32 * R, -3 * R / 4)

[0096] (-13 / 32 * R, -3 * R / 8)

[0097] (-1 / 4 * R, 1 / 32 * R)

[0098] (0, 1 / 32 * R)

[0099] (1 / 4 * R, 1 / 32 * R)

[0100] (3 / 8 * R, -3 * R / 8)

[0101] (1 / 4 * R, -3 * R / 4)

[0102] Formula for calculating the coordinates of each side label range

[0103] (-1 / 8*R, -1 / 8*R, 1 / 4*R, 1 / 4*R)

[0104] Formula for calculating the coordinates of the text area within the box:

[0105] (-1 / 8*R, -3*R / 2, 1 / 4*R, 1 / 4*R)

[0106] (-1 / 2*R, -3*R / 8, 1 / 4*R, 1 / 4*R)

[0107] (1 / 4*R, -3*R / 8, 1 / 4*R, 1 / 4*R)

[0108] (0, -3*3*R / 8, 1 / 4*R, 1 / 4*R)

[0109] (-1 / 4*R, -3*3*R / 8, 1 / 4*R, 1 / 4*R)

[0110] (-3 / 8*R, -3*R / 4, 1 / 4*R, 1 / 4*R)

[0111] (-1 / 4*R, -3*R / 8, 1 / 4*R, 1 / 4*R)

[0112] (0, -3*R / 8, 1 / 4*R, 1 / 4*R)

[0113] (1 / 8*R, -3*R / 4, 1 / 4*R, 1 / 4*R)

[0114] (-1 / 8*R, -3*R / 4, 1 / 4*R, 1 / 4*R)

[0115] Method for plotting sample particle size in a Shepard triangle

[0116] L sand + L silt + L clay = 100

[0117] ((100 - L clay - 2 * L sand) / 200 * R, -3 * L clay * R / 200).

[0118] S1. Collect marine geological samples and automatically acquire particle size data using a laser particle size analyzer;

[0119] S2. Transmit granular data to a remote server in real time via a wireless communication module;

[0120] S3. The remote server receives granular data and performs preprocessing, including detecting and removing outliers and using an interpolation algorithm to fill data gaps.

[0121] S3 includes the following steps:

[0122] First, determine if there are outliers in the granularity data;

[0123] If outliers are found, an outlier detection procedure is triggered, which uses the Z-Score method to calculate the Z value of each data point and removes outliers according to a preset threshold.

[0124] Based on the data gaps, an interpolation algorithm is triggered to fill the data gaps. The interpolation algorithm is either K-nearest neighbor interpolation or spline interpolation.

[0125] The outlier detection program's judgment logic is as follows: by calculating the Z value of each granularity data point, and then setting a threshold for the Z value, when the Z value is greater than the set threshold, the data is determined to be an outlier and removed.

[0126] S3 further includes the following steps:

[0127] Based on the missing data in the dataset, the interpolation algorithm selected includes:

[0128] If the K-nearest neighbor interpolation algorithm is used, the nearest neighbor sample data is selected for interpolation based on the distance between the samples;

[0129] If spline interpolation is used, a smooth interpolation curve between data points is calculated to fill in the missing data.

[0130] Specifically, outlier detection is first performed on the granular data to identify and remove outliers, improving the accuracy of subsequent analysis. Outliers can be caused by sensor malfunctions, data transmission errors, or natural phenomena. Therefore, the Z-Score method is used to determine whether each data point is an outlier. The outlier status is measured by calculating the standardized deviation between each data point and the dataset mean. The Z-Score formula is as follows:

[0131] ;

[0132] in, It is the first in the granularity dataset Data points, It is the mean of the dataset. It is the standard deviation and mean of the dataset. and standard deviation The results are obtained using the following formulas:

[0133] ;

[0134] By calculating the Z-value of each data point, if the Z-value of a point exceeds a preset threshold, such as 2.5 or 3, the data point is considered an outlier and will be removed from the dataset. When a data point is missing, the system will estimate the missing value based on the context information of the missing point. There are two interpolation methods: K-nearest neighbor interpolation (KNN) and spline interpolation. For KNN interpolation, the system first finds the K nearest neighbor data points near the missing data point, and then estimates the value of the missing data point based on the average of these neighbor points. The interpolation formula is as follows:

[0135] ;

[0136] in, Is with missing points The most recent 1 known data point This refers to the selected number of neighboring points. For more accurate interpolation, the system uses a weighted KNN interpolation method, where closer neighbors are assigned higher weights. The weighting formula is:

[0137] ;

[0138] in, Is it a missing point and the first The distance between neighboring points is considered; points that are closer together have a greater impact. If the range of missing data is large or the data itself exhibits a smooth trend, the system will choose spline interpolation for imputation. Spline interpolation fits a smooth curve, connects the known data points, and estimates the value of the missing data using the curve. The function form of cubic spline interpolation is:

[0139] ;

[0140] in, ,and The coefficients are obtained by using the boundary conditions, continuity conditions, and derivative consistency of known data points, making the entire curve not only smooth but also continuous, and are suitable for situations where the data fluctuations are small and the distribution is uniform.

[0141] By employing two key steps—outlier detection and missing data imputation—a stable and continuous granular dataset was ultimately obtained. This data will serve as the foundation for subsequent analysis and modeling, helping to improve the accuracy and reliability of the analysis results.

[0142] S4. The remote server analyzes the granular data according to the preset granularity analysis algorithm, obtains the granularity distribution results, and feeds the results back to the field equipment or user terminal.

[0143] S4 includes the following steps:

[0144] By introducing a deep learning model and training it based on historical granular data, it is used to automatically classify and predict the characteristics of different types of geological samples.

[0145] The server collects granular analysis data from multiple regions in the past as a training set.

[0146] Convolutional neural network models are constructed and trained on different granularity distribution features to learn the relationship between sample types and granularity distributions;

[0147] The newly collected granular data is input into the trained model, which outputs the predicted geological sample type and its possible environmental significance.

[0148] The granularity analysis algorithm includes the following steps:

[0149] First, cluster analysis is performed based on the distribution characteristics of the granular data to divide the data into multiple categories;

[0150] Granularity characteristics analysis is performed on each category based on distribution features, and the analysis results are output.

[0151] Specifically, the system first trains on historical granular data by introducing a deep learning model to automatically classify and predict the characteristics of different types of geological samples. The server collects historical granular analysis data from multiple regions as a training set. This data contains typical granular distribution patterns from different geological sources and sedimentary environments. By constructing a convolutional neural network model, the system automatically extracts local features and overall patterns from granular distribution graphics or granular parameter sequences. This enables the model to learn the correlation between sample type and granular distribution. The core structure of the convolutional neural network includes convolutional layers for feature extraction, pooling layers for feature compression, and fully connected layers for classification output, thus forming a training framework suitable for geological granular pattern recognition.

[0152] Once training is complete, the system will input the newly collected granular data into the trained model, which will automatically output the predicted geological sample type and further provide the possible sedimentary environment significance of that type. In addition to the deep learning model, it also includes an independent granular analysis algorithm to provide interpretable analytical basis beyond the model classification results. The system will perform cluster analysis based on the distribution characteristics of the granular data, and divide the overall data into multiple granular categories by calculating the similarity between data points based on the granular distribution pattern to identify the internal structure of the data.

[0153] The essence of cluster analysis is to make data within the same category have higher distributional similarity, while showing significant differences between different categories, thus reflecting the structural characteristics of the sample group. After clustering, the system performs granularity characteristic analysis based on the distribution characteristics of each category, including the extraction and interpretation of granularity parameters such as mean, skewness, kurtosis, and sorting coefficient. This achieves the goal of inferring sedimentary dynamics, transport environment, or provenance characteristics based on grain size morphology, and ultimately outputs the granularity analysis results.

[0154] When the classification results of the deep learning model are combined with the results of clustering granularity analysis, the geological sample interpretation provided by the system will have higher reliability. The deep learning part provides automated and patterned classification prediction, while the clustering and granular feature analysis part provides visualized and interpretable geological significance support, thus forming an intelligent granularity analysis framework that is both automated and interpretable.

[0155] The particle size analysis in S4 also includes:

[0156] Based on the distribution characteristics of the data, a hierarchical clustering algorithm is used to divide granular data into multiple categories;

[0157] Output the granularity characteristics for each category, perform statistical analysis, and provide a granularity distribution map and data summary.

[0158] Specifically, hierarchical clustering is performed based on the distribution characteristics of granular data. According to the similarity of each data point in the granular distribution dimension, the most similar sample clusters are gradually merged to form a tree-like classification structure that aggregates from small to large. This corresponds to the basic idea of ​​hierarchical clustering algorithm. By calculating the distance or similarity between any two samples, clustering criteria such as the shortest distance method, the average distance method, or the Ward minimum variance method are used to enable granular data to be naturally grouped according to distribution characteristics.

[0159] After clustering is completed, the system outputs the corresponding granularity characteristics for each category, including statistical analysis of the central tendency, dispersion, and morphological parameters of each granularity distribution, such as typical parameters like mean, skewness, kurtosis, and sorting coefficient. The system also generates a visualized granularity distribution map and category-level data summaries, enabling users to intuitively understand the characteristics of each granularity group. Through histograms, cumulative probability curves, or kernel density curves, the system displays the typical distribution pattern of each category, thereby providing more interpretive auxiliary evidence for subsequent environmental explanations or deep learning predictions.

[0160] S5. Based on deep learning models, analyze historical granular data to automatically classify and predict newly collected data, and output the predicted geological sample type and its environmental significance.

[0161] S5 includes the following steps:

[0162] Training on historical granular data based on deep learning models;

[0163] Use the trained model to classify and predict newly acquired granular data;

[0164] Based on the prediction results, output the geological sample types and their environmental impact analysis.

[0165] Specifically, firstly, the system trains a deep learning model based on historical granular data. The model can automatically identify the complex relationship between granular data and different geological sample types. By using granular data from multiple regions as the training set, which includes granular distribution data of various geological environments and different sedimentary features, the deep learning model can extract meaningful features from the granular data and form the ability to identify features of different granular distributions. The training process of the model relies on a standard convolutional network architecture, which includes multiple convolutional layers, pooling layers and fully connected layers, to extract multi-level granular features and complete the automatic classification of geological sample types.

[0166] Once trained, the deep learning model will be used to classify and predict newly acquired granular data. When new data is input into the model, the model will automatically calculate features similar to historical data and predict the geological sample type that the data point may belong to. It can automatically classify geological samples, thereby improving the efficiency and accuracy of granular data analysis. The deep learning model determines the sample type and its geological characteristics by classifying granular data and predicting new data based on learned association patterns.

[0167] Based on the prediction results output by the model, further environmental impact analysis is conducted, and the environmental significance of this geological sample type is output. By analyzing the relationship between sample type and sedimentary environment, the sedimentation process, transport pathway, and even possible geological evolution background of the sample can be inferred, which will provide key decision support for fields such as geological exploration, resource assessment, and environmental protection.

[0168] The S5 prediction output includes the following steps:

[0169] Based on the model prediction results, output the geological sample type;

[0170] Output the predicted environmental impact analysis report, including the types of mineral resources present in the sample and their ecological and environmental impacts;

[0171] The geological sample types include sandy and muddy.

[0172] Specifically, firstly, the prediction results based on the deep learning model output the type of geological sample. By inferring the newly collected granular data using the model, the prediction results will include geological sample types such as sandy or muddy. The determination of these types is based on the correlation between various granular characteristics in the granular data, such as particle size, shape, and distribution uniformity, and historical data samples, ensuring the accuracy and stability of the classification.

[0173] In addition to geological sample types, the system will also output an environmental impact analysis report based on the prediction results. It will combine the geological characteristics of the samples to analyze their possible mineral resource types and potential impacts on the ecological environment. For sandy samples, the report indicates that the mineral resource type in the area may be sandstone or certain types of metallic ores, while for muddy samples, it may be inferred that the minerals they contain are clay or other sedimentary minerals. The report will further assess the potential impact of these samples on the surrounding ecological environment, such as soil and water conservation capacity and land use impact, to help technical personnel in related fields make more accurate environmental assessments and management decisions.

[0174] By combining the prediction results of various geological samples, we can provide a scientific basis for subsequent geological exploration, mineral resource development, environmental protection and other work, and ensure that relevant activities can be decided based on accurate geological data and environmental prediction results.

[0175] Example 2:

[0176] Reference Figure 3 In a second embodiment of the present invention, the present invention provides a particle size assessment system based on Shepard triangle diagram, the system comprising: a data acquisition module for collecting marine geological samples and acquiring particle size data;

[0177] The wireless communication module is used to transmit granular data to a remote server in real time.

[0178] The remote server module is used to receive and process granular data, and perform data preprocessing, granular analysis, and deep learning classification.

[0179] The user terminal module is used to receive analysis results from the remote server module and provide them to the user for decision-making, including on mobile devices or computers.

[0180] Specifically, the data acquisition module is used to acquire granular data of marine geological samples and transmit the acquired granular data to the wireless communication module. The wireless communication module is responsible for transmitting the granular data to the remote server module in real time via wireless means. After receiving the data, the remote server module performs data preprocessing, granular analysis and deep learning classification. Specifically, it includes outlier detection, missing data imputation and standardization of granular data, visualization analysis of granular distribution through Shepard triangle, and classification and sample type prediction of granular data based on deep learning model.

[0181] Finally, the user terminal module receives the analysis results from the remote server module and displays the classification results of the granular data, the granularity distribution map, and related analysis reports for user decision support. The user terminal module supports access to the system via mobile devices or computers.

[0182] Example 3

[0183] In a third embodiment of the present invention, based on the same inventive concept, the present invention proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a granularity evaluation method based on a Shepard triangle diagram as described in the above embodiments.

[0184] Example 4

[0185] In the fourth embodiment of the present invention, based on the same inventive concept, a computer device is proposed, comprising: a memory, a laser particle size analyzer, a wireless communication module, and a central processing unit; the processor and the memory communicate with each other, the memory is used to store instructions, and the processor is used to execute the instructions in the memory to perform a particle size assessment based on a Shepard triangle diagram as described in the above embodiment.

[0186] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0187] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A granularity evaluation method based on Shepard triangle diagrams, characterized in that, Includes the following steps: S1. Collect marine geological samples and automatically acquire particle size data using a laser particle size analyzer; S2. Transmit granular data to a remote server in real time via a wireless communication module; S3. The remote server receives granular data and performs preprocessing, including detecting and removing outliers and using an interpolation algorithm to fill data gaps. S4. The remote server analyzes the granular data according to the preset granularity analysis algorithm, obtains the granularity distribution results, and feeds the results back to the field equipment or user terminal. S5. Based on deep learning models, analyze historical granular data to automatically classify and predict newly collected data, and output the predicted geological sample type and its environmental significance.

2. The granularity evaluation method based on Shepard triangle diagram according to claim 1, characterized in that, S3 includes the following steps: First, determine if there are outliers in the granularity data; If outliers are found, an outlier detection procedure is triggered, which uses the Z-Score method to calculate the Z value of each data point and removes outliers according to a preset threshold. Based on the data gaps, an interpolation algorithm is triggered to fill the data gaps. The interpolation algorithm is either K-nearest neighbor interpolation or spline interpolation. The outlier detection program's logic is as follows: by calculating the Z value of each granularity data point, and then setting a threshold for the Z value, when the Z value is greater than the set threshold, the data is determined to be an outlier and removed.

3. The granularity evaluation method based on Shepard triangle diagram according to claim 1, characterized in that, S3 further includes the following steps: Based on the missing data in the dataset, the interpolation algorithm selected includes: If the K-nearest neighbor interpolation algorithm is used, the nearest neighbor sample data is selected for interpolation based on the distance between the samples; If spline interpolation is used, a smooth interpolation curve between data points is calculated to fill in the missing data.

4. The granularity evaluation method based on Shepard triangle diagram according to claim 1, characterized in that, S4 includes the following steps: By introducing a deep learning model and training it based on historical granular data, it is used to automatically classify and predict the characteristics of different types of geological samples. The server collects granular analysis data from multiple regions in the past as a training set. Convolutional neural network models are constructed and trained on different granularity distribution features to learn the relationship between sample types and granularity distributions; The newly collected granular data is input into the trained model, which outputs the predicted geological sample type and its possible environmental significance. The granularity analysis algorithm includes the following steps: First, cluster analysis is performed based on the distribution characteristics of the granular data to divide the data into multiple categories; Granularity characteristics analysis is performed on each category based on distribution features, and the analysis results are output.

5. The granularity evaluation method based on Shepard triangle diagram according to claim 1, characterized in that, S5 includes the following steps: Training on historical granular data based on deep learning models; Use the trained model to classify and predict newly acquired granular data; Based on the prediction results, output the geological sample types and their environmental impact analysis.

6. The granularity evaluation method based on Shepard triangle diagram according to claim 4, characterized in that, The particle size analysis in S4 also includes: Based on the distribution characteristics of the data, a hierarchical clustering algorithm is used to divide granular data into multiple categories; Output the granularity characteristics for each category, perform statistical analysis, and provide a granularity distribution map and data summary.

7. The granularity evaluation method based on Shepard triangle diagram according to claim 1, characterized in that, The S5 prediction output includes the following steps: Based on the model prediction results, output the geological sample type; Output the predicted environmental impact analysis report, including the types of mineral resources present in the sample and their ecological and environmental impacts; The geological sample types include sandy and muddy.

8. A granularity evaluation system based on Shepard triangle diagrams, applied to a granularity evaluation method based on Shepard triangle diagrams as described in any one of claims 1-7, characterized in that, The system includes: a data acquisition module, used to collect marine geological samples and obtain particle size data; The wireless communication module is used to transmit granular data to a remote server in real time. The remote server module is used to receive and process granular data, and perform data preprocessing, granular analysis, and deep learning classification. The user terminal module is used to receive analysis results from the remote server module and provide them to the user for decision-making, including on mobile devices or computers.

9. A computer device comprising a memory, a laser particle size analyzer, a wireless communication module, a central processing unit, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a granularity evaluation method based on a Shepard triangle diagram as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements a granularity evaluation method based on a Shepard triangle as described in any one of claims 1 to 7.

Images