Intelligent prospecting prediction method and device based on big data

Abstract

The invention provides an intelligent prospecting prediction method and device based on big data, and the method comprises the steps: constructing an intelligent prospecting prediction model which comprises a feature extraction module, a feature fusion module and an output module, and enabling the feature extraction module to comprise a first feature extraction branch, a second feature extraction branch, a third feature extraction branch and a fourth feature extraction branch; after multi-source geological data of a to-be-detected area is obtained, the multi-source geological data is input into a corresponding feature extraction branch to obtain corresponding features, and then after feature fusion is performed on the obtained features, an intelligent prospecting prediction result is output. The features of the data of each source are extracted through different feature extraction branches, the features extracted from the data of each source are fused, feature loss is avoided, the richness and accuracy of the fused features are improved, and therefore the accuracy of an intelligent prospecting prediction result is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data intelligent prospecting technology, and particularly relates to a big data-based intelligent prospecting prediction method and device. BACKGROUND

[0002] Mineral resources are the cornerstone of modern society, providing indispensable raw materials for energy, industry, high-tech and national security, and are the material basis for infrastructure construction, manufacturing upgrading and energy transformation, directly related to the country's economic lifeline and strategic security. With the rapid development of the global economy, the demand for mineral resources is rising, leading to the development and utilization of mineral resources becoming one of the major problems currently facing the world.

[0003] Currently, mineral prediction has entered the era of "big data intelligence", and the method has developed from traditional geological mapping to deep integration of geology, geophysics, geochemistry and remote sensing data, and widely applied artificial intelligence (such as convolutional neural network) for automatic feature learning and intelligent target prediction, realizing the paradigm shift from experience-driven to data and knowledge-driven, significantly improving the efficiency and accuracy of deep prospecting.

[0004] However, the current mineral prediction model, when used, is to fuse the multi-source data of geology, geophysics, geochemistry and remote sensing as input data into the mineral prediction model for feature extraction. Since different sources of data carry different prospecting information, it is difficult for a single input mineral prediction model to mine deep mineralization information from each source of data, thereby reducing the accuracy of prospecting prediction. SUMMARY

[0005] The present application provides a big data-based intelligent prospecting prediction method and device to solve the problem of low accuracy of prospecting prediction caused by the difficulty of a single input mineral prediction model to mine deep mineralization information from each source of data mentioned in the background.

[0006] In a first aspect, the present application provides a big data-based intelligent prospecting prediction method, comprising: obtaining multi-source geological data of a to-be-tested area, the multi-source geological data comprising geological data, geochemical data, geophysical data and remote sensing data; inputting the geological data into a first feature extraction branch of an intelligent prospecting prediction model to obtain geological features, wherein the intelligent prospecting prediction model comprises a feature extraction module, a feature fusion module and an output module, and the feature extraction module comprises the first feature extraction branch, a second feature extraction branch, a third feature extraction branch and a fourth feature extraction branch; inputting the geochemical data into a second feature extraction branch of the intelligent prospecting prediction model to obtain geochemical features; inputting the geophysical data into a third feature extraction branch of the intelligent ore-prospecting prediction model to obtain geophysical features; inputting the remote sensing data into a fourth feature extraction branch of the intelligent ore-prospecting prediction model to obtain remote sensing features; inputting the geological features, the geochemical features, the geophysical features and the remote sensing features into the feature fusion module to obtain fused features; obtaining an intelligent ore-prospecting prediction result according to the fused features and the output module.

[0007] Optionally, the first feature extraction branch comprises a plurality of cascaded block convolutional neural networks and a first feature fusion unit in series; The inputting the geological data into the first feature extraction branch of the intelligent ore-prospecting prediction model to obtain geological features comprises: inputting the geological data into a first cascaded block convolutional neural network to obtain corresponding geological sub-features; inputting the geological sub-features as input data into a second cascaded block convolutional neural network to obtain corresponding geological sub-features; sequentially circulating, obtaining the geological features according to the corresponding geological sub-features of each cascaded block convolutional neural network.

[0008] Optionally, the inputting the geochemical data into the second feature extraction branch of the intelligent ore-prospecting prediction model and the obtaining geochemical features comprise: extracting a first spatial feature and a correlation feature of chemical elements from the geochemical data according to the second feature extraction branch; obtaining the geochemical features according to the first spatial feature and the correlation feature of chemical elements.

[0009] Optionally, the third feature extraction branch comprises a plurality of data input channels and a feature extraction unit, and input data of the plurality of data input channels are different; The inputting the geophysical data into the third feature extraction branch of the intelligent ore-prospecting prediction model to obtain geophysical features comprises: classifying the geophysical data according to a classification rule to obtain geophysical data of different categories; inputting the geophysical data of different categories into corresponding data input channels respectively, and extracting features of the geophysical data input into the plurality of data input channels according to the feature extraction unit to obtain the geophysical features.

[0010] Optionally, inputting the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features includes: Based on the fourth feature extraction branch, second spatial features and spectral features are extracted from the remote sensing data; The remote sensing features are obtained based on the second spatial features and the spectral features.

[0011] Optionally, acquiring multi-source geological data of the area to be tested includes: Acquire multi-source raw geological data of the area to be measured, including: geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data; The geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data are subjected to data alignment processing to obtain the corresponding multi-source geological data, wherein the geological data, geochemical data, geophysical data, and remote sensing data are a spatiotemporally consistent geological feature matrix.

[0012] Secondly, this application provides an intelligent mineral exploration prediction device based on big data, comprising: The data acquisition module is used to acquire multi-source geological data of the area to be measured. The multi-source geological data includes: geological data, geochemical data, geophysical data, and remote sensing data. The mineral exploration prediction module is used to input geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features, input geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model to obtain geochemical features, input geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features, input remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features, input geological features, geochemical features, geophysical features and remote sensing features into the feature fusion module to obtain fused features, and obtain intelligent mineral exploration prediction results based on the fused features and the output module. The intelligent mineral exploration prediction model includes a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch.

[0013] Optionally, the first feature extraction branch includes multiple cascaded block convolutional neural networks and a first feature fusion unit; The mineral exploration prediction module inputs the geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features, specifically for: The geological data is input into the first cascaded block convolutional neural network to obtain the corresponding geological sub-features. The geological sub-features are input as input data into the second cascaded block convolutional neural network to obtain the corresponding geological sub-features; The geological features are obtained by iterating sequentially based on the geological sub-features corresponding to each cascaded block convolutional neural network.

[0014] Optionally, the mineral exploration prediction module inputs the geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model, and the acquisition of geochemical features is specifically used for: Based on the second feature extraction branch, the first spatial feature and the correlation feature of chemical elements are extracted from the geochemical data; The geochemical features are obtained based on the first spatial features and the correlation features of the chemical elements.

[0015] Optionally, the third feature extraction branch includes: multiple data input channels and a feature extraction unit, wherein the input data of the multiple data input channels are different; The mineral exploration prediction module inputs the geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features, specifically used for: According to the classification rules, the geophysical data is classified to obtain different categories of geophysical data; The different categories of geophysical data are input into their respective data input channels. The feature extraction unit then extracts features from the geophysical data input into the multiple data input channels to obtain the geophysical features.

[0016] Optionally, the mineral exploration prediction module inputs the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features, specifically for: Based on the fourth feature extraction branch, second spatial features and spectral features are extracted from the remote sensing data; The remote sensing features are obtained based on the second spatial features and the spectral features.

[0017] Optionally, the mineral exploration prediction module acquires multi-source geological data of the area to be tested, specifically for: Acquire multi-source raw geological data of the area to be measured, including: geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data; The geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data are subjected to data alignment processing to obtain the corresponding multi-source geological data, wherein the geological data, geochemical data, geophysical data, and remote sensing data are a spatiotemporally consistent geological feature matrix.

[0018] Thirdly, this application provides an electronic device, including: a processor and a memory; The memory stores the instructions that the computer executes; The processor executes computer execution instructions stored in memory, causing the processor to perform the method as described in any of the first aspects.

[0019] Fourthly, embodiments of this application provide a readable storage medium including a program or instructions that, when run on a computer, execute the method described in any of the first aspects above.

[0020] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the first aspects.

[0021] The intelligent mineral exploration prediction method and apparatus based on big data provided in this application constructs an intelligent mineral exploration prediction model including a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes four branches: a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch. After acquiring multi-source geological data (including geological data, geochemical data, geophysical data, and remote sensing data) of the area to be measured, the geological data is input into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features; the geochemical data is input into the second feature extraction branch to obtain geochemical features; the geophysical data is input into the third feature extraction branch to obtain geophysical features; and the remote sensing data is input into the fourth feature extraction branch to obtain remote sensing features. The geological, geochemical, geophysical, and remote sensing features are then input into the feature fusion module to obtain fused features. Based on the fused features and the output module, the intelligent mineral exploration prediction result is obtained. This invention enables the in-depth mining of mineral exploration information from different data through different feature extraction branches, extracting corresponding features to avoid feature loss, and fusing features extracted from each source through a feature fusion module. Compared with the existing technology that first fuses multi-source geological data and then extracts features, this application improves the richness and accuracy of the fused features, thereby improving the accuracy of intelligent mineral exploration prediction results. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating an embodiment of the intelligent mineral exploration prediction method based on big data provided in this application; Figure 2 This is a schematic diagram of the structure of an intelligent mineral exploration prediction model provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of the first feature extraction branch provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of the second feature extraction branch provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of the third feature extraction branch provided in an embodiment of this application; Figure 6 A schematic diagram of the structure of an intelligent mineral prospecting prediction device based on big data provided in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application are described clearly and completely below. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are also within the scope of protection of this application.

[0025] Artificial intelligence, with its powerful algorithms, can rapidly analyze massive amounts of geological data, uncovering hidden patterns and characteristics. This has spurred the rapid development of AI in mathematical geosciences and mineral exploration, providing more scientific and efficient technical means for mineral exploration. However, existing mineral prediction models typically use a single input channel, fusing multi-source data from geology, geophysics, geochemistry, and remote sensing before inputting it into the model, as seen in patents with application numbers 202510131918.8 and 202510173745.6. Because the fused multi-source data weakens the mineral exploration information carried by different sources, the mineral prediction model struggles to extract deeper mineralization information when mining the fused data, thus reducing the accuracy of mineral exploration.

[0026] Therefore, to address the technical problems existing in the prior art, this application proposes an intelligent mineral exploration prediction method and apparatus based on big data. According to the characteristics of the mineral exploration information carried in data from different sources, effective information and corresponding features are extracted from the data. Since the mineral exploration information carried in data from each source is different, extracting the mineral exploration features of each source separately can deeply mine the mineralization information in the data from each source. Thus, after fusing the features of data from each source, the fused features contain the features from each source, thereby improving the accuracy of mineral exploration prediction when outputting the prediction results through these features.

[0027] Figure 1 A flowchart illustrating a big data-based intelligent mineral exploration prediction method provided in an embodiment of this application. Figure 1 The execution subject of the method shown can be an electronic device such as a server that has built an intelligent mineral exploration prediction platform, such as... Figure 1 As shown, the method includes: S101. Obtain multi-source geological data of the area to be measured. Multi-source geological data includes: geological data, geochemical data, geophysical data, and remote sensing data.

[0028] In this step, the parameters included for each source of data are as follows: Geological data mainly includes spatial distribution information of strata and lithology, occurrence and properties of tectonic elements (such as faults, folds and joints), morphology and composition of magmatic rock bodies, as well as field observation records of mineral deposits and mineral occurrences and detailed logging of borehole cores. Geological data can be used to obtain the geological background of mineralization and the factors controlling mineralization.

[0029] Geochemical data mainly includes the content data of ore-forming elements and associated indicator elements in rocks, soils, stream sediments, water samples, and gases. Through these data, the migration and enrichment patterns of elements in the region can be obtained, thereby indicating the location and extent of mineralization.

[0030] Geophysical data mainly includes field data obtained from gravity, magnetic, electrical, seismic, and radioactive exploration. Geophysical data can be used to invert the three-dimensional distribution of physical properties such as density, magnetism, electrical properties, and elastic wave velocity of underground rocks, enabling the detection of concealed rock masses, geological structures, and the direct search for ore bodies with unique physical properties.

[0031] Remote sensing data encompasses multispectral, hyperspectral, thermal infrared, and synthetic aperture radar imagery. By performing spectral interpretation and spatial analysis on these images, alteration mineral information (such as iron oxides and clay minerals) of surface materials can be extracted, linear and ring structures can be identified, and mineralization-related anomaly zones and favorable structural frameworks can be rapidly delineated over a large area.

[0032] For the multi-source geological data here, since it needs to be input into the intelligent mineral exploration prediction model, the data format of this multi-source geological data—namely, geological data, geochemical data, geophysical data, and remote sensing data—needs to be recognizable and learnable by the intelligent mineral exploration prediction model. Therefore, for each source of data, the raw data needs to be processed to transform it into a data format that can be recognized and learned by the intelligent mineral exploration prediction model.

[0033] Therefore, S101 can be implemented as follows: S1011. Obtain multi-source geological raw data for the area to be measured. Multi-source geological raw data includes: geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data.

[0034] Specifically, geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data refer to the original form of the corresponding data, such as images, text, and numerical values. These data can be obtained through various means, including: data acquisition equipment, historical text and images, and experiments.

[0035] Among them, the data obtained through data acquisition equipment includes, for example, the gravity, magnetic and electrical data in the geophysical raw data can be obtained by the CG-6 gravimeter, G-858 cesium magnetometer and V8 electrical resistivity meter carried by the Geophysical Field Integrated Detector, respectively, and the remote sensing raw data are remote sensing images and radar images obtained by remote sensing equipment.

[0036] Historically, when the area to be surveyed was explored, relevant texts and pictures were recorded, and corresponding multi-source original geological data were obtained through these historical texts and pictures.

[0037] Experiments can yield some parameters from raw geochemical data (e.g., the content of ore-forming elements and associated indicator elements in water samples and gases), as well as geological data (e.g., the properties of rock strata).

[0038] This embodiment does not limit the method of using multi-source geological raw data.

[0039] S1012. Perform data alignment processing on the original geological data, geochemical data, geophysical data, and remote sensing data to obtain corresponding multi-source geological data. Among them, the geological data, geochemical data, geophysical data, and remote sensing data are a spatiotemporally consistent geological feature matrix.

[0040] Specifically, because the original formats of multi-source geological raw data differ, including images, text, and numerical values, after obtaining the multi-source geological raw data, the first step is to convert these different formats into a unified format, generally numerical. During the conversion to numerical form, geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data can be aligned according to location. For example, for remote sensing data, for the same location, the data information of each remote sensing raw data source at that location is obtained and converted into numerical values, thus converting remote sensing raw data represented in different formats into remote sensing raw data represented numerically.

[0041] After converting the raw geological, geochemical, geophysical, and remote sensing data into numerical representations, data alignment processing is required to ensure consistency among these four sources of raw geological data. This will result in spatiotemporally consistent multi-source geological data. This alignment can be achieved through methods such as gridding, missing value imputation, outlier detection, and data standardization, which will not be elaborated upon here.

[0042] S102. Input the geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features.

[0043] The intelligent mineral exploration prediction model includes a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch.

[0044] S103. Input the geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model to obtain geochemical features.

[0045] S104. Input the geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features.

[0046] S105. Input the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features.

[0047] For S102-S105, the existing intelligent mineral exploration prediction model only contains one feature extraction branch. Thus, when multi-source geological data is input into the existing intelligent mineral exploration prediction model, the multi-source geological data is first fused, and then features are extracted through a feature extraction branch. Since the multi-source geological data is fused before feature extraction, fewer features are extracted when features are extracted through the feature extraction branch.

[0048] The structure of the intelligent mineral exploration prediction model in this embodiment is as follows: Figure 2 As shown, it includes feature extraction branches that correspond one-to-one with the model's input data (i.e., geological data, geochemical data, geophysical data, and remote sensing data), as well as a feature fusion module and an output module. Compared with existing intelligent mineral exploration prediction models, it has multiple feature extraction branches. Each feature extraction branch is used to extract features from the corresponding data. Then, the features extracted by each feature extraction branch are fused. Extracting features separately results in more features being extracted, avoiding feature loss, and thus making the fused features richer and improving the accuracy of mineral exploration prediction.

[0049] The training method of the intelligent mineral exploration prediction model provided in this embodiment can refer to existing technologies, and will not be described in detail here.

[0050] S106. Input geological features, geochemical features, geophysical features and remote sensing features into the feature fusion module to obtain the fused features.

[0051] In this step, the feature fusion module can achieve feature fusion by either element-wise addition or channel stitching when fusing geological features, geochemical features, geophysical features, and remote sensing features. Figure 2 The diagram shows the channel splicing method.

[0052] S107. Based on the fused features and output modules, obtain intelligent mineral exploration prediction results.

[0053] In this step, the intelligent mineral exploration prediction results may include the mineralization location, predicted resource quantity, and mineral type in the area to be tested, which are specifically related to the output results of the intelligent mineral exploration prediction model during training.

[0054] In this embodiment, an intelligent mineral exploration prediction model is constructed, comprising a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes four branches: a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch. After acquiring multi-source geological data (including geological data, geochemical data, geophysical data, and remote sensing data) of the area to be measured, the geological data is input into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features. The geochemical data is input into the second feature extraction branch to obtain geochemical features. The geophysical data is input into the third feature extraction branch to obtain geophysical features. The remote sensing data is input into the fourth feature extraction branch to obtain remote sensing features. The geological, geochemical, geophysical, and remote sensing features are then input into the feature fusion module to obtain fused features. Based on the fused features and the output module, the intelligent mineral exploration prediction result is obtained. This invention enables the in-depth mining of mineral exploration information from different data through different feature extraction branches, extracting corresponding features to avoid feature loss, and fusing features extracted from each source through a feature fusion module. Compared with the existing technology that first fuses multi-source geological data and then extracts features, this application improves the richness and accuracy of the fused features, thereby improving the accuracy of intelligent mineral exploration prediction results.

[0055] Optionally, one implementation of S102 is as follows: S1021. Based on the geological data, input it into the first cascaded block convolutional neural network to obtain the corresponding geological sub-features.

[0056] S1022. Input the geological sub-features as input data into the second cascaded block convolutional neural network to obtain the corresponding geological sub-features.

[0057] S1023. Repeat the process sequentially to obtain geological features based on the geological sub-features corresponding to each cascaded block convolutional neural network.

[0058] For S1021-S1023, specifically, the structure of the first feature extraction branch is as follows: Figure 3 As shown, it includes multiple cascaded block convolutional neural networks (i.e., cascaded block CNNs) and a first feature fusion unit, with the output of each cascaded block CNN connected to the first feature fusion unit. Each cascaded block CNN has the same structure, including convolutional layers, pooling layers, and connection layers; specific details can be found in existing technologies. Figure 3 The structure of the first feature extraction branch shown includes three cascaded CNN blocks. The number of cascaded CNN blocks can be set according to actual needs.

[0059] based on Figure 3The structure of the first feature extraction branch shown is used to extract data features from geological data to obtain geological features.

[0060] First, the geological data is input into the first cascaded block CNN, which extracts features from the geological data to obtain the corresponding geological sub-features.

[0061] Then, the first-level CNN block uses the corresponding geological sub-features as input data for the second-level CNN block to extract features and obtain the corresponding geological sub-features.

[0062] Following the above method, the geological sub-features extracted by each cascaded block CNN are used as the input data of the next cascaded block CNN for feature extraction to obtain geological sub-features. Finally, the geological sub-features of each cascaded block CNN are fused in the first feature fusion unit. This feature fusion is an element-wise addition to obtain geological features.

[0063] In this embodiment, by designing the structure of the first feature extraction branch as a series of cascaded CNN blocks, the output data of the previous cascaded CNN block is used as the input data of the next cascaded CNN block for feature extraction. This achieves residual cascading, improves the depth of feature extraction, and fuses the geological sub-features of each cascaded CNN block while preserving the features of the original address data as much as possible, making the geological features contain richer and more complete geological mineralization information.

[0064] Alternatively, one implementation of S103 is as follows: S1031. Based on the second feature extraction branch, extract the first spatial feature and the correlation feature of chemical elements from the geochemical data.

[0065] S1032. Based on the spatial characteristics and the correlation characteristics of chemical elements, geochemical characteristics are obtained.

[0066] For S1031-S1032, when analyzing geochemical data, the analysis generally involves examining changes in the chemical elemental composition, identifying correlations between elements, and analyzing the spatial distribution of geochemical variables to predict mineral exploration. Therefore, this implementation designs the second feature extraction branch as follows: Figure 4 The structure shown includes: a first spatial feature extraction channel, a chemical element connection feature extraction channel, and a second feature fusion unit. The first spatial feature and the chemical element connection feature are extracted through the first spatial feature extraction channel and the chemical element connection feature extraction channel, respectively.

[0067] Specifically, the first spatial feature extraction channel extracts features from the input geochemical data to obtain the first spatial features. T specThis represents the spatial distribution and variation of geochemical data; the chemical element association feature extraction channel extracts features from the input geochemical data to obtain chemical element association features. T corr This indicates a deep connection between the characteristics of chemical element associations and the formation process of mineral deposits.

[0068] Then, the first spatial features T spec Characteristics related to chemical elements T corr The second feature fusion unit passes through a fully connected layer f gc By fusing the samples, geochemical characteristics can be obtained. T gc The process of extracting geochemical features in the second feature extraction branch is expressed by the following formula: T gc = f gc ( W 1·( T spec || T corr )+ b 1) Formula 1 Where W1 and b1 represent the weights and biases of the fully connected layer in the second feature fusion unit, respectively, and "||" represents feature fusion.

[0069] In this embodiment, by means of... Figure 4 When the second feature extraction branch shown extracts geochemical features, it can obtain the first spatial features and chemical element relationship features. It not only obtains the relationship features between elements, but also the spatial characteristics of elements, thereby maximizing the extraction of effective information from the geochemical dataset.

[0070] Alternatively, one implementation of S104 is as follows: S1041. According to the classification rules, geophysical data are classified to obtain different categories of geophysical data.

[0071] S1042. Different types of geophysical data are input into the corresponding data input channels. The feature extraction unit extracts features from the geophysical data input into multiple data input channels to obtain geophysical features.

[0072] For S1041-S1042, the structure of the third feature extraction branch is as follows: Figure 5As shown, the system includes multiple data input channels and a feature extraction unit, with different input data for each channel. Specifically, geophysical data is classified, with classification rules based on data type, such as gravity, magnetic, and electrical data. These classifications are then input into their respective data input channels. Each pixel in each channel corresponds to the same location on the Earth's surface. The feature extraction unit extracts features from the different categories of data obtained through the multiple input channels to acquire geophysical features.

[0073] In this embodiment, different geophysical data are input through multiple data input channels, which transforms the original and mixed geophysical data into multiple independent feature channels with clear and complementary geophysical significance. This reduces the difficulty of geophysical feature extraction in the model, improves the expressive ability of the extracted geophysical features, and thus effectively improves the accuracy and reliability of intelligent mineral exploration.

[0074] Alternatively, one implementation of S105 is as follows: S1051. Based on the fourth feature extraction branch, extract the second spatial feature and spectral feature from the remote sensing data; S1052. Obtain remote sensing features based on the second spatial features and spectral features.

[0075] For S1051-S1052, the high spatial resolution of remote sensing data provides rich spatial information, and its spectral information plays a crucial role in identifying surface materials, including rock alteration and mineral composition. Therefore, the fourth feature extraction branch is designed to include a second spatial feature extraction channel, a spectral feature extraction channel, and a third feature fusion unit. A detailed structural diagram can be found in the provided diagram. Figure 4 .

[0076] Specifically, remote sensing data mainly refers to remote sensing images. These images are input into a second spatial feature extraction channel and a spectral feature extraction channel. The second spatial feature extraction channel divides the remote sensing image into image patches and extracts second spatial features. F spec The spectral feature extraction channel includes convolutional layers, activation functions, pooling layers, and normalization operations, used to extract spectral features from remote sensing images. F spat .

[0077] Second space features F spec Spectral characteristics F spat In the third feature fusion unit, a fully connected layer is used. f rs Achieve feature fusion to obtain remote sensing features Frs The fourth feature extraction branch extracts remote sensing features. F rs The formula is: F rs = f rs ( W 2·( F spec || F spat )+ b 2) Formula 2 in, W 2. b 2 represents the weights and biases of the fully connected layer in the third feature fusion unit.

[0078] In this embodiment, by designing the fourth feature extraction branch to include a second spatial feature extraction channel, a spectral feature extraction channel, and a third feature fusion unit, spectral information and spatial information in remote sensing data can be obtained separately, thereby achieving effective fusion of spatial and spectral features.

[0079] Figure 6 This is a schematic diagram of the structure of an intelligent mineral prospecting prediction device based on big data provided in an embodiment of this application, as shown below. Figure 6 As shown, the intelligent mineral exploration prediction device based on big data includes: a data acquisition module 601 and a mineral exploration prediction module 602.

[0080] Among them, the data acquisition module 601 is used to acquire multi-source geological data of the area to be measured. The multi-source geological data includes: geological data, geochemical data, geophysical data, and remote sensing data. The mineral exploration prediction module 602 is used to input geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features, input geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model to obtain geochemical features, input geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features, input remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features, input geological features, geochemical features, geophysical features and remote sensing features into the feature fusion module to obtain fused features, and obtain intelligent mineral exploration prediction results based on the fused features and the output module. The intelligent mineral exploration prediction model includes a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch.

[0081] Optionally, the first feature extraction branch includes multiple cascaded block convolutional neural networks and a first feature fusion unit; The mineral exploration prediction module 602 inputs the geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features, specifically for: The geological data is input into the first cascaded block convolutional neural network to obtain the corresponding geological sub-features. The geological sub-features are input as input data into the second cascaded block convolutional neural network to obtain the corresponding geological sub-features; The geological features are obtained by iterating sequentially based on the geological sub-features corresponding to each cascaded block convolutional neural network.

[0082] Optionally, the mineral exploration prediction module 602 inputs the geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model, wherein obtaining the geochemical features is specifically used for: Based on the second feature extraction branch, the first spatial feature and the correlation feature of chemical elements are extracted from the geochemical data; The geochemical features are obtained based on the first spatial features and the correlation features of the chemical elements.

[0083] Optionally, the third feature extraction branch includes: multiple data input channels and a feature extraction unit, wherein the input data of the multiple data input channels are different; The mineral exploration prediction module 602 inputs the geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features, specifically for: According to the classification rules, the geophysical data is classified to obtain different categories of geophysical data; The different categories of geophysical data are input into their respective data input channels. The feature extraction unit then extracts features from the geophysical data input into the multiple data input channels to obtain the geophysical features.

[0084] Optionally, the mineral exploration prediction module 602 inputs the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features, specifically for: Based on the fourth feature extraction branch, second spatial features and spectral features are extracted from the remote sensing data; The remote sensing features are obtained based on the second spatial features and the spectral features.

[0085] Optionally, the mineral exploration prediction module 602 acquires multi-source geological data of the area to be tested, specifically for: Acquire multi-source raw geological data of the area to be measured, including: geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data; The geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data are subjected to data alignment processing to obtain the corresponding multi-source geological data, wherein the geological data, geochemical data, geophysical data, and remote sensing data are a spatiotemporally consistent geological feature matrix.

[0086] The intelligent mineral exploration prediction device based on big data provided in this application embodiment can be referred to the above method embodiment for its specific implementation process. The implementation principle and technical effect are similar, and will not be repeated here.

[0087] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may be a server, such as... Figure 7 As shown, the electronic device includes a processor 701 and a memory 702.

[0088] The memory 702 stores computer-executed instructions.

[0089] The processor 701 executes the computer execution instructions stored in the memory 702, causing the processor 701 to perform the method described in any of the above embodiments.

[0090] The electronic device provided in this application embodiment can be referred to the above method embodiment for its specific implementation process. The implementation principle and technical effect are similar, and will not be repeated here.

[0091] In the above Figure 7 In the illustrated embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0092] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage.

[0093] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0094] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method shown in the above-described method embodiments.

[0095] The aforementioned computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0096] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0097] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A smart mineral exploration prediction method based on big data, characterized in that, include: Acquire multi-source geological data of the area to be measured, including geological data, geochemical data, geophysical data, and remote sensing data; The geological data is input into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features. The intelligent mineral exploration prediction model includes a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch. The geochemical data is input into the second feature extraction branch of the intelligent mineral exploration prediction model to obtain geochemical features; The geophysical data is input into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features; The remote sensing data is input into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features; The geological features, geochemical features, geophysical features, and remote sensing features are input into the feature fusion module to obtain the fused features; Based on the fused features and the output module, intelligent mineral exploration prediction results are obtained.

2. The method according to claim 1, characterized in that, The first feature extraction branch includes multiple cascaded block convolutional neural networks and a first feature fusion unit; The step of inputting the geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features includes: The geological data is input into the first cascaded block convolutional neural network to obtain the corresponding geological sub-features. The geological sub-features are input as input data into the second cascaded block convolutional neural network to obtain the corresponding geological sub-features; The geological features are obtained by iterating sequentially based on the geological sub-features corresponding to each cascaded block convolutional neural network.

3. The method according to claim 1, characterized in that, The step of inputting the geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model, and obtaining geochemical features, includes: Based on the second feature extraction branch, the first spatial feature and the correlation feature of chemical elements are extracted from the geochemical data; The geochemical features are obtained based on the first spatial features and the correlation features of the chemical elements.

4. The method according to claim 1, characterized in that, The third feature extraction branch includes: multiple data input channels and a feature extraction unit, wherein the input data of the multiple data input channels are different; The step of inputting the geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features includes: According to the classification rules, the geophysical data is classified to obtain different categories of geophysical data; The different categories of geophysical data are input into their respective data input channels. The feature extraction unit then extracts features from the geophysical data input into the multiple data input channels to obtain the geophysical features.

5. The method according to claim 1, characterized in that, The step of inputting the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features includes: Based on the fourth feature extraction branch, second spatial features and spectral features are extracted from the remote sensing data; The remote sensing features are obtained based on the second spatial features and the spectral features.

6. The method according to any one of claims 1-5, characterized in that, The acquisition of multi-source geological data for the area to be tested includes: Acquire multi-source raw geological data of the area to be measured, including: geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data; The geological raw data, geochemical raw data, geophysical raw data, and remote sensing raw data are subjected to data alignment processing to obtain the corresponding multi-source geological data, wherein the geological data, geochemical data, geophysical data, and remote sensing data are a spatiotemporally consistent geological feature matrix.

7. A smart mineral exploration prediction device based on big data, characterized in that, include: The data acquisition module is used to acquire multi-source geological data of the area to be measured, including geological data, geochemical data, geophysical data, and remote sensing data. The mineral exploration prediction module is used to input the geological data into the first feature extraction branch of the intelligent mineral exploration prediction model to obtain geological features, input the geochemical data into the second feature extraction branch of the intelligent mineral exploration prediction model to obtain geochemical features, input the geophysical data into the third feature extraction branch of the intelligent mineral exploration prediction model to obtain geophysical features, input the remote sensing data into the fourth feature extraction branch of the intelligent mineral exploration prediction model to obtain remote sensing features, input the geological features, geochemical features, geophysical features and remote sensing features into the feature fusion module to obtain fused features, and obtain intelligent mineral exploration prediction results based on the fused features and the output module. The intelligent mineral exploration prediction model includes a feature extraction module, a feature fusion module, and an output module. The feature extraction module includes a first feature extraction branch, a second feature extraction branch, a third feature extraction branch, and a fourth feature extraction branch.

8. An electronic device, characterized in that, include: Processor and memory; Memory is used to store instructions executed by the computer; A processor for executing computer execution instructions stored in memory, causing the processor to perform the method according to any one of claims 1-6.

9. A readable storage medium, characterized in that, include: A program or instruction that, when run on a computer, executes the method described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method as described in any one of claims 1-6.

Images