Coverage area iron-rich ore information analysis method, device and equipment and storage medium

Abstract

The invention relates to the technical field of mineralization analysis, in particular to a coverage area iron-rich ore information analysis method, device and equipment and a storage medium, and the method integrates multi-dimensional information such as gravity data, magnetic aviation delta T data, remote sensing images and geological data of a to-be-predicted area, and constructs a feature system capable of comprehensively reflecting mineralization conditions. The limitation of single data source analysis is broken through, a rich and accurate data basis is provided for intelligent division of the metallogenic target area, and the accuracy of metallogenic prediction is remarkably improved; furthermore, according to the method, a fracture speculation diagram, a regional magnetic anomaly plane diagram, a magnetic body three-dimensional probability model and a hydroxyl anomaly distribution vector diagram are adopted for collaborative analysis, so that a rich and accurate data basis is provided for intelligent division of a metallogenic target region, and the accuracy of metallogenic prediction is further improved; the recognition systematicness of the hidden ore body is obviously enhanced; compared with a traditional method depending on artificial experience, the method has the advantage that the objectivity of metallogenic prediction is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of mineralization analysis, and in particular to a method, device, equipment and storage medium for analyzing information of rich iron ore in a coverage area. Background Art

[0002] In the field of iron-rich ore resource exploration in covered areas, regional geological structure analysis and gravity field characteristic analysis are the basic work supporting target area delineation and mineralization potential evaluation.

[0003] Regional structural analysis aims to reveal the regional structural framework, evolutionary history and deep geological mechanism through systematic analysis of the morphology, occurrence, contact relationship and tectonic activity relics of geological bodies, providing key geological basis for identifying ore-controlling structures and locating hidden mineralized zones. However, traditional structural analysis methods are highly dependent on field geological mapping, outcrop observation, drilling sampling and indoor petrological and structural analysis. They are not only time-consuming, labor-intensive and costly, but also show obvious deficiencies in the detection of covered areas and deep hidden structures. It is difficult to quickly obtain regional-scale structural distribution characteristics, which restricts the efficiency and accuracy of iron-rich ore target prediction.

[0004] Gravity exploration, as an important geophysical means of regional geological research, can indirectly reflect the spatial distribution of underground geological bodies of different densities by capturing subtle changes in the Earth's gravity field, namely gravity anomalies, and provide effective clues for inferring deep structural boundaries, dividing geological units, and identifying concealed rock bodies and mineralized bodies. Its data has the advantages of wide coverage, large detection depth, and relatively low acquisition cost, and is widely used in the exploration of rich iron ore. However, faced with massive gravity field data, traditional processing and interpretation processes have significant bottlenecks: on the one hand, data processing involves multiple links such as terrain correction, anomaly separation, and feature extraction, which are cumbersome and highly professional. On the other hand, the interpretation process is highly dependent on manual experience, and different people are prone to subjective differences in the identification and inference of anomalies, which makes it difficult to ensure the objectivity and efficiency of the initial screening of the gravity field, which directly affects the rapid identification of favorable target areas for rich iron ore in the covered area.

[0005] It can be seen that the existing technology still needs to be improved and enhanced. Summary of the Invention

[0006] In order to fill the gaps in the existing technology, the purpose of the present invention is to provide a method for analyzing iron-rich ore information in coverage areas, break through the limitations of traditional single data prediction, and significantly improve the accuracy and reliability of mineralization target area prediction through progressive fusion of multi-source data.

[0007] The first aspect of the present invention provides a method for analyzing iron-rich ore information in a covered area, comprising: obtaining gravity data of the area to be predicted and processing the data to obtain information on high-value gravity areas, and generating a fault speculation map based on the obtained information on high-value gravity areas; obtaining magnetic navigation ΔT data of the area to be predicted and processing the data to generate a regional magnetic anomaly plan map; confirming the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, obtaining magnetic profile data of the mineralization potential area and processing the data to obtain a three-dimensional probability model of the magnetic body; obtaining remote sensing images of the area to be predicted and processing the data to generate a hydroxyl anomaly distribution vector map; obtaining geological data of the area to be predicted, and generating a mineralization probability target area map based on the geological data, the fault speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body, and the hydroxyl anomaly distribution vector map.

[0008] Optionally, in a first implementation method of the first aspect of the present invention, the gravity data of the area to be predicted is obtained and processed to obtain information on high-value gravity areas, and a fault speculation map is generated based on the obtained information on high-value gravity areas, including: obtaining 1:50,000 gravity data of the area to be predicted after terrain interference processing; using db5 wavelet to perform a three-layer decomposition of the gravity data to separate regional fields with wavelengths greater than 10 km and local fields with wavelengths in the range of 2-5 km; obtaining a regional geological map, verifying the separated local fields based on the regional geological map, and obtaining information on high-value gravity areas; calculating the horizontal total gradient of the local field, and based on the calculated horizontal total gradient, using the Canny algorithm to extract the gradient mutation zone of the local field; and generating a fault speculation map based on the extracted gradient mutation zone.

[0009] Optionally, in a second implementation manner of the first aspect of the present invention, the obtaining of the magnetic aerobatic ΔT data of the area to be predicted and processing the data to generate a regional magnetic anomaly plan map includes: obtaining 1:50,000 magnetic aerobatic ΔT data of the area to be predicted, where ΔT is the total magnetic field intensity anomaly; performing polarization processing and upward extension 5 km processing on the magnetic aerobatic ΔT data to obtain processed magnetic anomaly data; calculating the horizontal total gradient of the processed magnetic anomaly data, and extracting the gradient mutation band based on the calculated horizontal total gradient using the Canny algorithm; and generating a regional magnetic anomaly plan map based on the extracted gradient mutation band.

[0010] Optionally, in a third implementation of the first aspect of the present invention, the method of confirming the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, obtaining the magnetic profile data of the mineralization potential area and processing it to obtain a three-dimensional probability model of the magnetic body includes: superimposing the fault speculation map and the regional anomaly plan map to confirm the mineralization potential area; obtaining 1:10,000 magnetic profile data in the confirmed mineralization potential area; using the sym8 wavelet to perform a two-layer decomposition of the magnetic profile data, and retaining 50-500nT of effective magnetic anomaly data through threshold denoising; inputting the effective magnetic anomaly data and gravity high value area information into the pre-trained UNet++ model to obtain regional probability information; and generating a three-dimensional probability model of the magnetic body based on the regional probability information.

[0011] Optionally, in a fourth implementation method of the first aspect of the present invention, the remote sensing image of the area to be predicted is obtained and processed to generate a hydroxyl anomaly distribution vector diagram, including: obtaining a remote sensing image of the area to be predicted, the remote sensing image being a Landsat 9 OLI image; performing radiation calibration processing, atmospheric correction processing, and geometric correction processing on the remote sensing image to obtain a preprocessed remote sensing image; calculating the spectral index of the preprocessed remote sensing image, the calculated spectral index including a hydroxyl index, an iron staining index, and a normalized vegetation index; superimposing the original RGB band of the preprocessed remote sensing image with the calculated spectral index to construct a 7-band data set; and inputting the constructed 7-band data set into a pre-trained U-Net classification model to obtain a hydroxyl anomaly distribution vector diagram.

[0012] Optionally, in a fifth implementation of the first aspect of the present invention, the geological data of the area to be predicted is obtained, and a mineralization probability target area map is generated based on the geological data, the fracture speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body, and the hydroxyl anomaly distribution vector map, including: obtaining geological data of the area to be predicted, the geological data being a top surface depth model of the Ordovician limestone; converting the geological data fracture speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body, and the hydroxyl anomaly distribution vector map into raster data in a unified coordinate system of ArcGIS Pro; extracting 12 indicator features reflecting the mineralization conditions from the raster data to construct a multi-source feature data set; inputting the multi-source feature data set into a pre-trained XGBoost classifier to obtain the mineralization probability; and constructing a mineralization probability target area map based on the mineralization probability.

[0013] Optionally, in a sixth implementation of the first aspect of the present invention, the method of constructing a mineralization probability target area map based on the mineralization probability further includes: obtaining drilling verification results, the drilling verification results including well logging data of the drilling location and ground geophysical data around the drilling location; constructing drilling constraints based on the drilling verification results, and based on the constructed drilling constraints, using the least squares inversion method to adjust the feature weights of the XGBoost classifier.

[0014] The second aspect of the present invention provides an information analysis device for rich iron ore in a coverage area, comprising: a first generation module for acquiring gravity data of the area to be predicted and processing it to obtain information on high-value gravity areas, and generating a fault speculation map based on the acquired information on high-value gravity areas; a second generation module for acquiring magnetic navigation ΔT data of the area to be predicted and processing it to generate a regional magnetic anomaly plan map; a first processing module for acquiring magnetic profile data of the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, processing the magnetic profile data of the mineralization potential area to obtain a three-dimensional probability model of the magnetic body; a second processing module for acquiring remote sensing images of the area to be predicted and processing them to generate a hydroxyl anomaly distribution vector map; a third generation module for acquiring geological data of the area to be predicted, and generating a mineralization probability target area map based on the geological data, the fault speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body, and the hydroxyl anomaly distribution vector map.

[0015] The third aspect of the present invention provides a coverage area iron-rich ore information analysis device, which includes: a memory and at least one processor, wherein the memory stores instructions; at least one processor calls the instructions in the memory to enable the coverage area iron-rich ore information analysis device to execute each step of the coverage area iron-rich ore information analysis method described above.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the various steps of any of the above-mentioned methods for analyzing information of rich iron ore in a coverage area.

[0017] The technical solution of the present invention integrates multi-dimensional information such as gravity data, magnetic flight ΔT data, remote sensing images and geological data of the area to be predicted, constructs a characteristic system that can comprehensively reflect the mineralization conditions, breaks through the limitations of traditional single data source analysis, and provides a rich and accurate data basis for the intelligent division of mineralization target areas, thereby significantly improving the accuracy of mineralization prediction; further, the method uses fault speculation maps, regional magnetic anomaly plane maps, three-dimensional probability models of magnetic bodies and hydroxyl anomaly distribution vector maps for collaborative analysis, which not only provides a rich and accurate data basis for the intelligent division of mineralization target areas, further improves the accuracy of mineralization prediction, but also significantly enhances the systematic nature of hidden ore body identification; compared with traditional methods that rely on manual experience, this method effectively improves the objectivity of mineralization prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A logic flow chart of a method for analyzing rich iron ore information in a coverage area provided by an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a device for analyzing information on rich iron ore in a coverage area provided by an embodiment of the present invention; Figure 3 A schematic diagram of the structure of the coverage area rich iron ore information analysis device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention provides a method, device, equipment and storage medium for analyzing information of rich iron ore in a coverage area. In the present invention, the terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or are inherent to these processes, methods, products or apparatus.

[0020] This application discloses a method for analyzing information of rich iron ore in a coverage area. For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 An embodiment of the method for analyzing information of rich iron ore in a coverage area according to the present invention includes: 101. Obtaining and processing gravity data of the area to be predicted to obtain information on high-value gravity areas, and generating a fault prediction map based on the obtained information on high-value gravity areas; In this embodiment, underground density differences in the area to be predicted are analyzed using gravity data, which is basic data for identifying geological bodies such as concealed rock masses and fault zones.

[0021] 102. Obtain and process the magnetic airborne ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map; In this embodiment, magnetic navigation data is used to reflect the distribution of underground magnetic bodies in the area to be predicted.

[0022] 103. Based on the fault inference map and the regional magnetic anomaly plan map, identify the mineralization potential area, obtain and process the magnetic profile data of the mineralization potential area, and obtain a three-dimensional probability model of the magnetic body; In this embodiment, the fault speculation map and the regional magnetic anomaly plan map are superimposed, and the structural-magnetic anomaly superposition area is screened as the mineralization potential area. Based on the magnetic profile data of the mineralization potential area, a three-dimensional probability distribution is output to achieve quantitative prediction of the spatial position of the ore body.

[0023] 104. Obtain and process remote sensing images of the area to be predicted to generate a hydroxyl anomaly distribution vector map; In this embodiment, the remote sensing image contains visible light-shortwave infrared bands, which are sensitive to the spectral characteristics of altered minerals; the generated hydroxyl anomaly distribution vector diagram reflects the relationship between surface alteration and mineralization hydrothermal activity.

[0024] 105. Obtain geological data of the area to be predicted, and generate a mineralization probability target area map based on the geological data, fault speculation map, regional magnetic anomaly plan map, magnetic body three-dimensional probability model and hydroxyl anomaly distribution vector map.

[0025] The present application discloses a method for analyzing information on rich iron ore in covered areas, which integrates multi-dimensional information such as gravity data, magnetic flight ΔT data, remote sensing images and geological data of the area to be predicted, and constructs a feature system that can comprehensively reflect the mineralization conditions. It breaks through the limitations of traditional single data source analysis, provides a rich and accurate data basis for the intelligent division of mineralization target areas, and thus significantly improves the accuracy of mineralization prediction; further, the method uses fault speculation maps, regional magnetic anomaly plane maps, three-dimensional probability models of magnetic bodies and hydroxyl anomaly distribution vector maps for collaborative analysis, which not only provides a rich and accurate data basis for the intelligent division of mineralization target areas, further improves the accuracy of mineralization prediction, but also significantly enhances the systematic nature of hidden ore body identification; compared with traditional methods that rely on manual experience, this method effectively improves the objectivity of mineralization prediction.

[0026] Furthermore, in an embodiment of the present invention, the step of acquiring and processing gravity data of the area to be predicted to obtain information on high-value gravity areas, and generating a fault prediction map based on the acquired information on high-value gravity areas, includes: 201. Obtain 1:50,000 gravity data of the area to be predicted after terrain interference processing; In this embodiment, 1:50,000 gravity data of the area to be predicted is obtained after only terrain interference processing. Topographic undulations, such as mountains and depressions, will interfere with the gravity measurement results, such as low gravity values ​​at high places and high gravity values ​​at low places. The gravity data is terrain corrected using the Bouguer correction method. The influence of terrain undulations on the gravity values ​​is calculated through a formula and eliminated to ensure that the data only reflects differences in underground material density, such as the difference between high-density rock mass and low-density surrounding rock. 202. Using db5 wavelet to perform a three-layer decomposition of gravity data to separate the regional field with wavelength > 10 km and the local field with wavelength in the range of 2-5 km; In this embodiment, the db5 wavelet is used to perform a three-layer decomposition of the gravity data: the first layer extracts the regional field with a wavelength greater than 10 km, revealing the undulating characteristics of the deep crustal basement; the second to third layers extract the local field with a wavelength of 2-5 km, reflecting the density anomalies of shallow rock and ore bodies; the wavelet decomposition realizes the layered identification of geological bodies at different depths through multi-scale analysis, effectively solving the problem of traditional methods in separating deep and shallow information.

[0027] 203. Obtain a regional geological map, verify the separated local field based on the regional geological map, and obtain information on high gravity value areas; In this embodiment, the local field is verified in combination with a regional geological map containing known rock outcrops and stratigraphic distribution information; if the high-value area of ​​the local field spatially overlaps with the known diorite outcrop, or is located in a tectonic intersection zone, it is determined to be a high-gravity area (potentially concealed rock body), thereby providing density constraints for subsequent mineralization area screening and improving the accuracy of identifying high-gravity areas. 204. Calculate the total horizontal gradient of the local field, and extract the gradient mutation band of the local field using the Canny algorithm based on the calculated total horizontal gradient; In this embodiment, the horizontal total gradient of the local field is calculated by derivatives, and the horizontal total gradient can accurately reflect the rate of change of the gravity value in space; then, the Canny algorithm is applied, with the help of double threshold detection and edge connection, to effectively extract the gradient mutation zone, which corresponds to the fracture boundary with a significant difference in density; finally, a fracture speculation map is generated, which details the strike and length of the fracture, providing a quantitative basis of great value for mineralization structure analysis.

[0028] 205. Generate a fracture speculation map based on the extracted gradient mutation band.

[0029] Furthermore, in an embodiment of the present invention, obtaining and processing the magnetic airborne ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map includes: 301. Obtain 1:50,000 magnetic navigation ΔT data for the area to be predicted, where ΔT represents the total magnetic field intensity anomaly. In this embodiment, ΔT represents the total magnetic field intensity anomaly, which is defined as the difference between the measured total magnetic field intensity and the theoretical normal magnetic field, and is measured in nanoteslas (nT). The magnetic aeromagnetic ΔT data obtained by airborne magnetic surveys has a wide coverage range and can reflect the distribution characteristics of underground magnetic bodies, such as magnetite and magnetic rock masses. The data scale is 1:50,000, which reflects the matching relationship between the spatial resolution of the data and the coverage range. 302. Perform polarization processing and upward extension 5 km processing on the magnetic airborne ΔT data to obtain processed magnetic anomaly data; In this embodiment, the magnetic anomaly is converted to the magnetic pole coordinate system through coordinate transformation, so as to eliminate the magnetic anomaly offset phenomenon caused by the geomagnetic inclination; for example, in the northern hemisphere, the magnetic anomaly usually offsets to the north. After the polarization processing, the anomaly peak can correspond to the plane position of the magnetic body; further, the magnetic anomaly is extended upward by 5 km using Fourier transform to suppress shallow magnetic interference; shallow magnetic factors such as ferromagnetic minerals in the soil will interfere with the magnetic anomaly, and their influence can be effectively suppressed by the upward extension processing, thereby highlighting the overall abnormal characteristics of the magnetic body in the deep area, such as the abnormal characteristics of the intrusive rock mass; by combining the polarization processing and the upward extension processing, the interpretability of the magnetic anomaly is significantly improved.

[0030] 303. Calculate the total horizontal gradient of the processed magnetic anomaly data, and extract the gradient mutation zone using the Canny algorithm based on the calculated total horizontal gradient; In this embodiment, the horizontal total gradient of the processed magnetic anomaly data is calculated, and the Canny algorithm is used to extract the gradient mutation zone, which is the boundary feature of the magnetic body. Subsequently, the boundary feature is superimposed on the regional geological map to generate a regional magnetic anomaly plan map. In the regional magnetic anomaly plan map, the intersection area of ​​the high-value area and the mutation zone is the potential magnetic ore body distribution area. By generating the regional magnetic anomaly plan map, key magnetic constraint conditions are provided for the preliminary delineation of the mineralization potential area. 304. Generate a regional magnetic anomaly plan map based on the extracted gradient mutation band.

[0031] Furthermore, in an embodiment of the present invention, the method of confirming the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, obtaining the magnetic profile data of the mineralization potential area and processing it to obtain the three-dimensional probability model of the magnetic body includes: 401. Overlay the fault speculation map and the regional anomaly plane map to confirm the mineralization potential area; In this embodiment, the fault speculation map and the regional magnetic anomaly plane map are superimposed on the ArcGIS platform (the unified coordinate system is set to CGCS2000), and the overlapping areas of the fault-intensive belt and the magnetic anomaly high-value area are screened to determine them as mineralization potential areas; the mineralization potential areas have both structural channels (faults) and magnetic materials (ore bodies / rock bodies), and the mineralization conditions are relatively favorable.

[0032] 402. Obtain 1:1 magnetic profile data within the confirmed mineralization potential area; In this embodiment, within the mineralization potential area, a survey line is arranged perpendicular to the fault strike direction, and magnetic profile data with a scale of 1:10,000 is obtained with the help of ground magnetic survey instruments. The measuring point spacing is 50 meters to ensure that local anomalies can be finely characterized; the obtained magnetic profile data is used to perform a fine inversion of the three-dimensional spatial distribution of magnetic bodies (ore bodies / rock bodies). 403. Use sym8 wavelet to perform two-layer decomposition on magnetic profile data, and retain the effective magnetic anomaly data of 50-500nT through threshold noise reduction; In this embodiment, the sym8 wavelet is used to perform a two-layer decomposition on the data, and the anomalies of 50-500nT are retained by means of threshold noise reduction, and the instrument noise and surface interference are removed, so as to obtain effective magnetic anomaly data. 404. Input the valid magnetic anomaly data and the high gravity value area information into the pre-trained UNet++ model to obtain regional probability information; 405. Generate a three-dimensional probability model of the magnetic body based on the regional probability information; In this embodiment, the UNet++ model selects 100 sets of magnetic anomaly profile data of known ore bodies as input, and uses the ore body distribution verified by drilling as output to form training samples; iterative training is carried out with the help of the Adam optimizer until the model error is reduced to below 5%, completing the pre-training of the UNet++ model; then, the valid magnetic anomaly data and gravity high-value area information (as prior constraints) are input into the trained UNet++ model, and the model outputs the probability of the existence of the magnetic body at each spatial point, and the probability value is between 0 and 1; then, an interpolation algorithm is used to construct a three-dimensional probability model of the magnetic body, and the X / Y / Z resolution of the model is 100m×100m×50m; by combining high-resolution magnetic profile data with a deep learning model, a three-dimensional quantitative prediction of the spatial distribution of the magnetic body is achieved, which effectively solves the problem that traditional two-dimensional inversion is difficult to reflect the vertical changes of the ore body, and provides an accurate spatial coordinate reference for subsequent drilling design.

[0033] Furthermore, in an embodiment of the present invention, the step of obtaining and processing a remote sensing image of the area to be predicted to generate a hydroxyl anomaly distribution vector map includes: 501. Acquire a remote sensing image of the area to be predicted, where the remote sensing image is a Landsat 9 OLI image; In this example, Landsat 9 OLI imagery was selected. Landsat 9 is an Earth observation satellite whose onboard Land Imager (OLI) can acquire remote sensing images from visible light to shortwave infrared bands, covering eight multispectral bands and one panchromatic band. Landsat 9 OLI imagery is highly sensitive to the absorption characteristics of hydroxyl minerals (such as kaolinite and montmorillonite) and is therefore suitable for identifying altered areas.

[0034] 502. Performing radiometric calibration processing, atmospheric correction processing, and geometric correction processing on the remote sensing image to obtain a preprocessed remote sensing image; In this embodiment, radiometric calibration converts the digital quantization value (DN value) of the image into surface reflectance or radiant brightness value, eliminating the influence of the sensor's own response differences and making the data physically meaningful. Atmospheric correction uses the FLAASH model to eliminate interference with electromagnetic wave transmission caused by atmospheric scattering, aerosol absorption, etc., and restore the true reflective properties of the surface, such as the spectral characteristics of altered minerals. Geometric correction: Based on the CGCS2000 coordinate system (38° zoning), the image is geometrically precisely corrected using ground control points to ensure that the spatial registration error of different bands is less than 0.5 pixels, thereby ensuring the accuracy of subsequent multi-band data fusion.

[0035] 503. Calculate the spectral index of the preprocessed remote sensing image, the calculated spectral index includes hydroxyl index, iron staining index and normalized vegetation index; In this embodiment, the hydroxyl index (NDSI) is used to identify hydroxyl alteration. Specifically, NDSI = (Green band - SWIR1 band) / (Green + SWIR1), with a threshold value greater than 0.2. The iron staining index (Fe) is used to identify iron oxides. Specifically, Fe = (Red band - Blue band) / (Red + Blue), with a threshold value greater than 0.1. The normalized difference vegetation index (NDVI) is used to distinguish vegetation cover interference. Specifically, NDVI = (NIR - Red) / (NIR + Red).

[0036] 504. Superimposing the original RGB bands of the pre-processed remote sensing image with the calculated spectral index to construct a 7-band dataset; In this embodiment, the original RGB bands (Bands2, Bands3, Bands4) are superimposed with three indices to construct a 7-band dataset containing spectral and alteration information.

[0037] 505. Input the constructed 7-band dataset into the pre-trained U-Net classification model to obtain a hydroxyl abnormal distribution vector diagram; In this embodiment, remote sensing image slices of known alteration areas are used as training samples to pre-train a U-Net classification model, and the trained U-Net classification model is used to classify the 7-band data set; based on the input 7-band data set, the U-Net classification model outputs a hydroxyl anomaly distribution vector diagram, wherein areas with a probability greater than 0.5 are determined to be alteration anomaly areas; through multi-band fusion technology and deep learning classification methods, automatic identification of hydroxyl alteration areas is achieved; compared with traditional visual interpretation methods, it not only significantly improves the recognition efficiency, but also has the ability to capture weak alteration signals, providing a reliable and effective basis for the precise delineation of mineralization hydrothermal activity areas.

[0038] Furthermore, in an embodiment of the present invention, the geological data of the area to be predicted is obtained, and a mineralization probability target area map is generated based on the geological data, the fracture speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body, and the hydroxyl anomaly distribution vector map, including: 601. Obtain geological data of the area to be predicted, wherein the geological data is a top surface depth model of the Ordovician limestone; In this embodiment, a top surface depth model of the Ordovician limestone in the area to be predicted is collected. The top surface depth model of the Ordovician limestone is generated by interpolation of drilling data and can reflect the spatial distribution of the mineralized surrounding rocks. The top surface depth model of the Ordovician limestone provides geological constraints for the ore body occurrence layer, because most iron-rich ores are related to the limestone contact zone.

[0039] 602. Convert the geological data fracture inference map, regional magnetic anomaly plan map, magnetic body 3D probability model, and hydroxyl anomaly distribution vector map into raster data in the unified coordinate system of ArcGIS Pro. In this embodiment, in ArcGIS Pro, geological data, fault speculation maps, regional magnetic anomaly plane maps, three-dimensional probability models of magnetic bodies, and hydroxyl anomaly distribution vector maps are uniformly converted into raster data (resolution of 50m, coordinate system CGCS2000 38° zoning) to ensure data spatial matching.

[0040] 603. Extract 12 indicator features reflecting metallogenic conditions from raster data to construct a multi-source feature dataset; In this embodiment, 12 indicator features reflecting the mineralization conditions are extracted from the raster data, including: structural density (fracture length / unit area), magnetic probability value, gravity gradient value, hydroxyl anomaly probability, distance from the fracture, resistivity value, iron staining index, rock mass burial depth, magnetic anomaly intensity, fracture strike consistency, annular structural density, and depth of the top surface of the Ordovician limestone; these indicator features cover multi-dimensional information of structure, geophysical exploration, remote sensing, and geology.

[0041] 604. Input the multi-source feature dataset into the pre-trained XGBoost classifier to obtain the mineralization probability; In this embodiment, the XGBoost classifier is used to realize mineralization prediction. First, the known ore body area is used as the positive sample and the non-ore area is used as the negative sample. The n_estimators parameter is set to 800 and the max_depth parameter is set to 5. The model parameters are optimized through 5-fold cross validation to complete the model training.

[0042] In the pre-trained XGBoost classifier, each feature has a different weight, which reflects its importance in mineralization prediction. For example, the weight of the magnetic probability value is 0.35, which can directly reflect the magnetic anomaly of the ore body. Because the magnetic characteristics of the ore body often lead to abnormal magnetic measurement results, the magnetic probability value plays an important role in judging whether the ore body exists or not. The weight of the gravity gradient value is 0.25, which can indicate the location of the rock contact zone. The rock contact zone is usually a favorable location for mineralization because the material exchange and chemical reaction between different rock bodies may promote the formation of ore. The weight of the base anomaly probability is 0.2, which reflects the relationship between surface alteration and deep mineralization. Surface alteration is the manifestation of deep mineralization on the surface. By analyzing the hydroxyl anomaly probability, we can infer whether there is mineralization at depth; the weight of the distance from the fault is 0.15. The fault structure provides a favorable structural space for mineralization. Ore-forming fluids may migrate and precipitate along the fault zone, so the distance from the fault has a certain influence on mineralization; the weight of the resistivity value is 0.05. Low-resistance areas often indicate fluid migration channels. The flow of ore-forming fluids will change the resistivity of the rock. Therefore, the resistivity value can be used as an indicator to judge the fluid migration situation.

[0043] The multi-source feature dataset is input into the pre-trained XGBoost classifier, which outputs the metallogenic probability, which is between 0 and 1. The target area to be predicted is divided according to the metallogenic probability, and a metallogenic probability target area map is generated. The specific division criteria are as follows: Class A target area: The mineralization probability is greater than 0.8, the magnetic probability is greater than 0.7, and the hydroxyl probability is greater than 0.6. This type of target area has a higher mineralization possibility and requires priority drilling to verify the existence of ore bodies; Class B target area: The probability of mineralization is between 0.6 and 0.8. For this type of target area, additional verification work such as soil magnetism or induced polarization sounding is needed to further determine its mineralization potential.

[0044] 605. Constructing a mineralization probability target area map based on the mineralization probability; In this embodiment, the application of multi-source feature fusion technology effectively overcomes the significant limitations of using single data for prediction. In traditional prediction scenarios, relying solely on single data for analysis often leads to large deviations and lack of comprehensiveness in prediction results due to factors such as missing data dimensions and incomplete information. Multi-source feature fusion integrates data from different channels and different types, fully exploring the potential correlations and complementary information between various data sources, thereby constructing a richer and more accurate data set, providing a solid foundation for subsequent analysis and prediction.

[0045] As an advanced machine learning algorithm, the XGBoost model can conduct in-depth analysis of data after the fusion of multi-source features by automatically learning feature weights. During the learning process, the XGBoost model dynamically assigns corresponding weights based on the influence of each feature on the target variable (i.e., the probability of mineralization). This automatic learning mechanism enables the model to accurately capture key information related to mineralization in the data, thereby achieving accurate quantitative calculation of the probability of mineralization.

[0046] Compared with traditional target area delineation methods, target area delineation based on the quantitative calculation results of multi-source feature fusion and XGBoost model can identify potential mineralization areas more scientifically and accurately, greatly improving the accuracy of target area delineation.

[0047] Furthermore, in an embodiment of the present invention, the method of constructing a mineralization probability target area map based on the mineralization probability further includes: 701. Obtain drilling verification results, where the drilling verification results include well logging data at the drilling location and ground geophysical data around the drilling location; In this embodiment, logging data at the drilling location are collected, such as magnetic logging curves to identify the depth of magnetic ore bodies and gamma logging to distinguish alteration types, as well as ground geophysical data, that is, gravity and magnetic re-survey data within 500m around the drilling hole, to form a well-ground joint verification data set, that is, the drilling verification results.

[0048] 702. Constructing drilling constraints based on the drilling verification results, and adjusting the feature weights of the XGBoost classifier using a least squares inversion method based on the constructed drilling constraints; In this embodiment, constraints are constructed based on the drilling results to provide hard constraints for model optimization. For example, the constraints include: the top depth of the magnetic body is set to ±100m of the top depth of the ore body interpreted by logging, and the resistivity constraint uses the apparent resistivity value of the formation revealed by drilling.

[0049] The least squares inversion method is used to adjust the feature weights of the XGBoost classifier with the drilling verification results as the true value. For example, if the gravity value in a certain area is high but no minerals are found during actual drilling, the gravity feature weight is reduced by 5%-10%, that is, the weight of the gravity gradient value is adjusted to 0.15-0.2, so that the prediction results of the XGBoost classifier are more consistent with the actual drilling situation.

[0050] The XGBoost classifier is optimized using actual drilling data as feedback, forming a closed loop of prediction, verification, and iteration. This effectively solves the possible deviation problem in the initial training data of the model, enables the subsequent target area prediction accuracy to be continuously improved, and provides technical support with dynamic optimization capabilities for mineral exploration.

[0051] The above describes the method for analyzing the information of rich iron ore in the coverage area according to the embodiment of the present invention. The following describes the device for analyzing the information of rich iron ore in the coverage area according to the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an apparatus for analyzing information of rich iron ore in a coverage area includes: The first generation module 801 is used to obtain gravity data of the area to be predicted and process it to obtain information of high-value gravity areas, and generate a fault prediction map based on the obtained high-value gravity area information; The second generating module 802 is used to obtain and process the magnetic aeronautical ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map; The first processing module 803 is used to obtain magnetic profile data of the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, and process the magnetic profile data of the mineralization potential area to obtain a three-dimensional probability model of the magnetic body; The second processing module 804 is used to obtain and process the remote sensing image of the area to be predicted to generate a hydroxyl anomaly distribution vector map; The third generation module 805 is used to obtain geological data of the area to be predicted, and generate a mineralization probability target area map based on the geological data, fault speculation map, regional magnetic anomaly plan map, magnetic body three-dimensional probability model and hydroxyl anomaly distribution vector map.

[0052] Based on the same idea as the method in the above embodiment, the device provided in this application can implement the method in the above embodiment.

[0053] above Figure 2 The coverage area rich iron ore information analysis device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The coverage area rich iron ore information analysis device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0054] Figure 3This is a schematic diagram of the structure of a device for analyzing coverage area iron-rich ore information provided by an embodiment of the present invention. This device 900 may vary significantly depending on configuration or performance. It may include one or more processors (central processing units, CPUs) 910, a memory 920, and one or more storage media 930 (e.g., one or more mass storage devices) storing applications 933 or data 932. The memory 920 and storage medium 930 may be either transient or persistent storage. The program stored in the storage medium 930 may include one or more modules (not shown), each of which may include a series of instructions operating on the device 900. Furthermore, the processor 910 may be configured to communicate with the storage medium 930, executing the series of instructions stored in the storage medium 930 on the device 900 to implement the steps of the method for analyzing coverage area iron-rich ore information provided in the aforementioned method embodiments.

[0055] The coverage area rich iron ore information analysis device 900 may further include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input and output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the coverage area iron-rich ore information analysis device shown does not constitute a limitation to the coverage area iron-rich ore information analysis device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0056] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the method for analyzing information of rich iron ore in a coverage area.

[0057] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0058] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0059] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for analyzing information of rich iron ore in a coverage area, characterized in that: include: Obtaining and processing gravity data of the area to be predicted to obtain information on high-value gravity areas, and generating a fault prediction map based on the obtained high-value gravity area information; Obtain and process the magnetic airborne ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map; Based on the fault speculation map and regional magnetic anomaly plan map, the mineralization potential area is confirmed, and the magnetic profile data of the mineralization potential area is obtained and processed to obtain a three-dimensional probability model of the magnetic body; Obtain remote sensing images of the area to be predicted and process them to generate a hydroxyl anomaly distribution vector map; Obtain geological data of the area to be predicted, and generate a mineralization probability target area map based on geological data, fault speculation map, regional magnetic anomaly plan map, magnetic body three-dimensional probability model and hydroxyl anomaly distribution vector map.

2. The method for analyzing information of rich iron ore in coverage area according to claim 1, characterized in that: The method of obtaining gravity data of the area to be predicted and processing the data to obtain high-value gravity area information and generating a fault prediction map based on the obtained high-value gravity area information includes: Obtain 1:50,000 gravity data of the area to be predicted after terrain interference processing; The db5 wavelet was used to perform a three-layer decomposition of the gravity data to separate the regional field with a wavelength greater than 10 km and the local field with a wavelength in the range of 2-5 km. Obtain a regional geological map, verify the separated local field based on the regional geological map, and obtain information on high gravity value areas; The horizontal total gradient of the local field is calculated, and based on the calculated horizontal total gradient, the Canny algorithm is used to extract the gradient mutation band of the local field; A break prediction map is generated based on the extracted gradient mutation bands.

3. The method for analyzing information of rich iron ore in coverage area according to claim 1, characterized in that: The method of obtaining and processing the magnetic airborne ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map includes: Obtain 1:50,000 magnetic navigation ΔT data for the area to be predicted, where ΔT represents the total magnetic field intensity anomaly. The magnetic anomaly data were processed by polarization and 5km extension respectively. The horizontal total gradient of the processed magnetic anomaly data is calculated, and the gradient mutation zone is extracted using the Canny algorithm based on the calculated horizontal total gradient; A regional magnetic anomaly plan map is generated based on the extracted gradient mutation bands.

4. The method for analyzing information of rich iron ore in coverage area according to claim 1, characterized in that: The method of confirming the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, obtaining the magnetic profile data of the mineralization potential area and processing it to obtain the three-dimensional probability model of the magnetic body includes: Overlay the fault speculation map and regional anomaly plane map to confirm the mineralization potential area; Obtain 1:1 magnetic profile data within the confirmed mineralization potential area; The sym8 wavelet was used to perform a two-layer decomposition of the magnetic profile data, and the effective magnetic anomaly data of 50-500nT was retained through threshold noise reduction. Input the effective magnetic anomaly data and gravity high value area information into the pre-trained UNet++ model to obtain regional probability information; Generate a three-dimensional probability model of the magnetic body based on the regional probability information.

5. The method for analyzing information of rich iron ore in coverage area according to claim 1, characterized in that: The step of obtaining and processing a remote sensing image of the area to be predicted to generate a hydroxyl anomaly distribution vector map includes: Obtaining a remote sensing image of the area to be predicted, wherein the remote sensing image is a Landsat 9 OLI image; The remote sensing image is subjected to radiometric calibration, atmospheric correction and geometric correction respectively to obtain a pre-processed remote sensing image; Calculate the spectral index of the preprocessed remote sensing image, including the hydroxyl index, iron staining index and normalized vegetation index; The original RGB bands of the preprocessed remote sensing image are superimposed with the calculated spectral index to construct a 7-band dataset; The constructed 7-band dataset was input into the pre-trained U-Net classification model to obtain the hydroxyl abnormal distribution vector diagram.

6. The method for analyzing information of rich iron ore in coverage area according to claim 1, characterized in that: The method of obtaining geological data of the area to be predicted and generating a mineralization probability target area map based on the geological data, the fracture speculation map, the regional magnetic anomaly plan map, the three-dimensional probability model of the magnetic body and the hydroxyl anomaly distribution vector map includes: Acquiring geological data of the area to be predicted, wherein the geological data is a top surface depth model of the Ordovician limestone; In the unified coordinate system of ArcGIS Pro, the geological data fracture speculation map, regional magnetic anomaly plan map, magnetic body three-dimensional probability model and hydroxyl anomaly distribution vector map were converted into raster data; Twelve indicator features reflecting metallogenic conditions were extracted from raster data to construct a multi-source feature dataset; The multi-source feature dataset is input into the pre-trained XGBoost classifier to obtain the mineralization probability; A mineralization probability target area map is constructed based on the mineralization probability.

7. The method for analyzing information of rich iron ore in coverage area according to claim 6, characterized in that: The method further comprises: constructing a mineralization probability target area map based on the mineralization probability; and then: Obtaining drilling verification results, the drilling verification results including well logging data at the drilling location and ground geophysical data around the drilling location; Drilling constraints are constructed based on the drilling verification results, and based on the constructed drilling constraints, the least squares inversion method is used to adjust the feature weights of the XGBoost classifier.

8. A device for analyzing information of rich iron ore in a coverage area, characterized in that: include: The first generation module is used to obtain gravity data of the area to be predicted and process it to obtain information of high-value gravity areas, and generate a fault speculation map based on the obtained high-value gravity area information; The second generation module is used to obtain and process the magnetic aerobatic ΔT data of the area to be predicted to generate a regional magnetic anomaly plan map; The first processing module is used to obtain magnetic profile data of the mineralization potential area based on the fault speculation map and the regional magnetic anomaly plan map, and process the magnetic profile data of the mineralization potential area to obtain a three-dimensional probability model of the magnetic body; The second processing module is used to obtain and process the remote sensing image of the area to be predicted to generate a hydroxyl anomaly distribution vector map; The third generation module is used to obtain geological data of the area to be predicted, and generate a mineralization probability target area map based on geological data, fault speculation map, regional magnetic anomaly plan map, magnetic body three-dimensional probability model and hydroxyl anomaly distribution vector map.

9. A coverage area rich iron ore information analysis device, characterized in that: The coverage area rich iron ore information analysis device includes: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors calls the instructions in the memory to enable the coverage area iron-rich ore information analysis device to perform each step of the coverage area iron-rich ore information analysis method according to any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the method for analyzing the information of rich iron ore in the coverage area as claimed in any one of claims 1 to 7 are implemented.

Images