Multi-source data collaborative snake-green hybrid rock air-space-ground integrated interpretation method and system

Abstract

The embodiment of the invention provides a multi-source data collaborative snake-green hybrid rock air-space-ground integrated interpretation method and system, and relates to the technical field of geological survey, and the method comprises the steps: obtaining a satellite and unmanned plane remote sensing image of a target region, automatically identifying satellite and unmanned aerial vehicle remote sensing images through a convolutional neural network to extract a surface lithology initial boundary and a construction clue; obtaining a snake-green hybrid rock surface outcrop range; collecting gravity data, magnetic method data and electrical method data; performing underground physical property structure deduction on the gravity data, the magnetic method data and the electrical method data through a conjugate gradient inversion algorithm, and constructing an underground physical property structure model; and fusing the surface outcrop range of the snake-green hybrid rock with the underground geophysical prospecting body structure model based on an implicit modeling technology to obtain a geology-geophysical space structure interpretation map. The technical problems of long time consumption and low interpretation accuracy of snake-green hybrid rock structure analysis in the prior art are solved. And the effects of improving the credibility of deep structure inference and providing a reliable geological structure are achieved.

Description

Technical Field

[0001] This invention relates to the field of geological survey technology, and in particular to a method and system for integrated air-space-ground interpretation of ophiolite mélange based on multi-source data collaboration. Background Technology

[0002] Current investigations of ophiolite mélange mainly rely on manual surface geological work, which has significant limitations: on the one hand, the efficiency of the investigation is limited by the intensity of fieldwork, requiring personnel to trek through complex terrain to observe and record point by point, making it difficult to quickly cover large areas, especially in high mountain and canyon areas where access is difficult; on the other hand, traditional surface mapping cannot obtain information on the underground extension of the rock mass. For ophiolite mélange blocks that are buried by overburden or deformed deep by tectonic disturbance, it is difficult to accurately determine their spatial morphology, scale, and contact relationship with the surrounding rock based solely on surface outcrops, resulting in blind spots in the understanding of underground structures and seriously affecting the understanding of the overall structural framework of the mélange zone.

[0003] While existing geophysical exploration methods can obtain information on subsurface physical properties, the lack of reliable surface lithology and tectonic constraints leads to multiple interpretations in the inversion results, making it difficult to accurately characterize the spatial configuration relationship between ophiolite mélange blocks and the matrix. This results in significant uncertainties in the understanding of deep structures.

[0004] It should be noted that the information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a multi-source data collaborative, integrated air-space-ground interpretation method and system for ophiolite mélange. This solves the technical problems of time-consuming analysis and low interpretation accuracy in existing technologies for ophiolite mélange structure analysis, achieving the technical effect of improving the reliability of deep structural inferences and providing reliable geological structures. The specific technical solution is as follows:

[0006] According to a first aspect of the present invention, a method for integrated air-space-ground interpretation of ophiolite mélange based on multi-source data collaboration is provided, the method comprising:

[0007] Satellite and UAV remote sensing images of the target area are acquired. A convolutional neural network is used to automatically identify and extract the initial lithological boundaries and structural clues of the surface from these images. A spectral angle mapping algorithm is used to enhance the identification of the characteristic spectra of ophiolite in the satellite and UAV remote sensing images. The characteristic spectral identification results are fused with the initial lithological boundaries to obtain the outcrop range of the ophiolite. Within the outcrop range, gravity, magnetic, and electrical exploration lines and points are deployed to collect gravity, magnetic, and electrical data. Under geological constraints, a conjugate gradient inversion algorithm is used to extrapolate the subsurface physical structure from the gravity, magnetic, and electrical data to construct a subsurface physical structure model. The geological constraints include using the outcrop range of the ophiolite as a priori lithological boundary constraint and the structural clues as structural constraints. Based on implicit modeling technology, the outcrop range of the ophiolite is fused with the subsurface geophysical body structure model to obtain a geological-geophysical spatial structure interpretation map.

[0008] In one implementation, satellite and UAV remote sensing images of the target area are acquired. A convolutional neural network is used to automatically identify and extract initial boundaries of surface lithology and structural clues from the satellite and UAV remote sensing images. The following processing is also performed: acquiring initial satellite and UAV remote sensing images of the target area; performing image preprocessing to obtain satellite and UAV remote sensing images, wherein image preprocessing includes radiometric correction, geometric correction, and orthorectification; constructing a geological label sample library; using the geological label sample library to train the convolutional neural network to obtain a trained convolutional neural network; and loading the trained convolutional neural network to automatically identify and extract initial boundaries of surface lithology and structural clues from the satellite and UAV remote sensing images.

[0009] In one implementation, a pre-trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images, extract initial boundaries of surface lithology and tectonic clues, and the following processing is performed: the pre-trained convolutional neural network is loaded, satellite and UAV remote sensing images are input into the convolutional neural network, semantic segmentation is performed, and the initial boundaries of surface lithology and initial tectonic clues are automatically identified and output; post-processing is performed on the initial boundaries of surface lithology and initial tectonic clues to obtain the initial boundaries of surface lithology and tectonic clues; wherein, the post-processing includes removing small-scale noise patches, smoothing the initial boundaries of surface lithology, and performing connectivity restoration and skeleton extraction on the initial tectonic clues.

[0010] In one implementation, a spectral angle mapping algorithm is used to enhance the identification of characteristic spectra of ophiolite in satellite and UAV remote sensing images. The characteristic spectral identification results are fused with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite. The following processing is also performed: based on field measurement points and geological data, multiple remote sensing pixel samples representing ophiolite are acquired. The reflectance characteristics of multiple remote sensing pixel samples in multispectral bands are statistically analyzed to extract multiple ophiolite endmember spectra. These are then summarized to obtain an ophiolite characteristic spectral library. Satellite and UAV remote sensing images are used as input data. According to the algorithm, the spectral angle between the image pixel spectrum and the ophiolite endmember spectrum in the ophiolite characteristic spectral library is calculated pixel by pixel using a spectral angle filling algorithm to generate an ophiolite spectral similarity distribution map. The smaller the spectral angle, the higher the similarity between the pixel and the ophiolite characteristic spectrum. Based on a preset spectral angle threshold, the spectral similarity distribution map is thresholded, and pixels that meet the threshold condition are marked as potential ophiolite pixels to obtain the characteristic spectrum identification result. The characteristic spectrum identification result is fused with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite.

[0011] In one implementation, the ophiolite outcrop range is obtained by fusing the characteristic spectral identification results with the initial surface lithological boundary. The following processing is also performed: the characteristic spectral identification results are spatially superimposed with the initial surface lithological boundary; the characteristic spectral identification results are then filtered and corrected using the initial surface lithological boundary as a screening and correction range constraint to obtain a corrected characteristic spectral identification result; areas that simultaneously satisfy both the corrected characteristic spectral identification result and the initial surface lithological boundary are designated as high-confidence ophiolite outcrops; areas that only satisfy either the corrected characteristic spectral identification result or the initial surface lithological boundary are identified as low-confidence ophiolite outcrops, and these low-confidence ophiolite outcrops are screened through regional connectivity analysis to obtain the ophiolite outcrop range.

[0012] In one implementation, under geological constraints, a subsurface physical structure model is constructed by extrapolating gravity, magnetic, and electrical resistivity data using a conjugate gradient inversion algorithm. The geological constraints include using the outcrop range of the ophiolite as a priori lithological boundary constraint and the tectonic clues as structural constraints. The following processing is also performed: traversing the gravity, magnetic, and electrical resistivity data to obtain a standardized multi-source geophysical dataset; constructing a three-dimensional inversion grid based on the lithological boundary and structural constraints in the geological constraints; and iteratively solving the three-dimensional inversion network using the objective function in the conjugate gradient inversion algorithm based on the standardized multi-source geophysical dataset. When the convergence condition is met, the subsurface physical structure model is output.

[0013] In one implementation, the gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset. The following processing is also performed: drift correction, tidal correction, topographic correction, and anomaly separation are applied to the gravity data to obtain gravity anomaly data, which reflects changes in subsurface density; diurnal variation correction, normal field correction, and anomaly separation are applied to the magnetic data to obtain magnetic anomaly data, which reflects changes in subsurface magnetic susceptibility; noise suppression and preliminary inversion processing are applied to the electrical data to obtain electrical response data, which reflects changes in subsurface resistivity; and the coordinates of the gravity anomaly data, magnetic anomaly data, and electrical response data are unified to obtain a standardized multi-source geophysical dataset.

[0014] In one implementation, the surface outcrop range of ophiolite mélange is fused with the subsurface geophysical structure model based on implicit modeling technology to obtain a geological-geophysical spatial structure interpretation map. The following processing is also performed: spatial vector data of the ophiolite mélange surface outcrop range, three-dimensional volumetric data of the subsurface geophysical structure model, and topographic data are acquired; a three-dimensional implicit geological scalar field is constructed using a coordinate system; isosurfaces characterizing the interface between the ophiolite mélange and the surrounding rock are extracted from the three-dimensional implicit geological scalar field; and based on these isosurfaces, the three-dimensional space is cut and segmented to generate a three-dimensional geological structure model composed of ophiolite mélange block units and matrix units; the physical properties of the three-dimensional geological structure model are correlated and fused with those of the subsurface geophysical structure model to obtain the geological-geophysical spatial structure interpretation map.

[0015] In one embodiment, the method further performs the following processing: based on spatial vector data, the boundary and internal points of the ophiolite surface outcrop range are taken as first-type spatial control points and assigned a first preset field value; the three-dimensional volumetric data and topographic data of the underground geophysical body structure model are taken as spatial trend constraints; based on the radial basis function interpolation algorithm, combined with the first-type spatial control points, the first preset field value and the spatial trend constraints, a three-dimensional implicit geological scalar field is constructed.

[0016] According to a second aspect of the present invention, a multi-source data collaborative air-space-ground integrated interpretation system for ophiolite mélange is provided, the system comprising:

[0017] The boundary extraction module acquires satellite and UAV remote sensing images of the target area and automatically identifies and extracts the initial boundary of surface lithology and structural clues from these images using a convolutional neural network. The outcrop range acquisition module enhances the identification of characteristic spectra of ophiolite in satellite and UAV remote sensing images using a spectral angle mapping algorithm, and fuses the characteristic spectral identification results with the initial boundary of surface lithology to obtain the outcrop range of ophiolite. The data acquisition module deploys gravity, magnetic, and electrical exploration survey lines and survey lines within the outcrop range of ophiolite. The system includes: a point for collecting gravity, magnetic, and electrical resistivity data; a structural model construction module for using a conjugate gradient inversion algorithm to extrapolate the subsurface physical structure from the gravity, magnetic, and electrical resistivity data under geological constraints, and constructing a subsurface physical structure model; and an interpretation map acquisition module for fusing the subsurface outcrop range of the ophiolite with the subsurface geophysical body structure model based on implicit modeling technology to obtain a geological-geophysical spatial structure interpretation map.

[0018] Beneficial effects of the embodiments of the present invention:

[0019] In the solution provided by this invention, satellite and UAV remote sensing images of the target area are acquired. A convolutional neural network is used to automatically identify and extract the initial boundaries of surface lithology and structural clues from the satellite and UAV remote sensing images. Then, a spectral angle mapping algorithm is used to enhance the identification of the characteristic spectra of ophiolite in the satellite and UAV remote sensing images. The characteristic spectral identification results are fused with the initial boundaries of surface lithology to obtain the outcrop range of ophiolite. Within the outcrop range of ophiolite, gravity exploration, magnetic exploration, and electrical exploration are deployed. Gravity, magnetic, and electrical resistivity data were collected via survey lines and points. Then, under geological constraints, a conjugate gradient inversion algorithm was used to extrapolate the subsurface physical structure from these data, constructing a subsurface physical structure model. The geological constraints included using the outcrop extent of the ophiolite mélange as a priori lithological boundary constraint and tectonic clues as structural constraints. Implicit modeling techniques were used to fuse the outcrop extent of the ophiolite mélange with the subsurface geophysical body structure model, obtaining a geological-geophysical spatial structure interpretation map. This method achieves the goal of constructing a three-dimensional geological model in the ophiolite mélange zone, from detailed identification of surface outcrops to quantitative extrapolation of deep structures, through multi-source data collaborative processing. It clearly reveals the spatial configuration relationship between rock blocks and matrix, as well as the distribution of key tectonic interfaces, improving the reliability of deep structural inferences and providing reliable three-dimensional geological evidence for resource exploration and engineering geological evaluation. Of course, implementing any product or method of this invention does not necessarily require achieving all of the above-described advantages simultaneously. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This invention illustrates a schematic flowchart of the multi-source data collaborative interpretation method for ophiolite mélange using an integrated air-space-ground approach.

[0022] Figure 2 A schematic diagram of the structure of the multi-source data collaborative ophiolite integrated air-space-ground interpretation system provided by the present invention is shown.

[0023] Figure labeling: Boundary extraction module 11, Outcrop range acquisition module 12, Data acquisition module 13, Structural model construction module 14, Interpretation map acquisition module 15. Detailed Implementation

[0024] To facilitate understanding of the present invention, a more complete description of the invention will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein; rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the invention.

[0025] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0026] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0027] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0028] The present invention provides a multi-source data collaborative method and system for integrated air-space-ground interpretation of ophiolite mélange, which solves the technical problems of long time consumption and low interpretation accuracy in the existing technology for analyzing the structure of ophiolite mélange.

[0029] Example 1: See Figure 1 The flowchart of the multi-source data collaborative air-space-ground integrated interpretation method for ophiolite mélange provided in this embodiment of the invention includes:

[0030] A1: Acquire satellite and UAV remote sensing images of the target area, and automatically identify and extract the initial boundary of surface lithology and structural clues from the satellite and UAV remote sensing images using a convolutional neural network;

[0031] In one implementation, satellite and UAV remote sensing images of the target area are acquired, and the initial boundaries of surface lithology and structural clues are automatically identified and extracted from the satellite and UAV remote sensing images using a convolutional neural network. Step A1 may further include:

[0032] Acquire initial satellite and UAV remote sensing images of the target area, perform image preprocessing, and obtain satellite and UAV remote sensing images. The image preprocessing includes radiometric correction, geometric correction, and orthorectification.

[0033] Construct a geological label sample library, and use the geological label sample library to train a convolutional neural network to obtain a trained convolutional neural network;

[0034] The trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images and extract the initial boundary of surface lithology and tectonic clues.

[0035] In one implementation, a trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images, extracting initial boundaries of surface lithology and structural clues. Step A1 may further include:

[0036] Load a pre-trained convolutional neural network, input satellite and UAV remote sensing images into the convolutional neural network, perform semantic segmentation, and automatically identify and output the initial boundary of surface lithology and the initial tectonic clues;

[0037] Post-processing is performed on the initial surface lithological boundaries and initial structural clues to obtain the initial surface lithological boundaries and structural clues;

[0038] The post-processing includes removing small-scale noise patches, smoothing the initial boundaries of the initialized surface lithology, and repairing the connectivity of the initialized structural clues and extracting the skeleton.

[0039] It should be noted that satellite and UAV remote sensing imagery refers to multi-scale remote sensing data acquired through Earth observation satellites and low-altitude UAV platforms. Satellite remote sensing imagery focuses on providing continuous coverage information at the regional scale, while UAV remote sensing imagery is used to acquire high-resolution detail information of local areas. Initial lithological boundaries refer to the preliminary boundaries between different lithological units on the Earth's surface, obtained through automatic image feature identification and human-computer interactive interpretation. Tectonic clues refer to geological structural features that appear linear or banded in remote sensing imagery, including faults, shear zones, and folds.

[0040] In one embodiment, the acquired satellite and UAV remote sensing images undergo standardized preprocessing. For example, atmospheric correction is performed on the satellite remote sensing images to convert digital quantization values ​​into surface reflectance. Simultaneously, orthorectification is performed on the satellite and UAV remote sensing images in conjunction with a high-precision DEM to eliminate image point displacement caused by terrain undulations, ensuring that all image pixels have uniform geographic coordinates and true geometric shapes.

[0041] On the preprocessed images, those skilled in the art manually delineate different lithological units, such as gabbro and serpentinite, and multi-scale structures, such as regional faults and joint zones, using precise polygons or lines as labels, based on field measurement points or high-precision geological maps, thereby obtaining a geological label sample library. This geological label sample library is then used to train a convolutional neural network.

[0042] The trained convolutional neural network (CNN) can automatically perform semantic segmentation on new images. For each pixel in the input satellite and UAV remote sensing images, the CNN outputs a probability value indicating it belongs to ophiolite mélange, along with a binary map of the linear region of the image. The probability value is between 0 and 1. By setting a threshold, such as 0.7, the probability map is converted into a binary map, thus obtaining the initial boundary of the surface lithology.

[0043] Since the initial boundary of surface lithology is often jagged and contains many scattered small spots, post-processing is required to remove small-scale noise patches. Specifically, by calculating the area of ​​all connected regions and setting an empirical threshold (e.g., for areas with an actual ground area of ​​less than 100 square meters), all connected regions smaller than this threshold are removed from the binary map and smoothed. For the remaining binary regions, morphological closing operations are first performed to fill small holes and narrow gaps. Then, the boundary contours are extracted, and a polygon approximation algorithm is used to simplify the vertices of the boundary contours, thereby obtaining the initial boundary of surface lithology.

[0044] For binary images of linear regions, morphological closing operations are performed, such as using 3×3 linear structuring elements, to bridge closely spaced broken line segments. Then, thinning algorithms such as Zhang-Suen are applied to the repaired binary image, iteratively removing boundary pixels until the region width becomes a single pixel, thereby obtaining an accurate construct centerline skeleton. Finally, this skeleton is converted into construction cues.

[0045] By transforming the initial data-driven identification results into data that conforms to geological mapping standards and can be directly used for geological analysis and as constraints, the technical effect of providing reliable data support for subsequent work is achieved.

[0046] A2: The characteristic spectra of ophiolite in satellite and UAV remote sensing images are enhanced by using the spectral angle mapping algorithm. The characteristic spectral identification results are then fused with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite.

[0047] In one implementation, a spectral angle mapping algorithm is used to enhance the identification of characteristic spectra of ophiolite in satellite and UAV remote sensing images. The characteristic spectral identification results are then fused with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite. Step A2 may further include:

[0048] Based on field measurement points and geological data, multiple remote sensing pixel samples representing ophiolite were obtained. The reflectance characteristics of multiple remote sensing pixel samples under multispectral bands were statistically analyzed, and the end-member spectra of multiple ophiolite were extracted. After summarizing, a characteristic spectral library of ophiolite was obtained.

[0049] Using satellite and UAV remote sensing images as input data, a spectral angle mapping algorithm is used to calculate the spectral angle between the image pixel spectrum and the ophiolite endmember spectrum in the ophiolite characteristic spectral library pixel by pixel, generating an ophiolite spectral similarity distribution map. The smaller the spectral angle, the higher the similarity between the pixel and the ophiolite characteristic spectrum.

[0050] Based on a preset spectral angle threshold, a threshold determination is performed on the spectral similarity distribution map, and pixels that meet the threshold conditions are marked as potential ophiolite mélange pixels to obtain characteristic spectral identification results.

[0051] By integrating the characteristic spectral identification results with the initial boundary of surface lithology, the surface outcrop range of ophiolite mélange is obtained.

[0052] It should be noted that the endmember spectrum refers to the standard reflectance curve of a typical rock, such as diabase, basalt, or serpentinite, that constitutes the mixed pixel and represents a single pure land cover type. In this invention, it is an ophiolite mélange. The spectral similarity distribution map is a grayscale image generated by applying a spectral angle filling algorithm, where the grayscale value of each pixel represents the angle between the spectrum of that pixel and the target endmember spectrum.

[0053] In one embodiment, in conjunction with field surveys, clean pixel points are located on preprocessed remote sensing images in known ophiolite outcrop areas, and rock samples at the corresponding locations are collected for spectral measurements or the reflectance value of that point on the image is directly extracted. For example, in the eastern section of the Yarlung Tsangpo suture zone in the Tibet Autonomous Region, typical spectral curves of fresh peridotite, serpentinized peridotite, and gabbro are extracted from images, and multiple ophiolite end-member spectra are extracted to form a characteristic spectral library of ophiolite mélange.

[0054] Subsequently, satellite and UAV remote sensing imagery was used as input data, and a spectral angle mapping algorithm was employed to scan the entire area. For each pixel in the image, the algorithm calculated the angle between its spectral vector and the spectral vector of each endmember in the spectral library. Typically, the smallest angle value was taken as the spectral angle of that pixel. This generated a spectral similarity distribution map of ophiolite mélange; the smaller the value, the more similar the spectrum is to that of ophiolite mélange. By setting an empirical threshold, the spectral similarity distribution of ophiolite mélange was... Figure 2 Value-based identification yields the characteristic spectral identification results, marking all potential pixels similar to ophiolite mélange in spectral characteristics, thus obtaining the characteristic spectral identification results.

[0055] In one implementation, the characteristic spectral identification results are fused with the initial boundary of the surface lithology to obtain the outcrop range of the ophiolite mélange. Step A2 may further include:

[0056] The characteristic spectral identification results are spatially superimposed with the initial boundary of surface lithology. The initial boundary of surface lithology is used as a screening and correction range constraint to screen and correct the characteristic spectral identification results, thereby obtaining the corrected characteristic spectral identification results.

[0057] Regions that simultaneously meet the modified characteristic spectral identification results and the initial boundary of surface lithology are identified as high-confidence ophiolite outcrops.

[0058] Areas that only meet the modified characteristic spectral identification results or the initial boundary of surface lithology are identified as low-confidence ophiolite outcrops. The low-confidence ophiolite outcrops are then screened through regional connectivity analysis to obtain the range of ophiolite surface outcrops.

[0059] In one possible embodiment, the feature spectral recognition results generated by the spectral angle filling algorithm are spatially superimposed with the initial surface lithology boundary extracted and post-processed by the convolutional neural network. A buffer is created, for example, extending outward by 5-10 meters to accommodate boundary positioning errors, and the polygons of the polygon vector data of the initial surface lithology boundary and its buffer range are converted into a mask raster. Then, a bitwise AND operation is performed on the raster of the feature spectral recognition results, and consistency recognition is performed only on spectral recognition pixels falling within the mask range. If recognition fails, they are discarded, thus obtaining the corrected feature spectral recognition results.

[0060] Through spatial overlay analysis, the raster of the corrected characteristic spectral identification results is precisely compared with the polygon of the initial boundary of the surface lithology. Areas that are both within the initial boundary of the surface lithology and whose corresponding pixels are also in the corrected characteristic spectral identification results are designated as high-confidence ophiolite outcrops.

[0061] For regions that only meet the criteria for corrected feature spectral identification, regional connectivity analysis is performed. For example, the 8-neighborhood connectivity rule is used to identify all independent patches, and the area of ​​each patch is calculated. If the area of ​​a patch is less than a certain threshold, such as 100 square meters on the ground, it is considered to be noise or sporadic alteration and is discarded; if the area is large and the shape is relatively regular, it may represent ophiolite mélange that was missed by the convolutional neural network, which is thinly covered or severely weathered, and is retained but marked as low confidence.

[0062] For areas that only meet the initial lithological boundaries, this may be due to severe weathering, complete vegetation cover leading to spectral feature obliteration, or incorrect lithology identification. Such areas are typically marked directly as low-confidence areas, or manually verified using higher-resolution UAV imagery. Merging high-confidence areas with the selected low-confidence areas and vectorizing the data yields the complete outcrop extent of the ophiolite mélange. By establishing a dual verification mechanism and a hierarchical confidence system, the technical effects of information complementarity and error control are achieved.

[0063] A3: Within the area of ​​the ophiolite outcrop, lay out gravity exploration, magnetic exploration and electrical exploration survey lines and points, and collect gravity data, magnetic data and electrical data;

[0064] Preferably, the exploration network is planned with the outcrops of ophiolite mélange as the core target area, combined with topographic maps and access conditions. For example, in an ophiolite mélange belt in Xinjiang Uygur Autonomous Region, a regular grid covering a wider area, such as a point spacing of 100 meters × 100 meters, is deployed for a diabase outcrop area to capture the regional high gravity background and its boundary gradient zone caused by the rock mass. In the core area of ​​high-density outcrops, the spacing of measuring points can be increased to 50 meters. High-precision relative gravimeters (such as CG-5) are used to observe each point, and the precise latitude, longitude, elevation, and time are recorded to obtain gravity readings, thereby obtaining gravity data.

[0065] Continuous magnetic surveys can be conducted along gravity survey lines using a high-sensitivity proton magnetometer, either on foot or on a vehicle, with sampling intervals up to 0.1 seconds. Magnetometer readings are then used as magnetic data. The survey line direction should be as perpendicular as possible to the inferred strike of the rock mass. Considering the electrical differences between the ophiolite mélange and the surrounding rocks (e.g., serpentinization may lead to medium-to-low resistivity), magnetotelluric (MT) or controlled-source audio-frequency magnetotelluric (CSAMT) points are selected at key profiles, such as those traversing the outcrop center and extending to the surrounding rocks on both sides. The point spacing is set according to the detection depth and target size; for example, from the outcrop center outwards, the point spacing gradually increases from 50 meters to 200 meters. The electrical data are obtained based on the collected electromagnetic field components.

[0066] A4: Under geological constraints, the underground physical structure is deduced from gravity data, magnetic data and electrical data through the conjugate gradient inversion algorithm, and an underground physical structure model is constructed. The geological constraints include taking the surface outcrop range of the ophiolite as a priori lithological boundary constraint and the tectonic clues as structural constraints.

[0067] In one implementation, under geological constraints, a subsurface physical structure model is constructed by extrapolating gravity, magnetic, and electrical resistivity data using a conjugate gradient inversion algorithm. The geological constraints include using the surface outcrop range of the ophiolite mélange as a priori lithological boundary constraint and the tectonic clues as structural constraints. Step A4 may further include:

[0068] The gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset.

[0069] A three-dimensional inversion grid is constructed based on the a priori constraints of lithological boundaries and structural constraints in geological constraints.

[0070] The objective function in the conjugate gradient inversion algorithm is used to iteratively invert and solve the three-dimensional inversion network based on the standardized multi-source geophysical dataset. When the convergence condition is met, the underground physical property structure model is output.

[0071] In one implementation, the gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset. Step A4 may further include:

[0072] Gravity data is subjected to drift correction, tidal correction, topographic correction, and anomaly separation to obtain gravity anomaly data, wherein the gravity anomaly data is used to reflect changes in underground density.

[0073] The magnetic data is subjected to diurnal variation correction, normal field correction and anomaly separation processing to obtain magnetic anomaly data, wherein the magnetic anomaly data is used to reflect the changes in underground magnetic susceptibility.

[0074] The electrical resistivity data are subjected to noise suppression and preliminary inversion processing to obtain electrical resistivity response data, which is used to reflect changes in underground resistivity.

[0075] The coordinates of the gravity anomaly data, magnetic anomaly data, and electrical response data are unified to obtain a standardized multi-source geophysical dataset.

[0076] It should be noted that for gravity data, instrument drift correction is performed to eliminate changes in the instrument zero point over time, and solid tide correction is performed to eliminate gravity variations caused by solar and lunar gravitational forces. Then, a high-precision DEM is used, and the gravity effects caused by topographic undulations are removed through integration calculations. Topographic correction is typically achieved using intermediate layer correction and residual topographic correction. Finally, regional field separation is used to obtain gravity anomaly data reflecting local density volumes.

[0077] For magnetic data, base station data is used to eliminate the influence of diurnal variations in the geomagnetic field, and the Earth's main magnetic field is removed using the IGRF model to obtain magnetic anomaly data. For electrical data, noise removal is required, such as far reference trace processing and preliminary one-dimensional inversion, converting apparent resistivity and phase data into inversion impedance data or pseudo-profilometry data that more closely approximate the actual formation response. Finally, the coordinates of all measurement points are unified to the same projected coordinate system to obtain a standardized multi-source geophysical dataset.

[0078] A three-dimensional mesh was established based on the study area and exploration depth. The polygons representing the surface outcrops of the ophiolite mélange, interpreted by remote sensing, were used as prior constraints for lithological boundaries. In the modeling software, these polygons were vertically projected onto the surface of the three-dimensional mesh. Top mesh cells covered by the projection were marked as known ophiolite mélange cells and assigned a high prior density value (e.g., 2.8 g / cm³) and a narrow variation range (e.g., 2.7-3.0 g / cm³). During inversion, the physical property values ​​of these cells were strongly anchored. Simultaneously, interpreted structural clues, such as the main fault line, were used as structural constraints. Based on the fault attitude, a structural plane was constructed in three-dimensional space. Within this plane, the inversion algorithm applied anisotropic smoothing constraints to the mesh cells, meaning that physical property changes were gradual along the fault strike and dip directions, while properties perpendicular to the fault plane could change drastically, thus obtaining the three-dimensional inversion mesh.

[0079] Initialize the entire 3D mesh. For known ophiolite mélange elements, assign prior physical property values; for other areas, assign regional background values. Based on the current model, use forward modeling algorithms based on gravity, magnetism, and electrical resistivity to calculate the theoretical geophysical response, including gravity anomalies, magnetic anomalies, and apparent resistivity. Calculate the gradient of the objective function with respect to the model parameters. This gradient indicates the direction in which the model should be corrected to reduce the objective function. The efficiency of the conjugate gradient method depends on the accurate calculation of the gradient. Using the calculated gradient, update the model parameters according to the search direction and step size rules of the conjugate gradient method. After each iteration, check whether the model parameters exceed preset physical limits, such as density not being negative and geological constraint limits. For strongly constrained elements, if their values ​​deviate from prior values ​​beyond the allowable range, project them back into the allowable range. This step directly reflects the hard constraints. Determine whether the decrease in the objective function or the change in the model is less than a preset threshold, and whether the data fit reaches the expected level. If convergence is not achieved, proceed to the next iteration. When the iteration converges, output the final subsurface physical structure model.

[0080] A5: Based on implicit modeling technology, the surface outcrop range of ophiolite mélange is integrated with the underground geophysical body structure model to obtain a geological-geophysical spatial structure interpretation map.

[0081] In one implementation, the surface outcrop range of ophiolite mélange is fused with the subsurface geophysical structure model based on implicit modeling technology to obtain a geological-geophysical spatial structure interpretation map. Step A5 may further include:

[0082] Acquire spatial vector data of the surface outcrop range of ophiolite, three-dimensional volumetric data of the underground geophysical body structure model, and topographic data, and construct a three-dimensional implicit geological scalar field by executing a coordinate system.

[0083] From the three-dimensional implicit geological scalar field, the isosurface representing the interface between the ophiolite mélange and the surrounding rock is extracted, and the isosurface is cut and block-based in three-dimensional space to generate a three-dimensional geological structure model composed of ophiolite mélange block units and matrix units.

[0084] By correlating and fusing the physical properties of the three-dimensional geological structure model with those of the underground geophysical body structure model, a geological-geophysical spatial structure interpretation map is obtained.

[0085] In one implementation, step A5 may further include:

[0086] Based on spatial vector data, the boundary and internal points of the surface outcrop range of the ophiolite mélange are taken as the first type of spatial control points and assigned a first preset field value.

[0087] The three-dimensional volumetric data and terrain data of the underground geophysical body structure model are used as spatial trend constraints.

[0088] Based on the radial basis function interpolation algorithm, and combined with the first type of spatial control points, the first preset field value and the spatial trend constraints, a three-dimensional implicit geological scalar field is constructed.

[0089] It should be noted that all input data—spatial vector data of the ophiolite outcrop area, 3D volumetric data of the underground geophysical structure model, and topographic data—are mapped to a single coordinate system. Then, based on the spatial vector data, the boundary lines of the ophiolite outcrop area and a uniformly sampled set of points within it are extracted as first-class spatial control points. These control points are assigned a first preset field value; for example, all control points representing ophiolite outcrops are assigned a scalar field value G = 1. Simultaneously, additional control points can be set in known typical surrounding rock areas, such as based on regional geological maps, and assigned a value of G = -1. These points and their assigned values ​​constitute the hard constraints that the model must strictly adhere to. Next, the underground geophysical structure model (and topographic data) are used as spatial trend constraints. For example, high-density areas, such as volumetric regions >2.9 g / cm³, are guided towards a field value of +1, i.e., ophiolite; while low-density areas are guided towards -1, i.e., surrounding rock.

[0090] The terrain surface is used as the natural boundary constraint at the top of the model. Finally, based on the radial basis function interpolation algorithm, such as using the Gaussian kernel function, a smooth, continuous three-dimensional implicit geological scalar field is calculated using the first type of spatial control points and their preset field values ​​as the accurate interpolation targets, and the physical property trends and terrain as the background field. This field is exactly equal to the preset values ​​at the surface control points, and follows the physical property distribution trend at depth, achieving a mathematical unity between surface determinism and subsurface inference.

[0091] From the constructed 3D implicit geological scalar field, isosurfaces with field values ​​G=0 are extracted. Mathematically, these isosurfaces represent the transition interface from ophiolite mélange blocks (G>0) to the matrix (G<0), i.e., the geological contact boundary. Since the 3D implicit geological scalar field is generated by the joint constraints of surface outcrops and subsurface properties, this interface perfectly matches the input outcrop vector boundary at the surface and extends naturally along the gradient zone of property changes subsurface, forming a reasonable 3D extension of the geological interface. Then, using this isosurface with G=0 as a cutting tool, the entire 3D space is cut and segmented: all spatial regions with G>0 are defined as ophiolite mélange block units, and all regions with G<0 are defined as matrix units, thereby generating a 3D geological entity model with a clear topological structure.

[0092] The aforementioned three-dimensional geological structure model is then integrated with the subsurface geophysical body structure model. Specifically, this involves spatial location mapping, assigning physical attribute values ​​from the geophysical body model, such as the density value of each volume element, to the corresponding geological block unit. For example, for a single ophiolite mélange block unit, the values ​​of all mapped density volume elements within it can be statistically analyzed to calculate the average density, density standard deviation, and other attributes of the geological unit. These attributes are then used as the unit's physical property labels, thereby obtaining the aforementioned geological-geophysical spatial structure interpretation map. Through implicit modeling, deterministic surface geological maps, inferential subsurface geophysical models, and topographic information are integrated into a three-dimensional geological digital model, significantly improving the reliability of deep structural inferences and providing reliable three-dimensional geological evidence for resource exploration and engineering geological evaluation.

[0093] Example 2: Based on the same inventive concept as the multi-source data collaborative air-space-ground integrated interpretation method for ophiolite mélange in the foregoing examples, this invention provides a multi-source data collaborative air-space-ground integrated interpretation system for ophiolite mélange. See [link to relevant documentation]. Figure 2 As shown, the system includes:

[0094] Boundary extraction module 11 is used to acquire satellite and UAV remote sensing images of the target area, and automatically identify and extract the initial boundary of surface lithology and structural clues from the satellite and UAV remote sensing images through a convolutional neural network.

[0095] The outcrop range acquisition module 12 is used to enhance the identification of the characteristic spectra of ophiolite in satellite and UAV remote sensing images using the spectral angle mapping algorithm, and to fuse the characteristic spectral identification results with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite.

[0096] Data acquisition module 13 is used to lay out gravity exploration, magnetic exploration and electrical exploration survey lines and points within the outcrop area of ​​the ophiolite, and to collect gravity data, magnetic data and electrical data.

[0097] The structural model construction module 14 is used to perform subsurface physical structure inference on gravity data, magnetic data and electrical data under geological constraints by using the conjugate gradient inversion algorithm, and to construct a subsurface physical structure model. The geological constraints include taking the surface outcrop range of the ophiolite mélange as a priori lithological boundary constraint and the tectonic clues as structural constraints.

[0098] The interpretation map acquisition module 15 is used to integrate the surface outcrop range of ophiolite mélange with the underground geophysical body structure model based on implicit modeling technology to obtain a geological-geophysical spatial structure interpretation map.

[0099] In one implementation, the boundary extraction module 11 is used to perform the following steps:

[0100] Acquire initial satellite and UAV remote sensing images of the target area, perform image preprocessing, and obtain satellite and UAV remote sensing images. The image preprocessing includes radiometric correction, geometric correction, and orthorectification.

[0101] Construct a geological label sample library, and use the geological label sample library to train a convolutional neural network to obtain a trained convolutional neural network;

[0102] The trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images and extract the initial boundary of surface lithology and tectonic clues.

[0103] In one implementation, the boundary extraction module 11 is used to perform the following steps:

[0104] Load a pre-trained convolutional neural network, input satellite and UAV remote sensing images into the convolutional neural network, perform semantic segmentation, and automatically identify and output the initial boundary of surface lithology and the initial tectonic clues;

[0105] Post-processing is performed on the initial surface lithological boundaries and initial structural clues to obtain the initial surface lithological boundaries and structural clues;

[0106] The post-processing includes removing small-scale noise patches, smoothing the initial boundaries of the initialized surface lithology, and repairing the connectivity of the initialized structural clues and extracting the skeleton.

[0107] In one implementation, the outcrop range acquisition module 12 is used to perform the following steps:

[0108] Based on field measurement points and geological data, multiple remote sensing pixel samples representing ophiolite were obtained. The reflectance characteristics of multiple remote sensing pixel samples under multispectral bands were statistically analyzed, and the end-member spectra of multiple ophiolite were extracted. After summarizing, a characteristic spectral library of ophiolite was obtained.

[0109] Using satellite and UAV remote sensing images as input data, a spectral angle mapping algorithm is used to calculate the spectral angle between the image pixel spectrum and the ophiolite endmember spectrum in the ophiolite characteristic spectral library pixel by pixel, generating an ophiolite spectral similarity distribution map. The smaller the spectral angle, the higher the similarity between the pixel and the ophiolite characteristic spectrum.

[0110] Based on a preset spectral angle threshold, a threshold determination is performed on the spectral similarity distribution map, and pixels that meet the threshold conditions are marked as potential ophiolite mélange pixels to obtain characteristic spectral identification results.

[0111] By integrating the characteristic spectral identification results with the initial boundary of surface lithology, the surface outcrop range of ophiolite mélange is obtained.

[0112] In one implementation, the outcrop range acquisition module 12 is used to perform the following steps:

[0113] The characteristic spectral identification results are spatially superimposed with the initial boundary of surface lithology. The initial boundary of surface lithology is used as a screening and correction range constraint to screen and correct the characteristic spectral identification results, thereby obtaining the corrected characteristic spectral identification results.

[0114] Regions that simultaneously meet the modified characteristic spectral identification results and the initial boundary of surface lithology are identified as high-confidence ophiolite outcrops.

[0115] Areas that only meet the modified characteristic spectral identification results or the initial boundary of surface lithology are identified as low-confidence ophiolite outcrops. The low-confidence ophiolite outcrops are then screened through regional connectivity analysis to obtain the range of ophiolite surface outcrops.

[0116] In one implementation, the structural model building module 14 is used to perform the following steps:

[0117] The gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset.

[0118] A three-dimensional inversion grid is constructed based on the a priori constraints of lithological boundaries and structural constraints in geological constraints.

[0119] The objective function in the conjugate gradient inversion algorithm is used to iteratively invert and solve the three-dimensional inversion network based on the standardized multi-source geophysical dataset. When the convergence condition is met, the underground physical property structure model is output.

[0120] In one implementation, the structural model building module 14 is used to perform the following steps:

[0121] Gravity data is subjected to drift correction, tidal correction, topographic correction, and anomaly separation to obtain gravity anomaly data, wherein the gravity anomaly data is used to reflect changes in underground density.

[0122] The magnetic data is subjected to diurnal variation correction, normal field correction and anomaly separation processing to obtain magnetic anomaly data, wherein the magnetic anomaly data is used to reflect the changes in underground magnetic susceptibility.

[0123] The electrical resistivity data are subjected to noise suppression and preliminary inversion processing to obtain electrical resistivity response data, which is used to reflect changes in underground resistivity.

[0124] The coordinates of the gravity anomaly data, magnetic anomaly data, and electrical response data are unified to obtain a standardized multi-source geophysical dataset.

[0125] In one implementation, the interpretation map acquisition module 15 is used to perform the following steps:

[0126] Acquire spatial vector data of the surface outcrop range of ophiolite, three-dimensional volumetric data of the underground geophysical body structure model, and topographic data, and construct a three-dimensional implicit geological scalar field by executing a coordinate system.

[0127] From the three-dimensional implicit geological scalar field, the isosurface representing the interface between the ophiolite mélange and the surrounding rock is extracted, and the isosurface is cut and block-based in three-dimensional space to generate a three-dimensional geological structure model composed of ophiolite mélange block units and matrix units.

[0128] By correlating and fusing the physical properties of the three-dimensional geological structure model with those of the underground geophysical body structure model, a geological-geophysical spatial structure interpretation map is obtained.

[0129] In one implementation, the interpretation map acquisition module 15 is used to perform the following steps:

[0130] Based on spatial vector data, the boundary and internal points of the surface outcrop range of the ophiolite mélange are taken as the first type of spatial control points and assigned a first preset field value.

[0131] The three-dimensional volumetric data and terrain data of the underground geophysical body structure model are used as spatial trend constraints.

[0132] Based on the radial basis function interpolation algorithm, and combined with the first type of spatial control points, the first preset field value and the spatial trend constraints, a three-dimensional implicit geological scalar field is constructed.

[0133] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0134] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A multi-source data collaborative method for integrated air-space-ground interpretation of ophiolite mélange, characterized in that, The method includes: Acquire satellite and UAV remote sensing images of the target area, and automatically identify and extract the initial boundary of surface lithology and tectonic clues from the satellite and UAV remote sensing images using a convolutional neural network; The characteristic spectra of ophiolite in satellite and UAV remote sensing images were enhanced by using spectral angle mapping algorithm. The characteristic spectral identification results were then fused with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite. Within the area of ​​the ophiolite outcrop, gravity exploration, magnetic exploration and electrical exploration survey lines and points were laid out to collect gravity data, magnetic data and electrical data; Under geological constraints, the subsurface physical structure is deduced from gravity data, magnetic data and electrical data using the conjugate gradient inversion algorithm, and a subsurface physical structure model is constructed. The geological constraints include taking the surface outcrop range of the ophiolite mélange as a priori lithological boundary constraint and the tectonic clues as structural constraints. By integrating the surface outcrop range of ophiolite mélange with the subsurface geophysical structure model using implicit modeling techniques, a geological-geophysical spatial structure interpretation map is obtained.

2. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 1, characterized in that, Acquire satellite and UAV remote sensing images of the target area, and automatically identify and extract initial boundaries of surface lithology and structural clues from the satellite and UAV remote sensing images using a convolutional neural network, including: Acquire initial satellite and UAV remote sensing images of the target area, perform image preprocessing, and obtain satellite and UAV remote sensing images. The image preprocessing includes radiometric correction, geometric correction, and orthorectification. Construct a geological label sample library, and use the geological label sample library to train a convolutional neural network to obtain a trained convolutional neural network; The trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images and extract the initial boundary of surface lithology and tectonic clues.

3. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 2, characterized in that, The trained convolutional neural network is loaded to automatically identify satellite and UAV remote sensing images, extracting initial boundaries of surface lithology and tectonic clues, including: Load a pre-trained convolutional neural network, input satellite and UAV remote sensing images into the convolutional neural network, perform semantic segmentation, and automatically identify and output the initial boundary of surface lithology and the initial tectonic clues; Post-processing is performed on the initial surface lithological boundaries and initial structural clues to obtain the initial surface lithological boundaries and structural clues; The post-processing includes removing small-scale noise patches, smoothing the initial boundaries of the initialized surface lithology, and repairing the connectivity of the initialized structural clues and extracting the skeleton.

4. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 1, characterized in that, The spectral angle mapping algorithm was used to enhance the identification of characteristic spectra of ophiolite mélange in satellite and UAV remote sensing images. The characteristic spectral identification results were then fused with the initial surface lithological boundaries to obtain the surface outcrop range of ophiolite mélange, including: Based on field measurement points and geological data, multiple remote sensing pixel samples representing ophiolite were obtained. The reflectance characteristics of multiple remote sensing pixel samples under multispectral bands were statistically analyzed, and the end-member spectra of multiple ophiolite were extracted. After summarizing, a characteristic spectral library of ophiolite was obtained. Using satellite and UAV remote sensing images as input data, a spectral angle mapping algorithm is used to calculate the spectral angle between the image pixel spectrum and the ophiolite endmember spectrum in the ophiolite characteristic spectral library pixel by pixel, generating an ophiolite spectral similarity distribution map. The smaller the spectral angle, the higher the similarity between the pixel and the ophiolite characteristic spectrum. Based on a preset spectral angle threshold, a threshold determination is performed on the spectral similarity distribution map, and pixels that meet the threshold conditions are marked as potential ophiolite mélange pixels to obtain characteristic spectral identification results. By integrating the characteristic spectral identification results with the initial boundary of surface lithology, the surface outcrop range of ophiolite mélange is obtained.

5. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 4, characterized in that, By integrating the characteristic spectral identification results with the initial surface lithological boundary, the surface outcrop range of the ophiolite mélange is obtained, including: The characteristic spectral identification results are spatially superimposed with the initial boundary of surface lithology. The initial boundary of surface lithology is used as a screening and correction range constraint to screen and correct the characteristic spectral identification results, thereby obtaining the corrected characteristic spectral identification results. Regions that simultaneously meet the modified characteristic spectral identification results and the initial boundary of surface lithology are identified as high-confidence ophiolite outcrops. Areas that only meet the modified characteristic spectral identification results or the initial boundary of surface lithology are identified as low-confidence ophiolite outcrops. The low-confidence ophiolite outcrops are then screened through regional connectivity analysis to obtain the range of ophiolite surface outcrops.

6. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 1, characterized in that, Under geological constraints, a subsurface physical structure model is constructed by extrapolating gravity, magnetic, and electrical resistivity data using a conjugate gradient inversion algorithm. The geological constraints include using the surface outcrop range of the ophiolite mélange as a priori lithological boundary constraint and the tectonic clues as structural constraints. The gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset. A three-dimensional inversion grid is constructed based on the a priori constraints of lithological boundaries and structural constraints in geological constraints. The objective function in the conjugate gradient inversion algorithm is used to iteratively invert and solve the three-dimensional inversion network based on the standardized multi-source geophysical dataset. When the convergence condition is met, the underground physical property structure model is output.

7. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 6, characterized in that, The gravity data, magnetic data, and electrical data are processed to obtain a standardized multi-source geophysical dataset, including: Gravity data is subjected to drift correction, tidal correction, topographic correction, and anomaly separation to obtain gravity anomaly data, wherein the gravity anomaly data is used to reflect changes in underground density. The magnetic data is subjected to diurnal variation correction, normal field correction and anomaly separation processing to obtain magnetic anomaly data, wherein the magnetic anomaly data is used to reflect the changes in underground magnetic susceptibility. The electrical resistivity data are subjected to noise suppression and preliminary inversion processing to obtain electrical resistivity response data, which is used to reflect changes in underground resistivity. The coordinates of the gravity anomaly data, magnetic anomaly data, and electrical response data are unified to obtain a standardized multi-source geophysical dataset.

8. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 1, characterized in that, By fusing the surface outcrop extent of ophiolite mélange with the subsurface geophysical structure model using implicit modeling techniques, a geological-geophysical spatial structure interpretation map is obtained, including: Acquire spatial vector data of the surface outcrop range of ophiolite, three-dimensional volumetric data of the underground geophysical body structure model, and topographic data, and construct a three-dimensional implicit geological scalar field by executing a coordinate system. From the three-dimensional implicit geological scalar field, the isosurface representing the interface between the ophiolite mélange and the surrounding rock is extracted, and the isosurface is cut and block-based in three-dimensional space to generate a three-dimensional geological structure model composed of ophiolite mélange block units and matrix units. By correlating and fusing the physical properties of the three-dimensional geological structure model with those of the underground geophysical body structure model, a geological-geophysical spatial structure interpretation map is obtained.

9. The method for integrated air-space-ground interpretation of ophiolite mélange using multi-source data collaboration as described in claim 8, characterized in that, include: Based on spatial vector data, the boundary and internal points of the surface outcrop range of the ophiolite mélange are taken as the first type of spatial control points and assigned a first preset field value. The three-dimensional volumetric data and terrain data of the underground geophysical body structure model are used as spatial trend constraints. Based on the radial basis function interpolation algorithm, and combined with the first type of spatial control points, the first preset field value and the spatial trend constraints, a three-dimensional implicit geological scalar field is constructed.

10. A multi-source data collaborative interpretation system for ophiolite mélange using an integrated air-space-ground approach, characterized in that: For implementing the method steps of any one of claims 1 to 9, including: The boundary extraction module is used to acquire satellite and UAV remote sensing images of the target area, and automatically identify and extract the initial boundary of surface lithology and structural clues from the satellite and UAV remote sensing images through a convolutional neural network. The outcrop range acquisition module is used to enhance the identification of the characteristic spectra of ophiolite in satellite and UAV remote sensing images using spectral angle mapping algorithms, and to fuse the characteristic spectral identification results with the initial boundary of surface lithology to obtain the surface outcrop range of ophiolite. The data acquisition module is used to lay out gravity exploration, magnetic exploration and electrical exploration survey lines and points within the outcrop area of ​​the ophiolite mélange, and to collect gravity data, magnetic data and electrical data. The structural model construction module is used to extrapolate the underground physical property structure from gravity data, magnetic data, and electrical data under geological constraints using a conjugate gradient inversion algorithm, and to construct an underground physical property structure model. The geological constraints include using the surface outcrop range of the ophiolite mélange as a priori lithological boundary constraint and the tectonic clues as structural constraints. The interpretation map acquisition module is used to integrate the surface outcrop range of ophiolite mélange with the subsurface geophysical structure model based on implicit modeling technology to obtain a geological-geophysical spatial structure interpretation map.

Images