Hyperspectral remote sensing image analysis of hydrothermal alteration zone identification method and device

Abstract

The present application relates to the technical field of hyperspectral remote sensing image analysis, and particularly relates to a method and device for identifying and delineating a hydrothermal alteration zone based on hyperspectral remote sensing image analysis, which solves the technical problem that it is difficult to accurately delineate the real distribution range of a hydrothermal alteration zone under complex surface conditions due to the interference of same-spectrum different objects in the surface environment in the prior art. The method comprises: obtaining time-series hyperspectral data of a target region; extracting dynamic spectral feature parameters from the time-series hyperspectral data, and generating initial alteration mineral distribution information of the target region according to the dynamic spectral feature parameters; constructing a spatial permeability field according to geological structure data of the target region, and determining hydrothermal fluid migration path information according to the spatial permeability field and a preset pressure boundary condition; and performing spatial topological constraint correction on the initial alteration mineral distribution information and the hydrothermal fluid migration path information to obtain a delineation result of a hydrothermal alteration zone of the target region.

Description

Technical Field

[0001] This invention relates to the field of hyperspectral remote sensing image analysis technology, specifically to a method and apparatus for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis. Background Technology

[0002] Hyperspectral remote sensing technology can acquire detailed spectral information of ground objects within a continuous narrow band. Due to this technical characteristic, it is widely used in the field of mineral resource exploration. It has important application value in the identification of hydrothermal alteration minerals and the delineation of alteration zones. Different types of alteration minerals, such as sericite, kaolinite, and chlorite, have characteristic absorption peaks in the shortwave infrared band. By analyzing the spectral characteristics of hyperspectral images, alteration minerals can be identified and potential mineralization areas can be inferred.

[0003] Most existing hyperspectral remote sensing alteration identification methods rely on the analysis of static reflectance spectra of ground objects under natural lighting conditions. The identification process mainly uses spectral matching, statistical classification or machine learning algorithms to compare and judge the morphological features of spectral curves, and completes the identification and judgment of altered minerals only from the spectral perspective.

[0004] In practical applications, the Earth's surface is easily affected by factors such as vegetation cover, weathering crust, bare soil deposition, and changes in surface moisture, which can significantly alter the spectral response characteristics of ground features. This can cause different ground features to exhibit highly similar spectral curves, resulting in severe homonymy, which directly restricts the accuracy and reliability of alteration mineral identification. Traditional technical approaches only perform pixel-level identification from a single spectral dimension, lacking a comprehensive consideration of geological control factors such as hydrothermal fluid migration processes and fault structures. Consequently, the identification results often exhibit a discrete and isolated spatial distribution, making it difficult to form continuous and complete alteration zone boundaries. Summary of the Invention

[0005] To address the technical problem of existing technologies being susceptible to interference from foreign matter of the same spectrum in the surface environment, making it difficult to accurately delineate the true distribution range of hydrothermal alteration zones under complex surface conditions, the present invention aims to provide a method and apparatus for identifying and delineating hydrothermal alteration zones through hyperspectral remote sensing image analysis. The specific technical solution adopted is as follows: Firstly, a method for identifying and delineating hydrothermal alteration zones using hyperspectral remote sensing image analysis is provided, comprising: acquiring temporal hyperspectral data of the target area during the transmission of a preset energy pulse toward the target area; the temporal hyperspectral data includes static spectral frames before excitation, transient spectral frames after excitation, and spectral frames during the relaxation and recovery phase; extracting dynamic spectral feature parameters from the temporal hyperspectral data, and generating initial alteration mineral distribution information of the target area based on the dynamic spectral feature parameters; constructing a spatial permeability field based on the geological structure data of the target area, and determining hydrothermal fluid migration path information based on the spatial permeability field and preset pressure boundary conditions; and performing spatial topological constraint correction on the initial alteration mineral distribution information and the hydrothermal fluid migration path information to obtain the delineation result of the hydrothermal alteration zone in the target area.

[0006] Based on the above technical solution, in the hydrothermal alteration zone identification and delineation method provided by the present invention using hyperspectral remote sensing image analysis, temporal spectral information including pre-excitation, excitation transient, and relaxation recovery stages is obtained through active energy excitation. Feature parameters reflecting the dynamic response of minerals are extracted from this information to generate an initial alteration mineral distribution. Simultaneously, a spatial permeability field is constructed based on geological structural data, and the hydrothermal fluid migration path is simulated. Then, the initial alteration mineral distribution and fluid migration path are spatially topologically constrained and corrected. This makes the alteration identification process not only dependent on static spectral features but also controlled by the physical migration laws of hydrothermal activity. Thus, under complex surface conditions, it effectively suppresses interference from homospectral foreign objects, eliminates isolated anomalous patches, and completes the alteration information of obscured areas. Finally, it obtains spatially continuous and geologically consistent hydrothermal alteration zone delineation results, significantly improving the accuracy and reliability of alteration zone identification and delineation.

[0007] In conjunction with the first aspect above, in one possible implementation, the aforementioned dynamic spectral feature parameters include: absorption peak shift, which characterizes the degree of change in absorption peak position before and after stimulation; absorption depth change rate, which characterizes the relative change in absorption depth before and after stimulation; and relaxation recovery time, which characterizes the time required for the spectrum to recover to a specified proportion of the resting state after stimulation. The method for extracting dynamic spectral feature parameters from time-series hyperspectral data specifically includes: for each pixel in the time-series hyperspectral data, extracting the average spectrum of the resting state before stimulation from the static spectral frame before stimulation, and extracting the average spectrum of the resting state before stimulation from the transient spectral frame after stimulation. Extract the stimulated transient peak spectrum; determine the absorption peak positions and absorption peak depths of the pre-stimulated resting state average spectrum and the stimulated transient peak spectrum within the preset characteristic absorption bands of altered minerals; determine the absorption peak shift based on the absorption peak positions of the pre-stimulated resting state average spectrum and the stimulated transient peak spectrum; determine the absorption depth change rate based on the absorption peak depths of the pre-stimulated resting state average spectrum and the stimulated transient peak spectrum; determine the relaxation recovery time based on the recovery process of each frame of relaxation spectrum after stimulation relative to the pre-stimulated resting state average spectrum.

[0008] In conjunction with the first aspect above, in one possible implementation, the method for generating initial alteration mineral distribution information of a target region based on dynamic spectral feature parameters specifically includes: constructing a dynamic spectral feature vector based on absorption peak shift, absorption depth change rate, and relaxation recovery time; matching the dynamic spectral feature vector with standard dynamic features of known alteration mineral categories to obtain the probability distribution of each pixel in the target region belonging to each known alteration mineral; and determining the alteration probability characterizing the possibility that each pixel belongs to a hydrothermal alteration region based on the probability distribution, as the initial alteration mineral distribution information.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, the method of matching the dynamic spectral feature vector with the standard dynamic features of known alteration mineral categories to obtain the probability distribution of each pixel in the target area belonging to an alteration mineral specifically includes: normalizing the standard feature vector formed by the dynamic spectral feature vector and the standard dynamic features of known alteration mineral categories; analyzing the similarity measure between the normalized dynamic spectral feature vector and the normalized standard feature vector; and determining the probability distribution of each pixel corresponding to each known alteration mineral category based on the similarity measure using a preset normalization exponential function.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the method for constructing a spatial permeability field based on geological structural data of the target area specifically includes: extracting fault structure distribution information and fracture development zone distribution information from the geological structural data; setting permeability reference values ​​for the fault structure region, fracture development region, and surrounding rock region respectively; the permeability reference value for the fault structure region is greater than the permeability reference value for the fracture development region, and the permeability reference value for the fracture development region is greater than the permeability reference value for the surrounding rock region; using the fault structure distribution information and fracture development zone distribution information as spatial constraints, and employing a preset spatial interpolation method to interpolate the discrete permeability reference values ​​to a grid that matches the spatial resolution of the time-series hyperspectral data, thereby forming a continuous spatial permeability field.

[0011] In conjunction with the first aspect above, in one possible implementation, the method for determining hydrothermal fluid migration path information based on a spatial permeability field and preset pressure boundary conditions specifically includes: setting pressure source points and pressure sink points on the spatial permeability field to form preset pressure boundary conditions; the pressure source points are located at fracture intersections or rock mass contact zones, and the pressure sink points are located at the boundary of the target area; by using Darcy's law, which describes fluid migration in porous media, a fluid velocity vector field is determined based on the spatial permeability field, preset fluid viscosity parameters, and preset pressure boundary conditions; and the set of pixel locations in the fluid velocity vector field where the velocity is greater than a preset velocity threshold is determined as hydrothermal fluid migration path information.

[0012] In conjunction with the first aspect mentioned above, in one possible implementation, the method for spatially topologically constraining the initial alteration mineral distribution information and hydrothermal fluid migration path information to obtain the hydrothermal alteration zone delineation result of the target area specifically includes: identifying high-probability non-path pixels in the initial alteration mineral distribution information whose alteration probability is higher than a first probability threshold but whose spatial location does not fall within the coverage area of ​​the hydrothermal fluid migration path information, and performing isolation determination and processing on the alteration probability of the high-probability non-path pixels based on the pixels in the preset spatial neighborhood of the high-probability non-path pixels; identifying suspected occlusion pixels within the coverage area of ​​the hydrothermal fluid migration path information whose alteration probability is lower than a second probability threshold, and performing completion processing on the alteration probability of the suspected occlusion pixels based on the pixels in the preset spatial neighborhood of the suspected occlusion pixels; and extracting the boundary of the continuous region with an alteration probability higher than the first probability threshold after processing to obtain the hydrothermal alteration zone delineation result.

[0013] In conjunction with the first aspect above, in one possible implementation, the method for isolating and processing the erosion probability of a high-probability non-path pixel based on pixels in a preset spatial neighborhood of the high-probability non-path pixel specifically includes: counting the number of pixels in the preset spatial neighborhood of the high-probability non-path pixel whose erosion probability is higher than a first probability threshold; if the number of pixels is less than a preset isolation determination threshold, classifying the high-probability non-path pixel as an isolated anomalous pixel and adjusting the erosion probability of the isolated anomalous pixel to a specified value lower than the first probability threshold; if the number of pixels is greater than or equal to the preset isolation determination threshold, keeping the erosion probability of the high-probability non-path pixel unchanged.

[0014] In conjunction with the first aspect above, in one possible implementation, the method for completing the erosion probability of a suspected occluded pixel based on pixels in a preset spatial neighborhood of the suspected occluded pixel specifically includes: filtering valid pixels in the preset spatial neighborhood of the suspected occluded pixel whose erosion probability is higher than a first probability threshold; determining interpolation weights based on the spatial distance between valid pixels and suspected occluded pixels; and weighting the erosion probability of valid pixels according to the interpolation weights to obtain the erosion probability after the suspected occluded pixel completion processing.

[0015] Secondly, a device for identifying and delineating hydrothermal alteration zones using hyperspectral remote sensing image analysis is provided, comprising: a temporal spectral acquisition unit, used to acquire temporal hyperspectral data of the target area during the transmission of a preset energy pulse to the target area; the temporal hyperspectral data includes pre-excitation static spectral frames, excitation transient spectral frames, and relaxation recovery stage spectral frames; a dynamic feature extraction unit, used to extract dynamic spectral feature parameters from the temporal hyperspectral data, and generate initial alteration mineral distribution information of the target area based on the dynamic spectral feature parameters; a fluid path determination unit, used to construct a spatial permeability field based on the geological structure data of the target area, and determine hydrothermal fluid migration path information based on the spatial permeability field and preset pressure boundary conditions; and a spatial constraint correction unit, used to perform spatial topological constraint correction on the initial alteration mineral distribution information and the hydrothermal fluid migration path information to obtain the hydrothermal alteration zone delineation result of the target area.

[0016] The present invention has the following beneficial effects: By actively exciting temporal spectral information including pre-excitation, excitation transient, and relaxation recovery stages, characteristic parameters reflecting the dynamic response of minerals are extracted to generate the initial alteration mineral distribution. Simultaneously, a spatial permeability field is constructed based on geological structural data, and hydrothermal fluid migration paths are simulated. The initial alteration mineral distribution and fluid migration paths are then spatially topologically constrained and corrected, making the alteration identification process not only dependent on static spectral features but also controlled by the physical migration laws of hydrothermal activity. This effectively suppresses interference from heterogeneous objects with the same spectrum, eliminates isolated anomalous patches, and completes the alteration information of obscured areas under complex surface conditions. Ultimately, spatially continuous and geologically consistent hydrothermal alteration zone delineation results are obtained, significantly improving the accuracy and reliability of alteration zone identification and delineation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions and advantages 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.

[0018] Figure 1 This is a schematic diagram of a hydrothermal alteration zone identification and delineation device for hyperspectral remote sensing image analysis according to an embodiment of the present invention; Figure 2 The present invention provides a flowchart of a method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis, as an embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method and apparatus for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method and apparatus for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis provided by the present invention.

[0022] Please see Figure 1 The diagram shows a schematic of a hydrothermal alteration zone identification and delineation device for hyperspectral remote sensing image analysis according to an embodiment of the present invention. The hydrothermal alteration zone identification and delineation device for hyperspectral remote sensing image analysis includes: a time-series spectrum acquisition unit 1, a dynamic feature extraction unit 2, a fluid path determination unit 3, and a spatial constraint correction unit 4.

[0023] The time-series spectral acquisition unit 1 is the core of the device's raw data acquisition. It consists of a tunable mid-infrared pulsed laser, a visible to short-wave infrared hyperspectral imaging camera, and a field-programmable gate array (FPGA) synchronous control board. This unit emits preset energy pulses matching the absorption bands of altered minerals to the target area through the tunable mid-infrared pulsed laser. With the help of the FPGA synchronous control board, the pulse emission and spectral acquisition are aligned at the nanosecond level. The unit synchronously acquires the static spectral frames before stimulation, the transient spectral frames after stimulation, and the spectral frames during the relaxation and recovery phase of the target area, generating raw time-series hyperspectral data containing spatiotemporal information. This provides a basic data source for subsequent feature analysis. The generated raw time-series hyperspectral data is directly transmitted to the dynamic feature extraction unit 2.

[0024] The dynamic feature extraction unit 2 is the core of the device for spectral data processing and feature analysis. The entity consists of an edge computing unit and a spectral feature analysis algorithm module. This unit receives the raw time-series hyperspectral data transmitted by the time-series spectral acquisition unit 1, first completes radiometric correction, geometric registration, noise removal and data standardization to obtain high-quality standardized time-series hyperspectral data, and then extracts three types of dynamic spectral feature parameters from the standardized data: absorption peak shift, absorption depth change rate and relaxation recovery time. Through feature matching and probability calculation, it generates the initial alteration mineral distribution information of the target area. The generated initial alteration mineral distribution information is directly transmitted to the spatial constraint correction unit 4 as the basic data for delineating the alteration zone.

[0025] The fluid path determination unit 3 is the core of the device for geological structure analysis and fluid migration simulation. The entity consists of a geological data processing terminal and a fluid numerical simulation calculation module. This unit loads the geological structure data of the target area, extracts the distribution information of fracture structures and fissure development zones, constructs a continuous spatial permeability field that matches the resolution of hyperspectral data, and completes fluid dynamics calculations in combination with preset pressure boundary conditions to determine the hydrothermal fluid migration path information of the target area. The generated hydrothermal fluid migration path information will be transmitted to the spatial constraint correction unit 4 to provide the core basis for the geological rationality constraint of alteration distribution.

[0026] The spatial constraint correction unit 4 is the core of the spatial coupling processing and final output of the device. The entity consists of a GIS spatial analysis module and an image contour extraction and processing module. This unit simultaneously receives the initial alteration mineral distribution information transmitted by the dynamic feature extraction unit 2 and the hydrothermal fluid migration path information transmitted by the fluid path determination unit 3. Through spatial topological constraints, it completes the identification of isolated abnormal pixels and adjustment of alteration probability, and the completion of alteration probability of suspected occlusion areas. Then, it extracts the continuous boundary of the high alteration probability area after processing, and finally outputs the complete hydrothermal alteration zone delineation result of the target area, completing the final processing and output of the entire process of the device.

[0027] Please see Figure 2 The diagram illustrates a flowchart of a method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis according to an embodiment of the present invention. This method includes: S1. During the process of transmitting a preset energy pulse to the target area, acquire the time-series hyperspectral data of the target area.

[0028] The time-series hyperspectral data includes pre-excitation static spectral frames, excitation transient spectral frames, and relaxation recovery phase spectral frames.

[0029] In some implementations, a coaxial system mounted on a drone or vehicle-mounted mobile platform is used to dynamically scan the target area and emit high-energy broadband flash pulses into it. The spectral range of these high-energy broadband flash pulses covers at least the visible light to short-wave infrared bands to provide full-band broadband illumination. The pulse width is set to 50 ms, and the energy density is controlled at 10 J / cm². 2 This energy pulse can induce transient energy level transitions or lattice vibrations in hydrothermal alteration minerals, causing hydrothermal alteration minerals containing hydroxyl or carbonate structures to produce detectable dynamic spectral responses. Background materials such as vegetation and soil do not have this specific response, thus enabling a preliminary distinction between alteration minerals and background materials.

[0030] During the emission of a preset energy pulse, pre-excitation static spectral frames, excitation transient spectral frames, and relaxation recovery phase spectral frames are simultaneously acquired. The acquisition process utilizes a hyperspectral imaging system covering the visible to short-wave infrared bands, such as a hyperspectral camera with a spectral resolution of 10 nm and a spatial resolution of 3 m. The pre-excitation static spectral frames consist of three continuously acquired resting-state spectra; the excitation transient spectral frames consist of ten time-series spectra acquired at the moment of excitation; and the relaxation recovery phase spectral frames consist of five relaxation spectra acquired after excitation. Geographic coordinates and excitation timestamps are recorded simultaneously during acquisition to ensure temporal and spatial alignment of the time-series data, forming a raw temporal hyperspectral data cube containing the time dimension. Each pixel corresponds to a time-series spectral curve composed of 18 spectral frames, fully covering the spectral changes during the resting, excitation, and relaxation phases.

[0031] The raw time-series hyperspectral data underwent preprocessing, including radiometric correction, geometric registration, and noise removal. Radiometric correction used a standard reflectance plate with 99% reflectance to radiometrically calibrate the raw data, converting the digital number (DN) values ​​of the remote sensing image into surface reflectance to eliminate radiometric bias. Geometric registration used the first frame of the spectrum in resting state as a reference, employing a pre-defined feature matching algorithm, such as the scale-invariant feature transform (SIFT) algorithm, to spatially register all time-series frames, controlling the registration error to within 0.5 pixels and eliminating geometric bias. Noise removal used a pre-defined wavelet transform algorithm to denoise each time-series spectrum, for example, using the db4 wavelet and a 3-level decomposition method, to eliminate random noise while preserving characteristic absorption peaks, thus reducing spectral noise. After preprocessing, an aligned and standardized time-series hyperspectral data cube was obtained, with consistent spatial positions across frames and effective suppression of spectral noise.

[0032] S2. Extract dynamic spectral feature parameters from time-series hyperspectral data, and generate initial alteration mineral distribution information for the target area based on the dynamic spectral feature parameters.

[0033] In some implementations, dynamic spectral characteristic parameters include: absorption peak shift, which characterizes the degree of change in the position of the absorption peak before and after stimulation; absorption depth change rate, which characterizes the relative change in absorption depth before and after stimulation; and relaxation recovery time, which characterizes the time required for the spectrum to recover to a specified proportion of the resting state after stimulation.

[0034] Specifically, from the preprocessed time-series hyperspectral data, the resting-state average spectrum and the stimulated transient peak spectrum of each pixel are extracted. The resting-state average spectrum is calculated from multiple frames of static spectra acquired before stimulation, and the stimulated transient peak spectrum is determined from the spectral frame with the largest reflectance change among the time-series spectra acquired during the stimulated transient period. Baseline correction is performed on the resting-state average spectrum and the stimulated transient peak spectrum to determine the absorption peak positions and absorption peak depths (defined as the vertical distance from the baseline to the absorption peak, eliminating the influence of spectral baseline drift) within the preset alteration mineral characteristic absorption bands. The preset alteration mineral characteristic absorption bands include the 2.2 μm ± 0.05 μm band corresponding to hydroxyl groups and the 2.35 μm ± 0.05 μm band corresponding to carbonate groups.

[0035] The absorption peak shift is determined by comparing the absorption peak positions in the average spectrum of the resting state before excitation with those in the transient peak spectrum after excitation. The absorption peak shift is the absolute value of the difference between the absorption peak positions before and after excitation, and it characterizes the degree of positional change of the absorption peak after excitation. The larger the shift, the more significant the change in the vibrational state of the mineral lattice.

[0036] The absorption depth change rate is determined based on the absorption peak depth of the average spectrum in the resting state before excitation and the absorption peak depth of the peak spectrum in the excited transient state. This rate characterizes the relative change in absorption peak depth between the excited transient state and the resting state. It is important to note that if the target region contains no target alteration minerals or has extremely low alteration mineral content, the absorption peak depth of the average spectrum in the resting state will be zero or close to zero. Based on this, a depth threshold for the resting state is preset (usually set to 3 times the preprocessed spectral noise level, which can be adaptively adjusted according to the actual situation). When the absorption peak depth of the average spectrum in the resting state is less than the depth threshold, the absorption depth change rate is calculated by dividing the difference between the absorption peak depth of the peak spectrum in the excited transient state and the absorption peak depth of the average spectrum in the resting state by the modulated denominator. The modulation method involves adding a preset minimum positive modulation factor (e.g., 10) to the absorption peak depth of the average spectrum in the resting state. -6 To avoid calculation failures caused by a zero or too small denominator; when the absorption peak depth of the resting average spectrum is greater than or equal to the depth threshold, the absorption depth change rate is calculated directly using the absorption peak depth of the resting average spectrum as the denominator, ensuring the calculation accuracy of normal alteration pixels.

[0037] The relaxation recovery time is determined by observing the recovery process of each frame of the relaxed spectrum after stimulation relative to the average spectrum at rest before stimulation. Using the average spectrum at rest as a reference, the time required for each frame of the relaxed spectrum after stimulation to recover to a specified percentage of the resting reflectance is calculated. This specified percentage can be set to 90%, and this is used to determine the relaxation recovery time. The relaxation recovery time characterizes the speed at which the spectrum recovers to its resting state after stimulation; different altered minerals exhibit different relaxation recovery time characteristics.

[0038] Furthermore, a dynamic spectral feature vector corresponding to each pixel is constructed based on the absorption peak shift, the rate of change of absorption depth, and the relaxation recovery time.

[0039] The dynamic spectral feature vector is matched with the standard dynamic features of known alteration mineral categories to obtain the probability distribution of each pixel in the target region belonging to each known alteration mineral. Specifically, this is achieved as follows: The dynamic spectral feature vector and the standard feature vector composed of standard dynamic features of known alteration mineral categories (including absorption peak shift, absorption depth change rate, and relaxation recovery time of standard minerals) are normalized, for example, through a maximum-min linear normalization method. To ensure that the normalization scale remains consistent and physically comparable in any scenario, the global maximum and minimum values ​​used for normalization are derived from a pre-built database of standard dynamic features of alteration minerals and combined with mineral spectral theory, serving as preset theoretical extrema. For example, the theoretical maximum value of absorption peak shift is set to 1 / 2 of the characteristic absorption band width, the absorption depth change rate is set to 100%, and the relaxation recovery time is set to the pulse interval duration. Using these preset theoretical extrema, normalization is performed on both the dynamic spectral feature vector and the standard feature vector to ensure that the normalization scale of the two types of features is completely consistent.

[0040] The similarity measure between the normalized dynamic spectral feature vector and the normalized standard feature vector is analyzed. Specifically, it is the inner product between the vectors. This can effectively reflect the degree of matching between feature vectors. The higher the similarity measure, the closer the pixel feature is to the standard feature of the corresponding altered mineral.

[0041] Based on the similarity metric, the probability distribution of each pixel corresponding to each known alteration mineral category is determined using a preset normalization exponential function (such as the Softmax function), expressed as: In the formula, This represents the probability that the p-th pixel belongs to the m-th known alteration mineral category; This represents the similarity measure between the dynamic spectral feature vector of the p-th pixel and the standard feature vector of the m-th known alteration mineral; e is a natural constant. This represents the indexed similarity measure between the dynamic spectral feature vector of the p-th pixel and the standard feature vector of the m-th known alteration mineral; This represents the similarity measure between the dynamic spectral feature vector of the p-th pixel and the standard feature vector of the k-th known alteration mineral; K represents the total number of known alteration mineral categories. This represents the sum of the indexed similarity measures between the dynamic spectral feature vector of the p-th pixel and the standard feature vectors of all K known alteration minerals.

[0042] For each pixel, the probability distribution of all known alteration mineral categories corresponding to it is traversed, and the maximum value is taken as the alteration probability that each pixel belongs to a hydrothermal alteration region. This represents the highest matching degree between the pixel and the standard dynamic characteristics of a certain alteration mineral, and directly quantifies the possibility that the pixel belongs to a hydrothermal alteration region.

[0043] Based on preset high-probability and low-probability thresholds, alteration probabilities are divided into regions. Regions with alteration probabilities greater than or equal to the high-probability threshold are considered suspected alteration mineral distribution areas; regions with alteration probabilities less than the low-probability threshold are considered background areas; and regions with alteration probabilities between the low-probability and high-probability thresholds are considered transition areas. This process yields initial alteration mineral distribution information for the target region. The high-probability and low-probability thresholds are set based on the discriminative power of dynamic spectral features, combined with the distribution characteristics of the Softmax probability. The high-probability threshold can be set to 0.6, and the low-probability threshold can be set to 0.4.

[0044] S3. Construct a spatial permeability field based on the geological structure data of the target area, and determine the hydrothermal fluid migration path information based on the spatial permeability field and the preset pressure boundary conditions.

[0045] In some implementations, fault distribution information and fracture development zone distribution information are extracted from geological structural data. Geological structural data includes high-resolution satellite imagery and digital elevation model (DEM) data of the target area. Specifically, linear features are extracted from 0.5m resolution satellite imagery using the Canny edge detection algorithm. The normalized difference vegetation index (NDVI) is then used to remove pseudo-linear features in vegetated areas, yielding the spatial distribution of fault zones. Based on 10m resolution DEM data, topographic curvature and slope are calculated, and areas with abrupt slope changes and negative curvature are extracted as fracture development zones. The extracted results are then manually verified and corrected using regional geological maps to obtain the fault distribution information and fracture development zone distribution information for the target area.

[0046] Based on the degree of geological fragmentation in the target area, laboratory rock permeability test data, and regional geological literature, permeability reference values ​​were established for fractured structural areas, fissure-developed areas, and surrounding rock areas. The permeability reference value for fractured structural areas was higher than that for fissure-developed areas, and vice versa. For example, the typical permeability value for fractured rock in the reference laboratory was set to 1 × 10⁻⁶. -12 m 2 The typical permeability value of the reference altered fractured rock in the fractured development zone is set to 1×10⁻⁶. -14 m 2 The typical permeability value of the surrounding rock zone, referencing dense bedrock, is set to 1×10⁻⁶. -16 m 2 Meanwhile, the reference values ​​are calibrated through sensitivity analysis of fluid migration simulation to ensure that the permeability values ​​of different structures remain significantly different, which is consistent with the geological law that hydrothermal fluids preferentially migrate along high-permeability structures.

[0047] Using the distribution information of fracture structures and fissure development zones as spatial constraints, a pre-defined spatial interpolation method (such as Kriging interpolation) is employed to interpolate discrete permeability reference values ​​to a grid matching the spatial resolution of the time-series hyperspectral data. During the interpolation process, extracted structural features are used as control points, and interpolation parameters, such as a variogram model, are adjusted and sensitivity analysis is performed to ensure that the interpolation results reflect the control of tectonic structures on fluid migration, forming a continuous spatial permeability field. The dimension of the spatial permeability field is consistent with the spatial dimension of the hyperspectral data, with each pixel corresponding to an equivalent permeability value characterizing the fluid permeability of the subsurface medium.

[0048] Furthermore, pressure source points and pressure sink points are set on the spatial permeability field to form preset pressure boundary conditions. Among them, the pressure source points are located at key structural locations with active hydrothermal activity and excellent fluid recharge conditions, such as fault intersections or rock mass contact zones, which conform to the origin and enrichment patterns of natural hydrothermal fluids. The pressure sink points are located at the boundary of the target area, serving as the conventional discharge boundary after the migration of hydrothermal fluids, thus constructing a spatial pattern consisting of internal recharge and external discharge.

[0049] A fixed initial pressure value is assigned to the deployed pressure source points, and a constant boundary pressure value is set for the pressure sink points at the boundary of the target area. The global pressure constraint configuration is completed through the fixed-point pressure constraint and the boundary pressure constraint. For example, the initial pressure of the pressure source points is set to 10 MPa and the boundary pressure of the pressure sink points is set to 0.1 MPa.

[0050] Constrained by preset pressure boundary conditions and a spatial permeability field, a globally continuous gridded pressure field is first obtained, with each pixel possessing an independent pressure value. The pressure difference between the current pixel and its right-hand neighbor, divided by the horizontal grid step size, yields the pressure change per unit horizontal distance; a positive difference indicates increasing horizontal pressure, while a negative difference indicates decreasing horizontal pressure. Similarly, the pressure difference between the current pixel and its upper-hand neighbor, divided by the vertical grid step size, yields the pressure change per unit vertical distance; positive and negative values ​​directly reflect the vertical pressure fluctuations. Finally, the pressure change components in the two orthogonal directions are combined into a vector form to obtain the pressure gradient vector for each pixel, fully representing the pressure trend in two-dimensional space. The larger the absolute value of the component, the more drastic the pressure change in the corresponding direction.

[0051] Simultaneously set preset fluid viscosity and fluid density parameters, such as setting the fluid viscosity to 1×10⁻⁶. -3 Pa・s.

[0052] Using Darcy's law, which describes fluid transport in porous media, and based on the spatial permeability field, preset fluid viscosity parameters, and preset pressure boundary conditions, the fluid velocity vector field is determined as follows: In the formula, The vector representing the fluid velocity of a pixel, in m / s, includes both the magnitude and direction of the velocity, describing the flow state of the fluid at a certain location in space; M represents the equivalent permeability value of a pixel in the spatial permeability field, in m³. 2 Characterizes the ability of a medium to allow fluid to pass through; This represents the preset fluid viscosity parameter, in Pa·s, which characterizes the internal frictional resistance during fluid flow. This represents the pressure gradient vector of a pixel.

[0053] It is a flow rate adjustment factor, determined by both permeability M and fluid viscosity μ: permeability M is positively correlated with flow rate, the higher the permeability of the medium, the easier it is for the fluid to pass through, and the greater the flow rate; fluid viscosity μ is negatively correlated with flow rate, the greater the fluid viscosity, the stronger the flow resistance, and the lower the flow rate; the negative sign only adjusts the direction of flow rate and does not affect the magnitude of the flow rate.

[0054] In the fluid velocity vector field, the velocity of each pixel is positively correlated with the permeability and pressure gradient at that location, with the flow direction pointing from high-pressure areas to low-pressure areas. The formula calculates the fluid velocity vector field at each pixel location within the target area, with each pixel corresponding to a velocity vector, fully describing the flow state of hydrothermal fluids in the underground medium.

[0055] The set of pixel locations in the fluid velocity vector field where the velocity exceeds a preset velocity threshold is identified as the hydrothermal fluid migration path information. The preset velocity threshold is set based on the geological patterns of hydrothermal fluid migration, the statistical characteristics of velocity in the global simulation, and the numerical range of seepage in porous media. For example, the preset velocity threshold is set to 1×10⁻⁶. -8 m / s. The set of pixel locations in the fluid velocity vector field where the velocity is greater than a preset velocity threshold is determined as the hydrothermal fluid migration path information. The region where the velocity is greater than the threshold is the main fluid migration channel. The path extends along high-permeability structures, and the pressure gradient gradually decreases from the source point to the sink point. This is how hydrothermal fluid migration paths that conform to geological mineralization laws are obtained.

[0056] S4. The initial alteration mineral distribution information and hydrothermal fluid migration path information are corrected by spatial topological constraints to obtain the hydrothermal alteration zone delineation results of the target area.

[0057] In some implementations, a first probability threshold, such as 0.6, is set based on the significant boundary between the probability distributions of the two types of pixels, according to the alteration and background differentiation of dynamic spectral characteristics. A second probability threshold, such as 0.5, is set slightly lower than the first probability threshold to accommodate weak alteration signals caused by vegetation / weathering layer obstruction. Furthermore, a preset isolation threshold is set based on the neighborhood size and the continuity requirement of patchy alteration. Taking a 3×3 pixel neighborhood as an example, the threshold is set to 2. When the number of pixels in the neighborhood is less than the preset isolation threshold, it indicates that the pixel lacks continuous alteration signal support.

[0058] This method identifies high-probability non-path pixels in the initial alteration mineral distribution information whose alteration probability is higher than a first probability threshold but whose spatial location does not fall within the coverage area of ​​hydrothermal fluid migration path information. It then performs isolation determination and processing on the alteration probability of these high-probability non-path pixels based on pixels within a preset spatial neighborhood (e.g., a 3×3 pixel neighborhood). Specifically, it counts the number of pixels within the preset spatial neighborhood of a high-probability non-path pixel whose alteration probability is higher than the first probability threshold. If the number of pixels is less than the preset isolation determination threshold, the high-probability non-path pixel is determined as an isolated anomalous pixel. Such pixels are often misjudged due to background interference, and their alteration probability needs to be adjusted to a specified value below the first probability threshold, such as 0.3. If the number of pixels is greater than or equal to the preset isolation determination threshold, it indicates the presence of a large-scale alteration signal in the area, and the alteration probability of the high-probability non-path pixel remains unchanged. This process eliminates isolated anomalies that do not conform to fluid migration patterns, reducing the impact of background interference on alteration zone identification.

[0059] Identify suspected occluded pixels within the coverage area of ​​hydrothermal fluid migration paths that have an alteration probability lower than a second probability threshold. These pixels often have a low alteration probability in dynamic spectral identification due to occlusion by vegetation or weathering layers, but altered minerals actually exist. The alteration probability of the suspected occluded pixels is supplemented based on pixels within a preset spatial neighborhood (e.g., a 5×5 pixel neighborhood, increasing the number of references). Specifically, this involves filtering valid pixels within the preset spatial neighborhood of the suspected occluded pixel whose alteration probability is higher than a first probability threshold. Interpolation weights are determined based on the spatial distance between valid pixels and suspected occluded pixels. For example, an inverse distance-weighted interpolation method can be used, where closer valid pixels have higher weights, or an exponentially decaying distance weight function can be used, such as a weight of e. -d / c Where d is the spatial distance and c is a preset distance control constant (e.g., 10m, used to adjust the weight decay rate). The erosion probability of effective pixels is weighted according to the interpolation weights to obtain the erosion probability after the suspected occluded pixel completion processing.

[0060] It's important to understand that the completion process targets suspected occlusion pixels along the fluid migration path. These pixels have weakened spectral signals due to vegetation or weathering layers, resulting in a low initial alteration probability. However, geologically, they are located in hydrothermal activity channels and should contain alteration minerals. Therefore, a completion threshold higher than the first probability threshold (e.g., 0.8) needs to be set. If the completed alteration probability is greater than or equal to the completion threshold, the completed alteration probability is used to restore the weak signal to a valid alteration signal. If the completed alteration probability is less than the completion threshold, the following further operations are performed: First, expand the interpolation neighborhood (e.g., adjust the original 5×5 pixel neighborhood to a 7×7 pixel neighborhood), re-select the effective pixels within the neighborhood, and recalculate the completion probability to supplement more samples and improve the reliability of the completion result.

[0061] If the alteration probability after expanding the neighborhood is still between the second probability threshold and the completion threshold, the pixel is marked as a weak alteration signal area, and its alteration probability remains unchanged. When delineating alteration zones in the future, if the pixel is continuous with the surrounding high-probability alteration areas, it is included in the edge of the alteration zone; if it is an isolated weak signal, it is removed. This balances the completion effect with geological accuracy and avoids misjudgment due to excessive inflation. If the alteration probability after completion is still lower than the second probability threshold, it means that although the location is on the fluid path, there is insufficient alteration signal support in the neighborhood, and the actual alteration potential is extremely low. It is directly judged as a background pixel, and its alteration probability is adjusted to a specified value lower than the first probability threshold, such as 0.3. It is not included in the alteration zone, thus avoiding misjudging areas without actual alteration as alteration zones and ensuring the accuracy of alteration zone delineation.

[0062] After completion, a preset smoothing algorithm (such as 3×3 Gaussian filtering) is used to smooth the overall alteration probability distribution, eliminate local noise, and ensure the continuity of the alteration probability distribution, thereby completing the occluded alteration area and restoring the true alteration signal distribution.

[0063] The boundaries of continuous regions with alteration probabilities higher than a first probability threshold are extracted. A pre-defined contour extraction algorithm (e.g., the Suzuki contour extraction algorithm) is used to extract the outer and inner boundaries of the altered regions. Then, a pre-defined smoothing algorithm (e.g., B-spline curve fitting) is used to smooth the extracted boundaries, eliminating jagged noise and obtaining continuous, smooth hydrothermal alteration zone spatial boundaries, thus achieving the hydrothermal alteration zone delineation result. Finally, a vector boundary dataset of the hydrothermal alteration zone is output, including parameters such as the alteration zone's direction, length, and area, completing the delineation of the hydrothermal alteration zone and obtaining continuous hydrothermal alteration zone identification results that conform to geological mineralization patterns.

[0064] Based on the above technical solution, temporal spectral information including pre-excitation, excitation transient, and relaxation recovery stages is obtained through active energy excitation. Feature parameters reflecting the dynamic response of minerals are extracted to generate the initial alteration mineral distribution. At the same time, a spatial permeability field is constructed based on geological structural data and the hydrothermal fluid migration path is simulated. Then, the initial alteration mineral distribution and fluid migration path are spatially topologically constrained and corrected. This makes the alteration identification process not only dependent on static spectral features, but also controlled by the physical migration law of hydrothermal activity. Thus, under complex surface conditions, it effectively suppresses interference from homospectral foreign objects, eliminates isolated anomalous patches, and completes the alteration information of obscured areas. Finally, it obtains spatially continuous and geologically consistent hydrothermal alteration zone delineation results, which significantly improves the accuracy and reliability of alteration zone identification and delineation.

[0065] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0066] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0067] In this embodiment of the invention, the hydrothermal alteration zone identification and delineation device for hyperspectral remote sensing image analysis can be divided into functional units according to the above method example. For example, each function can be divided into its own functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or software. It should be noted that the unit division in this embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0068] Although the invention has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings and the disclosure, will understand and implement other variations of the disclosed embodiments in carrying out the claimed invention. In this invention, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several of the functions listed in this invention.

[0069] Although the invention has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made therein without departing from the scope of the invention. Accordingly, this specification and drawings are merely illustrative of the invention and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Thus, if such modifications and modifications of the invention fall within the scope of the invention and its equivalents, the invention is also intended to include such modifications and modifications.

Claims

1. A method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis, characterized in that, include: During the process of transmitting a preset energy pulse to the target area, a temporal hyperspectral image of the target area is acquired; The time-series hyperspectral images include pre-excitation static spectral frames, excitation transient spectral frames, and relaxation recovery phase spectral frames. Dynamic spectral feature parameters are extracted from the time-series hyperspectral images, and initial alteration mineral distribution information of the target region is generated based on the dynamic spectral feature parameters; Fault structure distribution information and fracture development zone distribution information are extracted from the geological structure data of the target area. Permeability reference values ​​are set for the fault structure area, fracture development area, and surrounding rock area respectively. The permeability reference value of the fault structure area is greater than that of the fracture development area, and the permeability reference value of the fracture development area is greater than that of the surrounding rock area. Using the fault structure distribution information and the fracture development area distribution information as spatial constraints, a preset spatial interpolation method is used to interpolate the discrete permeability reference values ​​to a grid that matches the spatial resolution of the time-series hyperspectral image, forming a continuous spatial permeability field. Pressure source points and pressure sink points are set on the spatial permeability field to form preset pressure boundary conditions; the pressure source points are located at fracture intersections or rock mass contact zones, and the pressure sink points are located at the boundary of the target area; by using Darcy's law to describe fluid transport in porous media, a fluid velocity vector field is determined based on the spatial permeability field, preset fluid viscosity parameters, and the preset pressure boundary conditions; the set of pixel locations in the fluid velocity vector field with velocities greater than a preset velocity threshold is determined as hydrothermal fluid transport path information; Identify high-probability non-path pixels in the initial alteration mineral distribution information whose alteration probability is higher than the first probability threshold but whose spatial location does not fall within the coverage area of ​​the hydrothermal fluid migration path information, and perform isolation determination and processing on the alteration probability of the high-probability non-path pixels based on the pixels in the preset spatial neighborhood of the high-probability non-path pixels. The isolation determination and processing of the erosion probability of the high-probability non-path pixel is performed based on the pixels in the preset spatial neighborhood of the high-probability non-path pixel, including: counting the number of pixels in the preset spatial neighborhood of the high-probability non-path pixel whose erosion probability is higher than the first probability threshold; if the number of pixels is less than the preset isolation determination threshold, the high-probability non-path pixel is determined as an isolated anomalous pixel, and the erosion probability of the isolated anomalous pixel is adjusted to a specified value lower than the first probability threshold; if the number of pixels is greater than or equal to the preset isolation determination threshold, the erosion probability of the high-probability non-path pixel remains unchanged. Identify suspected occlusion pixels within the coverage area of ​​the hydrothermal fluid migration path information that have an alteration probability lower than a second probability threshold, and perform alteration probability completion processing on the suspected occlusion pixels based on pixels in a preset spatial neighborhood of the suspected occlusion pixels; the alteration probability completion processing on the suspected occlusion pixels based on pixels in a preset spatial neighborhood of the suspected occlusion pixels includes: filtering valid pixels in the preset spatial neighborhood of the suspected occlusion pixels whose alteration probability is higher than the first probability threshold; determining interpolation weights based on the spatial distance between the valid pixels and the suspected occlusion pixels; and weighting the alteration probability of the valid pixels based on the interpolation weights to obtain the alteration probability of the suspected occlusion pixels after completion processing; Extract the boundaries of continuous regions whose alteration probability is higher than the first probability threshold after processing to obtain the hydrothermal alteration zone delineation result of the target region.

2. The method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis according to claim 1, characterized in that, The dynamic spectral feature parameters include: absorption peak shift, which characterizes the degree of change in absorption peak position before and after stimulation; absorption depth change rate, which characterizes the relative change in absorption depth before and after stimulation; and relaxation recovery time, which characterizes the time required for the spectrum to recover to a specified proportion of the resting state after stimulation. The dynamic spectral feature parameters extracted from the time-series hyperspectral image include: For each pixel in the time-series hyperspectral image, the average spectrum of the resting state before stimulation is extracted from the static spectral frame before stimulation, and the peak spectrum of the transient state before stimulation is extracted from the transient spectral frame after stimulation. Determine the absorption peak positions and absorption peak depths of the pre-excitation resting state average spectrum and the excitation transient peak spectrum within the preset characteristic absorption band of the altered mineral; The absorption peak shift is determined based on the absorption peak position of the average spectrum in the resting state before excitation and the absorption peak position of the peak spectrum in the transient state after excitation. The rate of change of absorption depth is determined based on the absorption peak depth of the average spectrum in the resting state before stimulation and the absorption peak depth of the peak spectrum in the transient state after stimulation. The relaxation recovery time is determined based on the recovery process of each frame of the relaxation spectrum after stimulation relative to the average spectrum of the resting state before stimulation.

3. The method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis according to claim 2, characterized in that, Initial alteration mineral distribution information for the target region is generated based on the dynamic spectral characteristic parameters, including: A dynamic spectral feature vector is constructed based on the absorption peak shift, the absorption depth change rate, and the relaxation recovery time. The dynamic spectral feature vector is matched with the standard dynamic features of known alteration mineral categories to obtain the probability distribution of each pixel in the target area belonging to each known alteration mineral. Based on the probability distribution, an alteration probability characterizing the likelihood that each pixel belongs to a hydrothermal alteration region is determined, which serves as the initial alteration mineral distribution information.

4. The method for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis according to claim 3, characterized in that, The dynamic spectral feature vector is matched with the standard dynamic features of known alteration mineral categories to obtain the probability distribution of each pixel in the target region belonging to each known alteration mineral, including: The standard feature vector formed by the dynamic spectral feature vector and the standard dynamic features of known alteration mineral categories is normalized. Analyze the similarity measure between the normalized dynamic spectral eigenvector and the normalized standard eigenvector; Based on the similarity metric, the probability distribution of each known alteration mineral category corresponding to each pixel is determined by a preset normalization exponential function.

5. A device for identifying and delineating hydrothermal alteration zones in hyperspectral remote sensing image analysis, characterized in that, include: The spectral image acquisition unit is used to acquire a temporal hyperspectral image of the target area during the process of emitting a preset energy pulse toward the target area; The time-series hyperspectral images include pre-excitation static spectral frames, excitation transient spectral frames, and relaxation recovery phase spectral frames. The dynamic feature extraction unit is used to extract dynamic spectral feature parameters from the time-series hyperspectral image and generate initial alteration mineral distribution information of the target area based on the dynamic spectral feature parameters. The fluid path determination unit is used to extract information on the distribution of fracture structures and the distribution of fracture development zones from the geological structural data of the target area. Permeability reference values ​​are set for fracture structure regions, fissure development regions, and surrounding rock regions, respectively. The permeability reference value for the fracture structure region is greater than that for the fissure development region, and the permeability reference value for the fissure development region is greater than that for the surrounding rock region. Using the distribution information of the fracture structure and the distribution information of the fissure development region as spatial constraints, a preset spatial interpolation method is used to interpolate the discrete permeability reference values ​​to a grid that matches the spatial resolution of the time-series hyperspectral image, forming a continuous spatial permeability field. Pressure source points and pressure sink points are set on the spatial permeability field to form preset pressure boundary conditions. The pressure source points are located at the fracture intersection or rock mass contact zone, and the pressure sink points are located at the boundary of the target region. Based on Darcy's law describing fluid transport in porous media, the fluid velocity vector field is determined according to the spatial permeability field, preset fluid viscosity parameters, and preset pressure boundary conditions. The set of pixel positions in the fluid velocity vector field with velocities greater than a preset velocity threshold is determined as hydrothermal fluid transport path information. The spatial constraint correction unit is used to identify high-probability non-path pixels in the initial alteration mineral distribution information whose alteration probability is higher than the first probability threshold, but whose spatial location does not fall within the coverage area of ​​the hydrothermal fluid migration path information, and to perform isolation determination and processing on the alteration probability of the high-probability non-path pixels based on the pixels in the preset spatial neighborhood of the high-probability non-path pixels. Based on the pixels within a preset spatial neighborhood of the high-probability non-path pixels, an isolation determination and processing of the erosion probability of the high-probability non-path pixels is performed, including: counting the number of pixels within the preset spatial neighborhood of the high-probability non-path pixels whose erosion probability is higher than a first probability threshold; if the number of pixels is less than a preset isolation determination threshold, the high-probability non-path pixels are determined as isolated anomalous pixels, and the erosion probability of the isolated anomalous pixels is adjusted to a specified value lower than the first probability threshold; if the number of pixels is greater than or equal to the preset isolation determination threshold, the erosion probability of the high-probability non-path pixels remains unchanged; identifying suspected occlusion pixels within the coverage area of ​​the hydrothermal fluid migration path information whose erosion probability is lower than a second probability threshold. The alteration probability of the suspected occluded pixel is completed based on pixels within a preset spatial neighborhood of the suspected occluded pixel. This completion process includes: filtering valid pixels within the preset spatial neighborhood of the suspected occluded pixel whose alteration probability is higher than a first probability threshold; determining interpolation weights based on the spatial distance between the valid pixels and the suspected occluded pixel; weighting the alteration probability of the valid pixels according to the interpolation weights to obtain the alteration probability after completion; and extracting the boundaries of continuous regions whose processed alteration probability is higher than the first probability threshold to obtain the hydrothermal alteration zone delineation result of the target region.

Images