Rapid Delineation Method of Mineralization Zones Based on Remote Sensing Technology

Abstract

This application provides a rapid delineation method for mineralization zones based on remote sensing technology, relating to the field of image recognition technology. The method includes: determining the characteristic bands of the target mineral based on standard spectral data of the target mineral and interfering spectral data of interfering substances within the area to be delineated; constructing a first evaluation model based on the actual mineral content of a calibration sample and its spectral values ​​at the characteristic bands to obtain a first mineral content estimate for each pixel in the remote sensing image data of the area to be delineated; determining the associated substance bands based on the correlation coefficients between each band and the characteristic band at the target pixel; constructing a second evaluation model based on the actual mineral content of the calibration sample and its spectral values ​​at the associated substance bands to obtain a second mineral content estimate; determining the extensibility based on the first mineral content estimate at neighboring pixels, the gradient vector per unit mineral content, and the unit direction vector, dynamically adjusting the proportion of the mineral content estimate, and delineating the mineralization zone accordingly. This application improves the accuracy of mineralization zone delineation.

Description

Technical Field

[0001] This application relates to the field of image recognition technology, specifically to a method for rapid delineation of mineralized zones based on remote sensing technology. Background Technology

[0002] As mineral resource exploration and development gradually advances into deeper and more concealed areas, the demand for efficient and accurate mineralization zone identification and delineation technologies is becoming increasingly urgent. Remote sensing technology, with its significant advantages such as wide coverage, high efficiency, and relatively low cost, has become an indispensable key technology in the fields of geological and mineral resource surveys and mineralization zone prediction.

[0003] In related technologies, the delineation of mineralization zones based on remote sensing methods mainly relies on the identification of surface or shallow information. However, due to the limited penetration capability of optical remote sensing and the obstruction by vegetation or overburden, it is difficult to accurately capture the true distribution characteristics of deep or concealed ore bodies. This leads to problems such as insufficient identification depth, inaccurate regional prediction, and unclear spatial continuity of mineralization zones during the delineation process, resulting in missed judgments, misjudgments, or large deviations in the delineation range. Summary of the Invention

[0004] In view of this, this application proposes a method for rapid delineation of mineralized zones based on remote sensing technology to solve the problems in related technologies, such as difficulty in accurately capturing the true distribution characteristics of deep or concealed ore bodies, which leads to insufficient identification depth, inaccurate regional prediction, and unclear spatial continuity of mineralized zones during the delineation process.

[0005] According to the first aspect of this application, a method for rapid delineation of mineralization zones based on remote sensing technology is provided, and the specific technical solution adopted is as follows:

[0006] Acquire remote sensing image data of the area to be delineated and perform preprocessing;

[0007] Based on the standard spectral data of the target mineral within the area to be delineated and the interference spectral data of the interfering substances, the mineral characterization coefficients of each band in the remote sensing image data are calculated, and the characteristic bands of the target mineral are determined based on the mineral characterization coefficients.

[0008] A first evaluation model is constructed based on the actual mineral content of the calibration samples and their spectral values ​​at the characteristic bands, and the first mineral content estimate of each pixel in the remote sensing image data is determined based on the first evaluation model; the calibration samples are the points in the area to be delineated where the actual mineral content of the target mineral is known.

[0009] Pixels whose estimated content of the first mineral is greater than a first preset threshold are identified as target pixels, and associated material bands are determined based on the correlation coefficients between each band at the target pixel and the feature bands.

[0010] A second evaluation model is constructed based on the actual mineral content of the calibration sample and its spectral value in the associated material band, and the second mineral content estimate of each pixel in the remote sensing image data is determined based on the second evaluation model.

[0011] The ductility of the target mineral is determined based on the first mineral content estimate, the unit mineral content gradient vector, and the unit direction vector at the neighboring pixels of each pixel. The proportion of the first mineral content estimate and the second mineral content estimate is dynamically adjusted based on the ductility to obtain the final mineral content estimate of each pixel.

[0012] Based on the final mineral content estimate, the mineralization zone of the target mineral is delineated in the area to be delineated.

[0013] For example, the step of calculating the mineral characterization coefficient of each band in the remote sensing image data based on the standard spectral data of the target mineral and the interference spectral data of the interfering substances in the area to be delineated includes: acquiring the spectral reflectance data of the pure mineral or standard mineral sample corresponding to the target mineral measured under laboratory conditions, and recording it as the standard spectral data; determining the distribution location of the interfering substances in the area to be delineated based on the remote sensing image data, and collecting the interfering substances at the distribution location as measurement samples, acquiring the spectral reflectance data of the measurement samples, and recording it as the interference spectral data; for each band in the remote sensing image data, calculating the absolute value of the difference between the current band and the band where each absorption peak is located in each of the interference spectral data, and recording the minimum value as the first minimum band distance between the current band and the corresponding interference spectral data; calculating the absolute value of the difference between the current band and the band where each absorption peak is located in the standard spectral data, and recording the minimum value as the second minimum band distance between the current band and the standard spectral data; and determining the mineral characterization coefficient of the current band based on the sum of the first minimum band distances and the second minimum band distance.

[0014] For example, determining the characteristic band of the target mineral based on the mineral characterization coefficient includes: constructing a wavelength-mineral characterization coefficient curve based on data pairs of each current band and the mineral characterization coefficient; obtaining each first local maximum point in the wavelength-mineral characterization coefficient curve; and determining the current band corresponding to the first local maximum point as the characteristic band.

[0015] For example, the construction of the first evaluation model based on the actual mineral content of the calibration sample and its spectral value at the characteristic band includes: extracting the spectral value of each calibration sample at the characteristic band and performing standardization processing to obtain the characteristic band spectral value; obtaining the actual mineral content of the calibration sample and performing standardization processing; and constructing the first evaluation model based on partial least squares regression, using the characteristic band spectral value of the calibration sample as the independent variable and the actual mineral content of the calibration sample as the dependent variable.

[0016] For example, determining the associated material band based on the correlation coefficient between each band at the target pixel and the feature band includes: extracting the full-band spectral value of the target pixel in the remote sensing image data; determining the associated material band probability of each non-feature band based on the spectral values ​​of the target pixel in the full-band non-feature band and the feature band; constructing a wavelength-associated material band probability curve based on the data pair of the non-feature band and the associated material band probability; obtaining each second local maximum point in the wavelength-associated material band probability curve, and determining the non-feature band corresponding to the second local maximum point as the associated material band.

[0017] For example, determining the probability of the associated material band in each of the non-feature bands based on the spectral values ​​of the target pixel in the non-feature bands and the feature bands of the entire band includes: for each of the non-feature bands, determining a first spectral value sequence of the target pixel in the current non-feature band; obtaining a second spectral value sequence of the target pixel in each of the feature bands; calculating the Pearson correlation coefficient between the first spectral value sequence of the current non-feature band and each of the second spectral value sequences, and calculating the average value of the Pearson correlation coefficient, which is denoted as the probability of the associated material band in the current non-feature band.

[0018] For example, the step of constructing a second evaluation model based on the actual mineral content of the calibration sample and its spectral value in the associated material band includes: extracting the spectral value of each calibration sample in the associated material band and performing standardization processing to obtain the spectral value of the associated material band; obtaining the actual mineral content of the calibration sample and performing standardization processing; and constructing the second evaluation model based on partial least squares regression, using the spectral value of the associated material band of the calibration sample as the independent variable and the actual mineral content of the calibration sample as the dependent variable.

[0019] For example, determining the extensibility of the target mineral based on the first mineral content estimate, mineral content gradient vector, and direction vector at neighboring pixels of each pixel includes: for each pixel in the remote sensing image data, determining the neighboring pixels of the current pixel based on a preset distance threshold; extracting the spectral values ​​of the neighboring pixels at the feature band and performing standardization processing, and inputting the standardized spectral values ​​into the first evaluation model to obtain the first mineral content estimate of the neighboring pixels; determining the direction of the fastest increase in the target mineral content of the neighboring pixels in the remote sensing image data plane as the direction of the unit mineral content gradient vector, determining the direction from the neighboring pixels to the current pixel as the direction of the unit direction vector, and calculating the cosine similarity between the unit mineral content gradient vector and the unit direction vector; and determining the extensibility of the target mineral at the current pixel based on the first mineral content estimate and the cosine similarity corresponding to each of the neighboring pixels.

[0020] For example, the step of dynamically adjusting the proportions of the first mineral content estimate and the second mineral content estimate based on the extensibility to obtain the final mineral content estimate for each pixel includes: normalizing the extensibility of the target mineral at the current pixel to obtain a weight adjustment factor; using the difference between 1 and the weight adjustment factor as a first weight, and determining a first final mineral content factor based on the first weight and the first mineral content estimate of the current pixel; using the weight adjustment factor as a second weight, and determining a second final mineral content factor based on the second weight and the second mineral content estimate of the current pixel; and determining the final mineral content estimate at the current pixel based on the first final mineral content factor and the second final mineral content factor.

[0021] For example, the step of delineating the mineralization zone of the target mineral in the area to be delineated based on the final mineral content estimate includes: determining the pixels whose final mineral content estimate is greater than a second preset threshold as delineated pixels, and determining the connected region formed by the delineated pixels to obtain the mineralization zone of the target mineral.

[0022] This application may have some or all of the following beneficial effects:

[0023] In the rapid delineation method for mineralized zones based on remote sensing technology provided in this application, the mineral characterization coefficients of each band in the remote sensing image data are calculated using the standard spectral data of the target mineral and the interfering spectral data of interfering substances within the area to be delineated. Based on these mineral characterization coefficients, the characteristic bands of the target mineral are determined. This allows for the selection of bands that highly match the spectral absorption peaks of the target mineral and are less affected by interfering substances, reducing spectral interference from non-mineralized materials. Furthermore, the bands of associated substances are determined by the correlation coefficients between each band and the characteristic bands at the target pixel, fully utilizing the genetic correlation and spatial coupling between the target mineral and associated rock strata. This enhances the detection of concealed ore bodies and mineralized zones in thick-covered areas. Furthermore, this application can construct a first evaluation model and a second evaluation model by calibrating the actual mineral content of the sample and its spectral values ​​at the characteristic band and the associated material band. The extensibility of the target mineral can be determined by the first mineral content estimate, the gradient vector of the unit mineral content, and the unit direction vector at the neighboring pixels of each pixel. Thus, the weights of the first and second mineral content estimates can be dynamically adjusted by combining the spatial extensibility analysis of the target mineral. By integrating the information of the characteristic band of the target mineral and the associated material band, the mineral content estimation results are effectively optimized, the probability of missed or misjudged cases is reduced, and the spatial accuracy and quantitative precision of mineralization zone delineation are greatly improved.

[0024] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0025] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart is shown of a method for rapid delineation of mineralization zones based on remote sensing technology according to an exemplary embodiment of this application;

[0027] Figure 2 A contour map showing the spatial distribution of mineral content in a method for rapid delineation of mineralized zones based on remote sensing technology according to an exemplary embodiment of this application is shown. Detailed Implementation

[0028] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive objective, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the method for rapid delineation of mineralized zones based on remote sensing technology proposed in this application. 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.

[0029] 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 application pertains.

[0030] The specific scheme of the rapid delineation method for mineralization zones based on remote sensing technology provided in this application is described in detail below with reference to the accompanying drawings.

[0031] Please see Figure 1 It illustrates a flowchart of a method for rapid delineation of mineralized zones based on remote sensing technology according to an embodiment of this application, as shown below. Figure 1 As shown, the method for rapid delineation of mineralization zones based on remote sensing technology specifically includes the following steps:

[0032] S110: Acquire remote sensing image data of the area to be delineated and perform preprocessing;

[0033] S120: Calculate the mineral characterization coefficients of each band in the remote sensing image data based on the standard spectral data of the target mineral and the interference spectral data of the interfering substances within the area to be delineated, and determine the characteristic bands of the target mineral based on the mineral characterization coefficients.

[0034] S130: Construct a first evaluation model based on the actual mineral content of the calibration samples and their spectral values ​​at characteristic bands, and determine the first mineral content estimate of each pixel in the remote sensing image data based on the first evaluation model; the calibration samples are the points in the area to be delineated where the actual mineral content of the known target minerals is known.

[0035] S140: Determine the pixel point whose estimated content of the first mineral is greater than the first preset threshold as the target pixel point, and determine the associated material band based on the correlation coefficient between each band and the feature band at the target pixel point;

[0036] S150: Construct a second evaluation model based on the actual mineral content of the calibration sample and its spectral value in the associated material band, and determine the second mineral content estimate of each pixel in the remote sensing image data based on the second evaluation model;

[0037] S160: Determine the extensibility of the target mineral based on the first mineral content estimate, the unit mineral content gradient vector, and the unit direction vector at the neighboring pixels of each pixel, and dynamically adjust the proportion of the first mineral content estimate and the second mineral content estimate based on the extensibility to obtain the final mineral content estimate of each pixel.

[0038] S170: Delineate the mineralization zone of the target mineral in the area to be delineated based on the final mineral content estimate.

[0039] The following is a detailed explanation of each step in the above-mentioned method for rapid delineation of mineralized zones based on remote sensing technology:

[0040] In step S110, remote sensing image data of the area to be delineated is acquired and preprocessed.

[0041] In the embodiments of this application, the aforementioned area to be delineated refers to the target geographical area where mineralized zones need to be identified and delineated in mineral exploration. It is usually an area with favorable mineralization geological conditions, such as the area around known mineral occurrences, along tectonic zones, alteration anomaly zones, etc.

[0042] In this embodiment, the aforementioned remote sensing image data is image data of the area to be delineated, formed by sensors on remote sensing platforms such as satellites and drones recording electromagnetic wave information reflected from the Earth's surface. For example, this remote sensing image data can be selected from hyperspectral data of the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). Specifically, the parameters of the ASTER hyperspectral data are as follows: the spectral range mainly uses the visible-near-infrared-shortwave infrared bands, where the short-wave infrared (SWIR) band (1.0–2.5 μm) has significant absorption characteristics for metallic mineral-related alteration minerals (such as sericite, kaolinite, and chlorite); the spectral resolution of the ASTER hyperspectral image is ≤10 nm; the spatial resolution of the ASTER hyperspectral image is 15-30 m; the detection range covers the target mining area and the surrounding prediction area, with a single image ≥60 km × 60 km.

[0043] After acquiring remote sensing image data of the area to be delineated, this application requires preprocessing the acquired remote sensing image data to remove non-geological interference and enhance mineralization-related information. For example, this preprocessing operation may include geometric correction and radiometric and atmospheric correction. Geometric correction is used to eliminate image distortion caused by terrain undulations and sensor attitude deviations, ensuring accurate matching between the image and actual geographic coordinates. Specifically, the precision setting of geometric correction can be no greater than 1 pixel to ensure spatial consistency of the data. Radiometric and atmospheric correction is used to eliminate the scattering and absorption of electromagnetic waves by water vapor, aerosols, and ozone in the atmosphere, restoring the true spectral information of the Earth's surface. Specifically, radiometric and atmospheric correction can employ physical models such as Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) or Quick Atmospheric Correction (QUAC). It should be noted that the implementation principles of the above-mentioned geometric correction and radiometric and atmospheric correction preprocessing are the same as those in related technologies, and will not be elaborated further here.

[0044] In step S120, the mineral characterization coefficients of each band in the remote sensing image data are calculated based on the standard spectral data of the target mineral and the interference spectral data of the interfering substances in the area to be delineated, and the characteristic bands of the target mineral are determined based on the mineral characterization coefficients.

[0045] In the embodiments of this application, the target mineral refers to a key mineral in mineral exploration that can directly or indirectly indicate the existence of a specific mineralization zone. For example, the specific mineralization zone can be a copper ore, iron ore, gold ore, or other mineralization zone. The target mineral can be the main mineral of the ore body corresponding to the mineralization zone (e.g., chalcopyrite in copper ore, hematite in iron ore), or alteration minerals (such as sericite, kaolinite, chlorite, etc.) that are associated with the mineralization process of the specific mineralization zone. These alteration minerals are formed by mineralization and their distribution is highly correlated with the spatial location of the mineralization zone.

[0046] In this embodiment, the standard spectral data refers to the spectral reflectance data of the target mineral, either pure or standard, measured under laboratory conditions in the visible, near-infrared, and short-wave infrared bands. This data is used to provide the standard absorption peak characteristics of the target mineral, serving as a reference for subsequent screening of the target mineral's characteristic bands.

[0047] In this embodiment, the aforementioned interfering substance refers to a substance within the delineated area whose spectral characteristics overlap or are similar to the characteristic spectrum of the target mineral, or which may mask the spectral signal of the target mineral, thereby interfering with the accurate identification of the target mineral. Exemplarily, the interfering substance may include various non-mineralized materials or surrounding rocks such as vegetation, weathering crust, aquifers, carbonate rocks, and siliceous rocks.

[0048] In this embodiment of the application, the aforementioned interference spectral data is the spectral reflectance data of the aforementioned interfering substances obtained through field investigation and geological structural data interpretation, which is used to quantify the degree of interference of the interfering substances on the target mineral.

[0049] In the embodiments of this application, the aforementioned mineral characterization coefficient is used to measure the ability of a corresponding band in remote sensing image data to characterize the target mineral. Specifically, the closer a certain band is to the absorption peak of the target mineral and the farther it is from the absorption peak of interfering substances, the larger the mineral characterization coefficient of that band, and the better the identification effect of the target mineral.

[0050] For example, the above-mentioned calculation of the mineral characterization coefficient of each band in the remote sensing image data based on the standard spectral data of the target mineral and the interference spectral data of the interfering substances in the area to be delineated can be achieved as follows: The spectral reflectance data of the pure mineral or standard mineral sample corresponding to the target mineral, measured under laboratory conditions, is obtained and recorded as the standard spectral data; the distribution location of the interfering substances in the area to be delineated is determined based on the remote sensing image data, and the interfering substances at the distribution location are collected as measurement samples, and the spectral reflectance data of the measurement samples is obtained and recorded as the interference spectral data; for each band in the remote sensing image data, the absolute value of the difference between the current band and the band containing each absorption peak in each interference spectral data is calculated, and the minimum value is recorded as the first minimum band distance between the current band and the corresponding interference spectral data; the absolute value of the difference between the current band and the band containing each absorption peak in the standard spectral data is calculated, and the minimum value is recorded as the second minimum band distance between the current band and the standard spectral data; the mineral characterization coefficient of the current band is determined based on the sum of the first minimum band distances and the second minimum band distance.

[0051] Specifically, taking the current band as band λ as an example, the mineral characterization coefficient of band λ can be calculated using the following formula:

[0052]

[0053] Among them, S λ λ represents the mineral characterization coefficient of the current band λ; n represents the number of interfering substances in the region to be delineated; n z λ represents the number of absorption peaks in the interfering spectral data of the z-th interfering substance; λ is the current spectral band; λ z,lλ represents the band containing the l-th absorption peak in the interfering spectral data of the z-th interfering substance; n0 represents the number of absorption peaks in the standard spectral data of the target mineral; λ r The band containing the r-th absorption peak in the standard spectral data of the target mineral; it should be noted that adding 1 to the denominator in the above formula is to avoid the denominator being 0.

[0054] In the above formula, |λ-λ z,l | represents the absolute value of the difference between the current band λ and the band containing the l-th absorption peak in the interference spectrum data of the z-th interfering substance, used to indicate the distance between the current band λ and the band containing the l-th absorption peak in the interference spectrum data of the z-th interfering substance. This is the minimum distance between the current band λ and the band containing each absorption peak in the interference spectrum data of the z-th interfering substance (i.e., the first minimum band distance mentioned above); This is the sum of the distances of each of the first minimum bands. The smaller this value, the closer the current band λ is to the absorption peak of the interfering substance, meaning that the current band λ is more strongly interfered with by the interfering substance; |λ-λ r | represents the absolute value of the difference between the current band λ and the band containing the r-th absorption peak in the standard spectral data of the target mineral, used to indicate the distance between the current band λ and the band containing the r-th absorption peak in the standard spectral data of the target mineral; This is the minimum distance between the current spectral band λ and the spectral bands containing the absorption peaks of the target mineral in the standard spectral data (i.e., the second minimum spectral band distance mentioned above). The smaller this value, the closer the current spectral band λ is to the absorption peaks of the target mineral, and the stronger its characterization ability of the target mineral. In summary, S λ The larger the value, the farther the current band λ is from the absorption peak of the interfering substance and the closer it is to the absorption peak of the target mineral, indicating that the current band λ has a stronger characterization ability for the target mineral.

[0055] In this embodiment, the aforementioned characteristic bands are the bands that have the strongest response to the spectral characteristics of the target mineral and are least affected by interference, selected based on the aforementioned mineral characterization coefficients. They are usually the bands in remote sensing image data that best reflect the characteristics of the target mineral.

[0056] For example, the above-mentioned determination of the characteristic band of the target mineral based on the mineral characterization coefficient can be achieved as follows: construct a wavelength-mineral characterization coefficient curve based on the data pairs of each current band and the mineral characterization coefficient; obtain each first local maximum point in the wavelength-mineral characterization coefficient curve, and determine the current band corresponding to the first local maximum point as the characteristic band. Specifically, establish a coordinate system with wavelength as the horizontal axis and mineral characterization coefficient as the vertical axis; plot the wavelength-mineral characterization coefficient curve in the established coordinate system based on each band in the remote sensing image data and its corresponding mineral characterization coefficient; filter the local maximum points (i.e., the first local maximum points mentioned above) in the wavelength-mineral characterization coefficient curve, and determine the band corresponding to the first local maximum point as the characteristic band.

[0057] In step S130, a first evaluation model is constructed based on the actual mineral content of the calibration sample and its spectral value at the characteristic band, and the first mineral content estimate of each pixel in the remote sensing image data is determined based on the first evaluation model; the calibration sample is the location of the actual mineral content of the known target mineral in the area to be delineated.

[0058] In this embodiment of the application, the aforementioned calibration samples are points where the actual mineral content of the known target mineral is located within the area to be delineated. For example, the calibration samples can be selected from surface outcrop sampling points of known iron ore deposits within the area to be delineated, or from verification borehole locations from historical exploration.

[0059] In the embodiments of this application, the above-mentioned actual mineral content refers to the true content of the target mineral in the calibration sample obtained by laboratory analysis (such as quantitative chemical analysis) or field detection methods (such as core drilling tests).

[0060] In the embodiments of this application, the first evaluation model is a relational model constructed based on the actual mineral content of the calibration sample and its spectral values ​​at the characteristic band. It is used to establish the mapping relationship between the spectral values ​​of the characteristic band and the mineral content, and can convert the spectral signal of the characteristic band into the content estimate of the target mineral.

[0061] For example, the above-mentioned construction of the first evaluation model based on the actual mineral content of the calibration samples and their spectral values ​​at the characteristic bands can be achieved as follows: Extract the spectral values ​​of each calibration sample at the characteristic bands and perform standardization to obtain the characteristic band spectral values; obtain the actual mineral content of the calibration samples and perform standardization; based on partial least squares regression, use the characteristic band spectral values ​​of the calibration samples as independent variables and the actual mineral content of the calibration samples as dependent variables to construct the first evaluation model. In the above process, the mean-standard deviation standardization method can be used to standardize the spectral values ​​at the characteristic bands of the calibration samples and the actual mineral content. Through standardization, the spectral values ​​at the characteristic bands of the calibration samples and the actual mineral content can be mapped to a unified data scale, eliminating dimensional differences. The above-mentioned construction of the first evaluation model can use the standardized characteristic band spectral values ​​of the calibration samples as input and the standardized actual mineral content of the calibration samples as output. The quantitative relationship between the characteristic band spectral values ​​of the target mineral and the actual mineral content is established through partial least squares regression (PLS), resulting in the following first evaluation model:

[0062]

[0063] Among them, F i is the estimated content of the first mineral at the i-th pixel in the remote sensing image data; m is the number of characteristic bands of the target mineral; Let be the spectral value of the i-th pixel at the first characteristic band of the target mineral. Let be the spectral value of the i-th pixel at the second characteristic band of the target mineral. is the spectral value of the i-th pixel at the m-th characteristic band of the target mineral; F is the quantitative relationship determined based on the actual mineral content of the calibration sample and its spectral values ​​at each characteristic band of the target mineral.

[0064] In this embodiment of the application, after the first evaluation model is constructed by the above method, the spectral value of a certain pixel in the remote sensing image data at each characteristic band of the target mineral can be input into the first evaluation model, and the first mineral content estimate of the pixel can be output through the first evaluation model.

[0065] In step S140, the pixel points whose estimated first mineral content is greater than the first preset threshold are determined as target pixel points, and the associated material bands are determined based on the correlation coefficients between each band and the characteristic band at the target pixel point.

[0066] The process described above for determining the estimated content of the first mineral at each pixel only considers the differences in spectral reflectance between different bands, without taking into account the genetic relationship between the target mineral and the mineralization process, rock strata assemblage, and associated materials in the area to be delineated. When the detection depth is limited and concealed ore bodies cannot be directly imaged, this can easily lead to problems such as decreased accuracy in delineating mineralization zones and missed detection of concealed ore bodies. Therefore, in this embodiment, after obtaining the estimated content of the first mineral, spectral information of rock strata or materials that are genetically related to the target mineral and spatially connected or associated with it can be introduced to improve the ability to identify the target mineral. For example, in the exploration of concealed copper deposits, the spectral characteristics of altered minerals near 2.2 μm are usually relied upon, but vegetation or thick weathering layers weaken this signal. By using the 2.3-2.35 μm band of carbonate rock strata spatially coupled with the mineralization zone as an auxiliary feature band, the accuracy of mineralization zone delineation can be improved.

[0067] In this embodiment, the target pixels are those with high mineral content estimates selected from all pixels in the remote sensing image data by using a first preset threshold. Their spectral characteristics are more indicative of the distribution of mineralization zones and can serve as the basis for analyzing associated material bands, thus compensating for the insufficient depth of remote sensing detection. For example, the first preset threshold can be set based on actual scenario requirements. Specifically, the first preset threshold can be set to the top 25% of the first mineral content estimate. That is, determining the pixels with first mineral content estimates greater than the first preset threshold as target pixels can be achieved as follows: pixels in the remote sensing image data whose first mineral content estimates are in the top 25% are identified as target pixels.

[0068] In this embodiment, the aforementioned associated material refers to rock strata or substances that are genetically related to the target mineral within the area to be delineated and spatially connected or associated with it. Its spectral characteristics are highly coupled with the spatial distribution of the mineralization zone of the target mineral, which can help compensate for the limitations of remote sensing technology in terms of detection depth and improve the accuracy of identifying and delineating concealed or deep mineralization zones. For example, in the exploration of concealed copper deposits, the carbonate rock strata overlying the alteration zone can serve as associated material, and its spectral information in specific bands (such as 2.3-2.35 μm) can help indicate the distribution of mineralization zones; the aforementioned associated material bands are the characteristic absorption bands of the associated material.

[0069] For example, the above-mentioned determination of the associated material band based on the correlation coefficient between each band and the feature band at the target pixel can be achieved as follows: extract the full-band spectral value of the target pixel in the remote sensing image data; determine the associated material band probability of each non-feature band based on the spectral values ​​of the target pixel in the full-band non-feature band and feature band; construct a wavelength-associated material band probability curve based on the data pairs of non-feature band and associated material band probability; obtain each second local maximum point in the wavelength-associated material band probability curve, and determine the non-feature band corresponding to the second local maximum point as the associated material band.

[0070] Specifically, taking band λ in remote sensing image data, the probability of associated material bands in band λ can be calculated using the following formula:

[0071]

[0072] Among them, P λ λ represents the band probability of associated material in band λ; m represents the number of characteristic bands of the target mineral; I represents the set of the aforementioned target pixels. For the target pixel set I, the i-th pixel in the p-th feature band λ′ p spectral values ​​at The spectral value X of the i-th pixel at band λ i,λ The correlation coefficient between them. For example, this correlation coefficient can be the Pearson correlation coefficient; The probability P of the associated material band is obtained by averaging the sum of the spectral correlations between band λ and the characteristic bands of the target mineral in the target pixel set I. λ P λ P reflects the average correlation between band λ and the target mineral feature bands in the target pixel set I. λ The larger the value, the higher the probability that the spectral characteristics of band λ are highly correlated with the characteristic bands of the target mineral in the target pixel set I, and the greater the probability that it is an associated substance with the target mineral.

[0073] In this embodiment of the application, after determining the probability of associated material bands corresponding to each band of the remote sensing image data through the above process, a coordinate system can be established with wavelength as the horizontal axis and associated material band probability as the vertical axis; based on each non-feature band in the remote sensing image data and its corresponding associated material band probability, a wavelength-associated material band probability curve is plotted in the established coordinate system; local maxima points (i.e., the above-mentioned second local maxima points) in the wavelength-associated material band probability curve are selected, and the band corresponding to the second local maxima point is determined as the associated material band.

[0074] In step S150, a second evaluation model is constructed based on the actual mineral content of the calibration sample and its spectral value in the associated material band, and the second mineral content estimate of each pixel in the remote sensing image data is determined based on the second evaluation model.

[0075] In the embodiments of this application, the second evaluation model is a quantitative relationship model constructed based on the actual mineral content of the calibration sample and its spectral value in the associated material band, which is used to convert the spectral signal of the associated material band into an estimated value of the target mineral content.

[0076] For example, the above-mentioned construction of the second evaluation model based on the actual mineral content of the calibration sample and its spectral value in the associated material band can be achieved as follows: extract the spectral value of each calibration sample in the associated material band and perform standardization processing to obtain the spectral value of the associated material band; obtain the actual mineral content of the calibration sample and perform standardization processing; based on the partial least squares regression method, use the spectral value of the associated material band of the calibration sample as the independent variable and the actual mineral content of the calibration sample as the dependent variable to construct the second evaluation model. In the above process, the mean-standard deviation standardization method can be used to standardize the actual mineral content and its spectral values ​​in the associated material band of the calibration sample. Standardization maps the actual mineral content and its spectral values ​​in the associated material band to a uniform data scale, eliminating dimensional differences. The second evaluation model can be constructed using the standardized spectral values ​​of the associated material band of the calibration sample as input and the standardized actual mineral content as output. A quantitative relationship between the spectral values ​​of the associated material band and the actual mineral content is established using partial least squares regression (PLS), resulting in the second evaluation model as follows:

[0077]

[0078] Among them, F′ i is the estimated value of the second mineral content of the i-th pixel in the remote sensing image data; m′ is the number of bands of associated material; Let be the spectral value of the i-th pixel at the first band of the associated material. Let be the spectral value of the i-th pixel at the second band of the associated material. F' represents the spectral value of the i-th pixel at the m′-th associated material band; F′ represents the quantitative relationship determined based on the actual mineral content of the calibration sample and its spectral value at the associated material band.

[0079] In this embodiment of the application, after the second evaluation model is constructed by the above method, the spectral value of a certain pixel in the remote sensing image data at each associated material band can be input into the second evaluation model, and the second evaluation model outputs the second mineral content estimate of the pixel. The second mineral content estimate is used to supplement the information of the concealed or deep mineralization zone.

[0080] In step S160, the extensibility of the target mineral is determined based on the first mineral content estimate, the unit mineral content gradient vector, and the unit direction vector at the neighboring pixels of each pixel. The proportion of the first mineral content estimate and the second mineral content estimate is dynamically adjusted based on the extensibility to obtain the final mineral content estimate of each pixel.

[0081] In this embodiment, the first evaluation model is based on the target mineral characteristic band, which is accurate in identifying shallow mineralization but weak in detecting deep mineralization. The second evaluation model is based on the associated material band, which indirectly reflects deep mineralization, but because the associated material band easily overlaps with other material bands, its estimation accuracy is low when used alone. Therefore, after constructing the first and second evaluation models using the above methods, in order to improve the estimation accuracy of mineral content at each pixel in remote sensing image data, it is necessary to dynamically adjust the weights of the two evaluation models. Since the target mineral content has continuity and extension in spatial distribution, and mineralization zones are mostly strip-shaped, connected, or clustered, this embodiment can provide a basis for the above-mentioned dynamic weight adjustment process by analyzing the gradient of pixel mineral content changes, local consistency, or extension characteristics in the first evaluation model in space.

[0082] In this embodiment, the aforementioned neighboring pixels refer to other pixels within a preset distance threshold range centered on the current pixel, used to analyze the spatial correlation between the current pixel and surrounding mineralization information. For example, if the preset distance threshold is ω, then for the i-th pixel in the remote sensing image data, pixels in the remote sensing image data whose distance to the i-th pixel is less than ω are recorded as neighboring pixels of the i-th pixel.

[0083] In this embodiment, the gradient vector of unit mineral content is the direction vector of the fastest change in target mineral content at the neighboring pixel, with a magnitude of 1, reflecting the spatial trend of mineral content change, and is used to measure the spatial extension direction of the mineralization zone.

[0084] In this embodiment, the aforementioned unit direction vector is a direction vector pointing from a neighboring pixel to the current pixel, with a magnitude of 1, used to characterize the spatial positional relationship between the neighboring pixel and the current pixel.

[0085] In this embodiment, the extensibility of the target mineral is an index used to measure the spatial continuity between the current pixel and its neighboring highly mineralized area (the area formed by the target pixels), reflecting whether the current pixel is on the natural extension path of the mineralization zone. The larger the extensibility value, the stronger the spatial correlation between the current pixel and the neighboring highly mineralized area, and the higher the reliability of the first mineral content estimate.

[0086] For example, the above-mentioned determination of the extensibility of the target mineral based on the estimated first mineral content, unit mineral content gradient vector, and unit direction vector at the neighboring pixels of each pixel can be achieved as follows: For each pixel in the remote sensing image data, the neighboring pixels of the current pixel are determined based on a preset distance threshold; the spectral values ​​of the neighboring pixels at the feature bands are extracted and standardized, and the standardized spectral values ​​are input into the first evaluation model to obtain the estimated first mineral content of the neighboring pixels; the direction of the fastest increase in the target mineral content of the neighboring pixels in the remote sensing image data plane is determined as the direction of the unit mineral content gradient vector, and the direction from the neighboring pixels to the current pixel is determined as the direction of the unit direction vector, and the cosine similarity between the unit mineral content gradient vector and the unit direction vector is calculated; the extensibility of the target mineral at the current pixel is determined based on the estimated first mineral content and cosine similarity of each neighboring pixel.

[0087] Specifically, taking the i-th pixel as an example, the ductility of the target mineral at the i-th pixel can be calculated using the following formula:

[0088]

[0089] Among them, H i Let be the ductility of the target mineral at the i-th pixel; d(i,j) is the distance between the i-th and j-th pixels, ω is the preset distance threshold, and the j-th pixel satisfying d(i,j)<ω is a neighboring pixel of the i-th pixel; F j G is the estimated mineral content of the first ore element at the j-th pixel; j V is the gradient vector of unit mineral content at the j-th pixel; i,j Let G be the unit direction vector pointing from the j-th pixel to the i-th pixel; cos{G j V i,j} represents the cosine similarity between the gradient vector of unit mineral content and the unit direction vector; F j The larger the value of cos{G}, the higher the mineral content of the j-th pixel, and the greater the probability that it is a shallow, highly mineralized point. j V i,j The larger the value of}, the closer the i-th pixel is to the mineralization extension direction of the j-th pixel, hence H iThe larger the value, the more high mineralization points there are around the i-th pixel, and the stronger the possibility that the i-th pixel is in the direction of mineralization extension, that is, the i-th pixel is more likely to belong to the extension part of the mineralization zone.

[0090] After determining the ductility of the target mineral at the current pixel, for example, the above-mentioned dynamic adjustment of the proportions of the first mineral content estimate and the second mineral content estimate based on ductility to obtain the final mineral content estimate of each pixel can be achieved as follows: normalize the ductility of the target mineral at the current pixel to obtain a weight adjustment factor; take the difference between 1 and the weight adjustment factor as the first weight, and determine the first final mineral content factor based on the first weight and the first mineral content estimate of the current pixel; take the weight adjustment factor as the second weight, and determine the second final mineral content factor based on the second weight and the second mineral content estimate of the current pixel; determine the final mineral content estimate at the current pixel based on the first final mineral content factor and the second final mineral content factor.

[0091] Specifically, taking the i-th pixel as an example, the final mineral content estimate at the i-th pixel can be calculated using the following formula:

[0092] F″ i =(1-norm(H) i ))×F i +norm(H i )×F i

[0093] Among them, F″ i H represents the final mineral content at the i-th pixel. i The ductility of the target mineral at the i-th pixel; norm is the normalization function, 1-norm(H i ) is the first weight, norm(H) i ) is the second weight; F i The first mineral content estimate at the i-th pixel is obtained from the first evaluation model constructed based on the target mineral feature bands, reflecting information about the shallow mineralization zone; F′ i The second mineral content estimate at the i-th pixel is obtained from the second evaluation model constructed based on the associated material band. It can indirectly reflect deep mineralization, but the estimation accuracy is low when used alone because the associated material band is prone to overlap with other materials. When the i-th pixel is far from the mineralization zone extension path and the shallow mineralization information is clear, the first weight needs to be increased, and the final mineral content estimate at the i-th pixel mainly depends on the direct mineralization information of the feature band. When the i-th pixel is on the natural extension path of the mineralization zone, the second weight needs to be increased, and the final mineral content estimate at the i-th pixel preferentially uses the information of the associated material band to compensate for the deep mineralization signal.

[0094] Specifically, such as Figure 2 As shown in the figure, regions A and B are core areas with high mineral content and clear shallow mineralization information (dense contour lines, with values ​​decreasing from the inside to the outside). Point C is located on the connecting zone between regions A and B, in the spatial gap between regions A and B. Point D is far from regions A and B and their connecting zone. If point C is obscured by a thick overburden or alteration zone, the spectral signal of the characteristic band will be weakened. Therefore, when estimating the mineral content of point C, the spectral signal of the deep associated rock layer can be captured through the associated material band to indirectly reflect deep mineralization, and the second mineral content estimate needs to be given higher weight. On the other hand, point D is far from the extension direction of regions A and B and is not on the main extension path of the mineralization zone, so there is no obvious need for deep mineralization compensation. Therefore, when estimating the mineral content of point D, the first mineral content estimate needs to be given higher weight.

[0095] In step S170, the mineralization zone of the target mineral is delineated in the area to be delineated based on the final mineral content estimate.

[0096] In the embodiments of this application, the mineralization zone of the target mineral is the area where the target mineral is concentrated. Due to mineralization, it is usually manifested as a continuous or semi-continuous spatial distribution in the form of strips, interconnected shapes, or clusters.

[0097] For example, the above-mentioned delineation of the mineralization zone of the target mineral in the area to be delineated based on the final mineral content estimate can be achieved as follows: the pixels with the final mineral content estimate greater than the second preset threshold are determined as delineated pixels, and the connected regions formed by the delineated pixels are determined to obtain the mineralization zone of the target mineral.

[0098] The foregoing mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, it includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0099] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

[0100] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. Multitasking and parallel processing may be advantageous in certain environments. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

Claims

1. A method for delineating mineralization zone rapidly based on remote sensing technology, characterized in that, The method comprises: acquiring remote sensing image data of a region to be delineated and performing preprocessing; calculating mineral representation coefficients of each band in the remote sensing image data based on standard spectral data of a target mineral and interference spectral data of interfering substances in the region to be delineated, and determining a characteristic band of the target mineral based on the mineral representation coefficients; constructing a first evaluation model based on actual ore content of calibration samples and spectral values thereof at the characteristic band, and determining first ore content estimation values of each pixel point in the remote sensing image data based on the first evaluation model; the calibration samples are points in the region to be delineated at which the actual ore content of the target mineral is known; determining a pixel point at which the first ore content estimation value is greater than a first preset threshold as a target pixel point, and determining a paragenetic substance band based on correlation coefficients of each band and the characteristic band at the target pixel point; constructing a second evaluation model based on the actual ore content of the calibration samples and spectral values thereof at the paragenetic substance band, and determining second ore content estimation values of each pixel point in the remote sensing image data based on the second evaluation model; determining the extensibility of the target mineral based on the first ore content estimation values of neighboring pixel points of each pixel point, a unit mineral content gradient vector and a unit direction vector, and dynamically adjusting the proportion of the first ore content estimation value and the second ore content estimation value based on the extensibility to obtain a final ore content estimation value of each pixel point; delineating a mineralization zone of the target mineral in the region to be delineated based on the final ore content estimation value; the method comprises: acquiring spectral reflectance data of a pure mineral or standard mineral sample corresponding to the target mineral measured under laboratory conditions, denoted as the standard spectral data; determining the distribution position of the interfering substances in the region to be delineated based on the remote sensing image data, and collecting the interfering substances at the distribution position as a measurement sample to acquire spectral reflectance data of the measurement sample, denoted as the interference spectral data; for each band in the remote sensing image data, calculating the absolute value of the difference between the current band and the band at each absorption peak in each of the interference spectral data, and recording the minimum value thereof as a first minimum band distance between the current band and the corresponding interference spectral data; calculating the absolute value of the difference between the current band and the band at each absorption peak in the standard spectral data, and recording the minimum value thereof as a second minimum band distance between the current band and the standard spectral data; determining the mineral representation coefficient of the current band based on the sum of each first minimum band distance and the second minimum band distance; the method comprises: constructing a wavelength-mineral representation coefficient curve based on data pairs of each current band and the mineral representation coefficient; Acquire each first local maximum point in the wavelength-mineral characterization coefficient curve, and determine the current wave band corresponding to the first local maximum point as the characteristic wave band; The first mineral content estimation value, the mineral content gradient vector and the direction vector of the neighbor pixel points of each pixel point are used to determine the ductility of the target mineral, including: For each pixel point in the remote sensing image data, the neighbor pixel points of the current pixel point are determined based on a preset distance threshold; The spectral values of the neighbor pixel points at the characteristic wave band are extracted and standardized, and the standardized spectral values are input into the first evaluation model to obtain the first mineral content estimation value of the neighbor pixel points; The direction in which the target mineral content of the neighbor pixel points increases fastest in the remote sensing image data plane is determined as the direction of the unit mineral content gradient vector, and the direction of the neighbor pixel points pointing to the current pixel point is determined as the direction of the unit direction vector. The cosine similarity of the unit mineral content gradient vector and the unit direction vector is calculated; The first mineral content estimation value and the cosine similarity corresponding to each neighbor pixel point are used to determine the ductility of the target mineral at the current pixel point; The proportion of the first mineral content estimation value and the second mineral content estimation value is dynamically adjusted based on the ductility to obtain the final mineral content estimation value of each pixel point, including: The ductility of the target mineral at the current pixel point is normalized to obtain a weight adjustment factor; The difference between 1 and the weight adjustment factor is taken as a first weight, and a first final mineral content factor is determined based on the first weight and the first mineral content estimation value of the current pixel point; The weight adjustment factor is taken as a second weight, and a second final mineral content factor is determined based on the second weight and the second mineral content estimation value of the current pixel point; The final mineral content estimation value at the current pixel point is determined based on the first final mineral content factor and the second final mineral content factor.

2. The method for rapid delineation of mineralized zones based on remote sensing technology according to claim 1, characterized in that, The first evaluation model is constructed based on the actual mineral content of the calibration sample and its spectral value at the characteristic wave band, including: The spectral value of each calibration sample at the characteristic wave band is extracted and standardized to obtain the characteristic wave band spectral value; The actual mineral content of the calibration sample is obtained and standardized; Based on the partial least squares regression method, the characteristic wave band spectral value of the calibration sample is taken as the independent variable, and the actual mineral content of the calibration sample is taken as the dependent variable to construct the first evaluation model.

3. The remote sensing based method of rapid delineation of mineralized zones according to claim 1, characterized in that, The associated substance wave band is determined based on the correlation coefficient of each wave band at the target pixel point and the characteristic wave band, including: The full wave band spectral value of the target pixel point in the remote sensing image data is extracted; The associated substance wave band probability of each non-characteristic wave band is determined based on the spectral value of the target pixel point in the non-characteristic wave band and the characteristic wave band in the full wave band; A wavelength-associated substance wave band probability curve is constructed based on the data of the non-characteristic wave band and the associated substance wave band probability; Obtaining each second local maximum point in the wavelength-companion material wave band probability curve, and determining the non-characteristic wave band corresponding to the second local maximum point as the companion material wave band.

4. The method for rapid delineation of mineralized zones based on remote sensing technology according to claim 3, characterized in that, The companion material wave band probability of each non-characteristic wave band is determined based on the spectral values of the target pixel point in the non-characteristic wave band and the characteristic wave band in the full wave band, including: For each non-characteristic wave band, a first spectral value sequence of the target pixel point in the current non-characteristic wave band is determined. A second spectral value sequence of the target pixel point in each characteristic wave band is obtained. The Pearson correlation coefficients of the first spectral value sequence of the current non-characteristic wave band and each second spectral value sequence are calculated, and the average value of the Pearson correlation coefficients is calculated, denoted as the companion material wave band probability of the current non-characteristic wave band.

5. The remote sensing based method of rapid delineation of mineralized zones according to claim 1, characterized in that, The second evaluation model is constructed based on the actual ore content of the calibration sample and the spectral value thereof at the companion material wave band, including: The spectral value of each calibration sample at the companion material wave band is extracted and standardized to obtain the companion material wave band spectral value. The actual ore content of the calibration sample is obtained and standardized. Based on the partial least squares regression method, the companion material wave band spectral value of the calibration sample is taken as the independent variable, and the actual ore content of the calibration sample is taken as the dependent variable to construct the second evaluation model.

6. The remote sensing based method of rapid delineation of mineralized zones according to claim 1, characterized in that, The mineralization zone of the target mineral in the to-be-delineated area is delineated based on the final ore content estimation value, including: The pixel point with the final ore content estimation value greater than a second preset threshold value is determined as a delineated pixel point, and a connected region formed by the delineated pixel points is determined to obtain the mineralization zone of the target mineral.

Images