Geological resource dynamic monitoring method and system based on RS image enhancement

Abstract

The invention relates to a geological resource dynamic monitoring method and system based on RS image enhancement, computer equipment and a medium. According to the method, a high-frequency edge image of an RS image is extracted through multi-scale edge detection, high-frequency edge blocks are further divided, and a gray level mutation sequence is calculated to determine the contrast ratio and representative gray level. And based on the parameters, adaptively adjusting sharpening parameters, performing sharpening processing on the high-frequency edge blocks and background pixels, finally generating an edge sharpening image, and superposing the edge sharpening image on an original image to identify a geological resource area. By means of the method, richer edge details and overall contour information can be captured, edge sharpening is more accurate, and the situation that the edge is lost or blurred is reduced; the method can achieve the automatic adjustment according to the local features of the image, better adapts to remote sensing images with different complexity degrees, and improves the performance of the method for dynamically monitoring geological resources in various scenes.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a method and system for dynamic monitoring of geological resources based on RS image enhancement. Background Art

[0002] In the field of dynamic monitoring of geological resources, remote sensing imaging technology has become one of the main monitoring methods due to its advantage of being able to obtain surface information over a large area and periodically. By analyzing remote sensing images from different periods, key information such as changes in geological resource reserves, expansion of mining areas, and changes in related geological environments can be timely grasped, thereby providing a decision-making basis for scientific planning, rational development, and environmental protection.

[0003] However, the current dynamic monitoring technology of geological resources based on remote sensing images faces many challenges. Among them, the edge clarity problem of geological resource areas in remote sensing images is particularly prominent. Poor edge clarity directly affects the definition of the accurate location and scope of geological resources, which in turn restricts the accuracy and reliability of monitoring.

[0004] On the one hand, traditional edge detection and enhancement methods have obvious defects in dealing with this problem. In terms of edge detection, most existing technologies use single-scale edge detection operators, such as the common Sobel operator. This single-scale detection method can only capture edge information at a specific scale, and has poor adaptability to the edge structure of complex and diverse geological resource areas. For example, in some areas containing small mineral veins or complex geological structures, a single scale may not be able to accurately detect all key edge details, resulting in the loss of some edge information; and in larger geological areas, the overall outline may not be fully presented due to scale issues, thus affecting the accurate judgment of the scope of geological resources.

[0005] On the other hand, traditional methods lack the ability to adaptively adjust the features of different regions of the image during the entire processing process. Whether it is edge detection or subsequent sharpening parameter calculation, most of them rely on fixed algorithm modes and parameter settings, and cannot be dynamically optimized according to changes in local image features. For example, when the texture complexity and grayscale distribution of different regions in the image vary greatly, the fixed parameters cannot fully adapt to these changes, resulting in excessive or insufficient edge enhancement in some areas, further reducing the accuracy of edge detection and enhancement, and limiting the application effect of dynamic monitoring technology of geological resources.

[0006] In summary, in order to improve the accuracy and reliability of dynamic monitoring of geological resources, an improved method that can overcome the shortcomings of existing technologies is urgently needed. Summary of the invention

[0007] Based on this, it is necessary to provide a geological resource dynamic monitoring method, system, computer equipment and storage medium based on RS image enhancement that can improve edge clarity and adapt to remote sensing images of different complexity levels in response to the above technical problems.

[0008] In the first aspect, the present application provides a method for dynamic monitoring of geological resources based on RS image enhancement, comprising:

[0009] Acquire the RS image, and obtain the high-frequency edge image of the RS image based on the multi-scale edge detection and fusion method; obtain the high-frequency edge block of the high-frequency edge image based on the high-frequency edge image;

[0010] Obtain the grayscale values ​​of all pixels of each high-frequency edge block, calculate the absolute value of the grayscale value difference between two horizontally adjacent pixels, and use it as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence;

[0011] According to the frequency of each grayscale mutation value in the grayscale mutation sequence, the contrast of the high-frequency edge block is obtained; the mean grayscale value of all pixels in the high-frequency edge block is calculated as the representative grayscale of the high-frequency edge block;

[0012] According to the contrast and representative grayscale of the high-frequency edge block, an adaptive adjustment parameter is obtained; the adaptive adjustment parameter is multiplied by the grayscale value of each pixel in the high-frequency edge block to obtain a sharpening parameter for each pixel in the high-frequency edge block;

[0013] According to the representative grayscale of all high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block, the sharpening value of each pixel in each high-frequency edge block is obtained; the gradient amplitude of the pixel outside the high-frequency edge block is used as its sharpening value;

[0014] The sharpening value of each pixel in all high-frequency edge blocks and the sharpening value of all background pixels are used to form an edge-sharpened image; the edge-sharpened image is superimposed on the original RS image, and the geological resource area is identified on the superimposed RS image to obtain the geological resource area.

[0015] In the second aspect, the present application also provides a geological resource dynamic monitoring system based on RS image enhancement, including:

[0016] The high-frequency edge block generation module is used to obtain the RS image, and obtain the high-frequency edge image of the RS image based on the multi-scale edge detection and fusion method; based on the high-frequency edge image, obtain the high-frequency edge block of the high-frequency edge image;

[0017] A grayscale mutation sequence construction module is used to obtain the grayscale values ​​of all pixels in each high-frequency edge block, calculate the absolute value of the grayscale value difference between two horizontally adjacent pixels, and use it as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence;

[0018] The high-frequency edge block parameter calculation module is used to obtain the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence; calculate the grayscale value mean of all pixels in the high-frequency edge block as the representative grayscale of the high-frequency edge block;

[0019] The pixel sharpening parameter acquisition module is used to obtain an adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block; the adaptive adjustment parameter is multiplied by the grayscale value of each pixel in the high-frequency edge block to obtain the sharpening parameter of each pixel in the high-frequency edge block;

[0020] The pixel sharpening value calculation module is used to obtain the sharpening value of each pixel in each high-frequency edge block according to the representative grayscale of all high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block; and take the gradient amplitude of the pixel outside the high-frequency edge block as its sharpening value;

[0021] The geological resource area identification module is used to use the sharpening value of each pixel in all high-frequency edge blocks and the sharpening value of all background pixels to form an edge-sharpened image; the edge-sharpened image is superimposed on the original RS image, and the geological resource area is identified on the superimposed RS image to obtain the geological resource area.

[0022] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, a dynamic monitoring method of geological resources based on RS image enhancement as in the first aspect is implemented.

[0023] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for dynamic monitoring of geological resources based on RS image enhancement as in the first aspect is implemented.

[0024] The above-mentioned method, system, computer equipment and storage medium for dynamic monitoring of geological resources based on RS image enhancement obtain high-frequency edge images by utilizing multi-scale edge detection and fusion methods, thereby being able to capture richer edge details and overall contour information, making edge sharpening more accurate, and reducing edge loss or blur; by utilizing adaptive adjustment parameters to calculate the sharpening parameters of each pixel in the high-frequency edge block, it is able to automatically adjust according to the local features of the image, better adapt to remote sensing images of different complexities, and improve the performance of the dynamic monitoring method of geological resources in various scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 A schematic diagram of a flow chart of a method for dynamic monitoring of geological resources based on RS image enhancement provided by the present invention;

[0027] Figure 2 A structural schematic diagram of a geological resource dynamic monitoring system based on RS image enhancement provided by the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0029] refer to Figure 1 , which shows a flow chart of a method for dynamic monitoring of geological resources based on RS image enhancement provided by the present application, the method comprising the following steps:

[0030] S101, obtaining a RS image, and obtaining a high-frequency edge image of the RS image based on a multi-scale edge detection and fusion method; and obtaining a high-frequency edge block of the high-frequency edge image based on the high-frequency edge image.

[0031] Specifically, RS images can be obtained through appropriate remote sensing technology. This involves using remote sensing equipment carried by satellites, aircraft, etc. to collect data according to predetermined orbits, routes, and set observation parameters. These observation parameters include but are not limited to band range, resolution, observation angle, etc. For example, common satellite remote sensing platforms such as the Landsat series and the Sentinel series can observe large areas of the earth's surface at different time intervals, thereby obtaining rich RS image data.

[0032] After the original RS image is acquired, it is preprocessed to ensure the quality and availability of the data. The preprocessing step can include radiation correction, which is used to eliminate the radiation error caused by the sensor's own characteristics, atmospheric scattering and absorption, so that the gray value of the image can accurately reflect the true radiation characteristics of the ground object; geometric correction is to correct the geometric deformation caused by the change of aircraft attitude and earth curvature during the acquisition process, so that the image can accurately match the actual geographic coordinate system.

[0033] Multi-scale edge detection is a technology that can capture image edge information at different scales. It uses filters of different sizes (such as Gaussian filters) to perform convolution operations on RS images to highlight edge features in images at different scales. Smaller-scale filters can detect finer edge details in images, while larger-scale filters can capture more macroscopic and more holistic edge structures.

[0034] Then, the edge information detected at each scale is fused. The fusion process can be done in a variety of ways, such as a fusion method based on weighted average, which assigns corresponding weights according to the importance of edge information at different scales, and then sums them up to obtain a high-frequency edge image that combines the advantages of each scale. This high-frequency edge image can more comprehensively and accurately reflect the edge features in the RS image.

[0035] In order to further analyze and process the high-frequency edge image, it needs to be divided into several high-frequency edge blocks. The division method can be determined according to specific needs and algorithms. A common method is to use a regular grid division method, such as dividing the high-frequency edge image into multiple small blocks of a fixed size (such as a square or rectangle), and each small block is a high-frequency edge block. Such division helps to perform more detailed quantitative analysis and processing of local edge features in the future.

[0036] S102, obtaining the grayscale values ​​of all pixels of each high-frequency edge block, calculating the absolute value of the grayscale value difference between two horizontally adjacent pixels as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence.

[0037] Specifically, for each divided high-frequency edge block, it is necessary to traverse each pixel in it and obtain its corresponding grayscale value. In digital images, the grayscale value is a value used to represent the brightness of a pixel, usually ranging from 0 to 255 (for 8-bit grayscale images), 0 represents black, 255 represents white, and the values ​​in between correspond to different degrees of gray. The grayscale values ​​of all pixels in the high-frequency edge block are obtained one by one.

[0038] After obtaining the grayscale values ​​of all pixels in each high-frequency edge block, the grayscale changes between adjacent pixels are analyzed. Specifically, for each high-frequency edge block, the grayscale value differences between two adjacent pixels are calculated in the horizontal direction, and their absolute values ​​are taken as the grayscale mutation value of the position. For example, for the i-th and i+1-th pixels (horizontally adjacent) in a high-frequency edge block, their grayscale values ​​are G and G, respectively. i and G i+1 , then the grayscale mutation value between them is |G i -G i+1 In this way, the intensity of grayscale changes between adjacent pixels in the horizontal direction in each high-frequency edge block can be quantified, thereby capturing the grayscale mutation characteristics of the edge area.

[0039] All grayscale mutation values ​​calculated in each high-frequency edge block according to the above method are arranged in a certain order (such as from left to right along the horizontal direction) to form a grayscale mutation sequence of the high-frequency edge block.

[0040] S103, obtaining the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence; and calculating the average grayscale value of all pixels in the high-frequency edge block as the representative grayscale of the high-frequency edge block.

[0041] Specifically, first, statistical analysis is performed on the constructed grayscale mutation sequence to calculate the frequency of occurrence of each grayscale mutation value. For example, if a grayscale mutation value appears n times in the entire grayscale mutation sequence, and the total length of the sequence is N, then the frequency of the grayscale mutation value is

[0042] Then, the contrast of the high-frequency edge block is calculated based on the frequency of these grayscale mutation values. Contrast is an indicator to measure the degree of difference between the bright and dark parts of an image. In this context, the contrast of the high-frequency edge block can be comprehensively calculated by a specific calculation formula, taking the frequency of different grayscale mutation values ​​as input parameters. For example, a weighted summation method can be used to assign corresponding weights according to the importance of the impact of different grayscale mutation values ​​on the contrast, and then each grayscale mutation value is multiplied by its corresponding weight and summed to obtain the contrast of the high-frequency edge block. The contrast calculated in this way can reflect the drastic degree of grayscale change in the high-frequency edge block, thereby reflecting the obviousness of the edge to a certain extent.

[0043] For each high-frequency edge block, the grayscale values ​​of all pixels in it are summed. For example, there are M pixels in a high-frequency edge block, and their grayscale values ​​are G1, G2, …, G M , then the sum of their gray values ​​is ∑G i(i=1,2,…,M).

[0044] Then, this sum is divided by the total number of pixels M, and the result is the mean grayscale value of the high-frequency edge block. This mean is defined as the representative grayscale of the high-frequency edge block, which can represent the overall grayscale level of the high-frequency edge block to a certain extent.

[0045] S104, obtaining an adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block; and multiplying the grayscale value of each pixel in the high-frequency edge block by the adaptive adjustment parameter to obtain a sharpening parameter for each pixel in the high-frequency edge block.

[0046] Specifically, according to the contrast and representative grayscale of the high-frequency edge block calculated above, the adaptive adjustment parameter is determined through a specific functional relationship. This functional relationship can be an empirical formula summarized through a large number of experiments and analyses, or can be derived based on a certain theoretical model.

[0047] For example, a function f(contrast, mean_gray) can be set, where contrast represents the contrast of the high-frequency edge block, and mean_gray represents the representative grayscale of the high-frequency edge block. Substituting these two parameters into the function, after a series of calculation operations (such as multiplication, addition, division, etc.), the adaptive adjustment parameter can be adaptively adjusted according to the contrast and representative grayscale characteristics of the high-frequency edge block itself to adapt to the specific conditions of different edge blocks.

[0048] For each high-frequency edge block, after obtaining the adaptive adjustment parameter, multiply it with the grayscale value of each pixel in the high-frequency edge block. For example, the grayscale value of the jth pixel in a high-frequency edge block is G j , the adaptive adjustment parameter is α, then the sharpening parameter of the pixel is α×G j In this way, each pixel is assigned a sharpening parameter that is related to its own grayscale value and adaptively adjusted according to the overall characteristics of the high-frequency edge block.

[0049] S105. Obtain a sharpening value for each pixel in each high-frequency edge block according to the representative grayscales of all high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block; and use the gradient amplitude of the pixel outside the high-frequency edge block as its sharpening value.

[0050] Specifically, considering the relationship between different high-frequency edge blocks and the overall image characteristics, it is necessary to integrate the representative grayscale information of all high-frequency edge blocks. For each pixel in each high-frequency edge block, its sharpening parameter is calculated with the representative grayscale of the high-frequency edge block to obtain the final sharpening value.

[0051] For example, an addition operation can be used, that is, the sharpening parameter of each pixel is added to the representative grayscale of the high-frequency edge block to obtain the sharpening value of the pixel. Alternatively, other more complex operation methods, such as weighted summation, can be used. The specific operation method depends on the overall algorithm design and the requirements for image enhancement effect.

[0052] For those pixels that are not within the high-frequency edge block range, that is, pixels that belong to the background area, their gradient amplitude is used as the sharpening value. The gradient amplitude is an indicator used to measure the grayscale change rate of pixels in the image. It can be calculated by calculating the grayscale change rate of pixels in different directions (such as horizontal and vertical directions), and then synthesized into a gradient amplitude using methods such as the Pythagorean theorem. The purpose of this is to perform appropriate processing on the background area while sharpening the edge area, so that the entire image can maintain good visual effects and information integrity after enhancement.

[0053] S106, using the sharpening value of each pixel in all high-frequency edge blocks and the sharpening values ​​of all background pixels to form an edge-sharpened image; superimposing the edge-sharpened image on the original RS image, identifying the geological resource area on the superimposed RS image, and obtaining the geological resource area.

[0054] Specifically, the sharpening values ​​of each pixel in all high-frequency edge blocks calculated above are arranged according to their positions in the original high-frequency edge blocks, and the sharpening values ​​of all background pixels are also arranged according to their positions in the original image, and then these two parts of data are combined to form an edge-sharpened image. This edge-sharpened image retains the basic information of the original RS image, and through the sharpening of the edge area and the background area, the edge of the image is clearer and the contrast is higher.

[0055] The edge-sharpened image is superimposed on the original RS image, and a simple addition operation or other more appropriate superposition method can be used to organically combine the information of the original RS image and the information of the edge-sharpened image. Then, the superimposed RS image is analyzed and processed using existing geological resource identification technologies and algorithms to identify the geological resource areas therein. These geological resource identification technologies can include classification algorithms based on machine learning (such as support vector machines, random forests, etc.), methods based on feature extraction and matching, etc. By comprehensively analyzing and identifying the superimposed RS images, the location, scope and other information of the geological resource area are finally determined, thereby achieving the purpose of dynamic monitoring of geological resources.

[0056] The above-mentioned method for dynamic monitoring of geological resources based on RS image enhancement obtains high-frequency edge images by utilizing multi-scale edge detection and fusion methods, thereby being able to capture richer edge details and overall contour information, making edge sharpening more accurate, and reducing edge loss or blurring; by utilizing adaptive adjustment parameters to calculate the sharpening parameters of each pixel in the high-frequency edge block, it can automatically adjust according to the local features of the image, better adapt to remote sensing images of different complexities, and improve the performance of the dynamic monitoring method of geological resources in various scenarios.

[0057] In an optional embodiment, obtaining a RS image and obtaining a high-frequency edge image of the RS image based on a multi-scale edge detection and fusion method includes the following steps:

[0058] RS images are collected by satellite and grayscaled to obtain grayscale remote sensing images;

[0059] The multi-scale edge detection method is adopted, and multiple edge detection operators of different scales are used to process the grayscale remote sensing image separately to obtain edge images of different scales.

[0060] Fusing edge images of different scales to obtain a fused edge image;

[0061] The fused edge image is filtered through a high-pass filter to obtain a high-frequency edge image.

[0062] Specifically, RS images are acquired using satellites specifically used for remote sensing observations. These satellites are usually equipped with advanced optical sensors or other types of remote sensing equipment. They operate in space according to preset orbits and perform periodic observations of specific areas on the earth's surface. For example, satellites like the Landsat series have sensors with multiple bands that can simultaneously collect reflection or radiation information from the earth's surface in different bands. During operation, the satellite will continuously acquire a large amount of remote sensing data according to the set time intervals and coverage range. These data constitute the original RS images. These original images contain rich information about land objects, such as topography, vegetation coverage, water distribution, etc.

[0063] The original RS image is often a multi-band color image, each band corresponds to a different spectral range, recording the reflection or radiation characteristics of the object in the spectral range. However, in the subsequent edge detection and other processing, in order to simplify the calculation and highlight the basic structural features of the image, it is usually necessary to convert it into a grayscale image, that is, to perform grayscale processing.

[0064] There are many methods for grayscale processing, and a common one is the weighted average method. For example, for an RGB color RS image (red, green, and blue bands), the pixel values ​​of the three bands can be weighted and summed according to certain weights to obtain the corresponding grayscale value. Generally speaking, the weight distribution method that can be used is such as the red band weight is 0.299, the green band weight is 0.587, and the blue band weight is 0.114. Then, for each pixel in the image, the RGB value is calculated and the result of the weighted summation according to the above weights is used as the grayscale value of the pixel. After such processing of all pixels in the entire image, the original color RS image is converted into a grayscale remote sensing image. This grayscale remote sensing image uses a single grayscale value to represent the brightness of each pixel, which is more convenient for subsequent edge detection and other operations.

[0065] Edge detection operator is a mathematical tool used to identify edge positions in images. Common edge detection operators include Sobel operator, Prewitt operator, Canny operator, etc. Different operators have different characteristics and detection effects. In multi-scale edge detection, multiple edge detection operators of the same or multiple types at different scales are selected to process grayscale remote sensing images separately.

[0066] For example, for the Sobel operator, different scale settings can be achieved by changing the size of its convolution kernel. Suppose we choose three different scales of Sobel operators, namely small scale, medium scale and large scale. For the small-scale Sobel operator, its convolution kernel may be relatively small, such as a 3×3 matrix form. When it performs convolution operations on grayscale remote sensing images, it can quickly detect local and relatively fine edge changes in the image; the convolution kernel of the medium-scale Sobel operator may be slightly larger, such as a 5×5 matrix form, which can detect edge features that are slightly more macroscopic than the small scale; the convolution kernel of the large-scale Sobel operator is even larger, such as a 7×7 matrix form, which focuses on capturing a more holistic edge structure in the image.

[0067] When the grayscale remote sensing images are convolved with the three Sobel operators at different scales, edge images at different scales are obtained. The edge images at each scale highlight the edge features of interest at that scale, and these edge images provide a rich source of edge information for subsequent fusion operations.

[0068] Since edge images of different scales focus on capturing edge information at different levels, using an edge image of a certain scale alone may not be able to fully and accurately reflect all edge features of the entire grayscale remote sensing image. Therefore, it is necessary to fuse these edge images of different scales to combine the advantages of each scale to obtain a more complete fused edge image that can more accurately reflect the edge of the image.

[0069] There are many fusion methods, and the most common one is the fusion method based on weighted average. In this method, corresponding weights are assigned according to the importance or reliability of edge images of different scales. For example, for the edge images of small, medium, and large scales mentioned above, the weight of small-scale edge images may be determined to be 0.3, the weight of medium-scale edge images to be 0.5, and the weight of large-scale edge images to be 0.2 based on experience or through some experimental analysis.

[0070] Then, for each pixel in the image, the edge value of the pixel is obtained from the edge images of the three scales (usually the value representing the edge strength or position obtained after edge detection operator processing), and then weighted summed according to their respective weights. For example, let the edge value of a pixel in the small-scale edge image be E1, the edge value of the pixel in the medium-scale edge image be E2, and the edge value of the pixel in the large-scale edge image be E3. Then, after weighted average fusion, the edge value E of the pixel in the fused edge image is: E = 0.3 × E1 + 0.5 × E2 + 0.2 × E3.

[0071] By performing such a weighted average fusion operation on all pixels in the image, the entire fused edge image can be obtained. This fused edge image combines the advantages of edge images of different scales and can more comprehensively display the edge information in the grayscale remote sensing image.

[0072] A high-pass filter is a filter that allows high-frequency signals to pass through while suppressing low-frequency signals. In the field of images, high-frequency signals usually correspond to the edges, details and other parts of the image that change dramatically, while low-frequency signals correspond to the smoother and slower-changing areas of the image, such as large areas of pure color.

[0073] When the fused edge image is processed by a high-pass filter, its main purpose is to further highlight the high-frequency information such as the edges in the image, remove or weaken the low-frequency information that may interfere with the edge detection results, and thus obtain a purer high-frequency edge image with more prominent high-frequency information.

[0074] There are many forms of high-pass filters, such as ideal high-pass filters, Butterworth high-pass filters, etc. Taking the Butterworth high-pass filter as an example, it needs to set some parameters, such as cutoff frequency.

[0075] When filtering the fused edge image, we must first determine the appropriate cutoff frequency and other parameters based on the specific conditions of the image and the requirements for edge prominence. Then, we take the grayscale value of each pixel in the fused edge image as input, and filter each pixel using the calculation formula of the Butterworth high-pass filter to obtain the filtered grayscale value.

[0076] After such filtering is performed on all pixels in the fused edge image, a high-frequency edge image is obtained. The high-frequency information (i.e., edge information) in this high-frequency edge image is further highlighted, which is more conducive to the subsequent further analysis and processing of the RS image based on high-frequency edges.

[0077] In an optional embodiment, fusing edge images of different scales to obtain a fused edge image includes the following steps:

[0078] Assign weights according to the reliability or importance of edges at different scales, and perform weighted fusion on edge images at different scales according to the weights to obtain a fused edge image;

[0079] The formula for weighted fusion is:

[0080]

[0081] Among them, E fusion is the fused edge image, E i is the edge image of the i-th scale, n is the number of scales, ω i is the edge image E of the i-th scale i The weight coefficient of .

[0082] Specifically, when determining the weights of edge images of different scales, multiple factors can be considered to evaluate their reliability or importance. For edge images of smaller scales, although it can detect fine edge details, these details may be more susceptible to interference from factors such as noise, and their reliability may be relatively low to a certain extent. However, in certain specific scenarios, such as the analysis of tiny geological structure features, these fine edge details may have higher importance.

[0083] For larger-scale edge images, it can capture more macroscopic edge structures, is relatively less affected by noise, and may be more reliable, but may ignore some local subtle edge information. For example, when analyzing the overall boundaries of a large geological area, large-scale edge information is more important.

[0084] The reliability and importance of edge images of different scales can be comprehensively evaluated by testing and analyzing them on actual geological sample data or simulated data, and statistically analyzing indicators such as the accuracy of edge detection and the ability to capture edges of different types of geological features. The general direction of weight distribution can also be determined based on the experience and knowledge of experts in the field and the previous understanding of the role of edge information of different scales in geological resource monitoring.

[0085] Assume that we use three different scales of edge detection results for fusion, namely small scale (S), medium scale (M) and large scale (L). After the above reliability and importance evaluation, it may be determined that the weight coefficient of the small scale edge image is ω1 = 0.2, the weight coefficient of the medium scale edge image is ω2 = 0.5, and the weight coefficient of the large scale edge image is ω3 = 0.3. The weight distribution here reflects that the medium scale edge image has a relatively high reliability and importance under comprehensive consideration, followed by the large scale edge image, and the small scale edge image is relatively low but still has a certain contribution.

[0086] For each pixel in the fused edge image, a weighted fusion calculation needs to be performed according to the weights determined above. Taking a certain pixel as an example, let the edge value (which can be a value representing edge strength, gradient, etc.) of the pixel in the small-scale edge image be E1, the edge value in the medium-scale edge image be E2, and the edge value in the large-scale edge image be E3.

[0087] According to the weighted fusion formula Here n = 3 (because there are three scales), then the fused edge value E of the pixel in the fused edge image is fusion For E fusion =ω1E1+ω2E2+ω3E3=0.2E1+0.5E2+0.3E3.

[0088] This calculation process is repeated for each pixel in the fused edge image to obtain the entire fused edge image. Through such weighted fusion, the fused edge image integrates the edge information of edge images of different scales at the pixel point, so that the final fused edge image can better reflect the edge characteristics related to geological resources at different scales.

[0089] The fused edge image has obvious advantages over the single-scale edge image. It can more comprehensively and accurately depict the edge conditions of the geological resource area. For example, in some complex geological and topographical areas, there are not only the edges of macroscopic geological structures such as large-scale mountains and basins, but also the edges of microscopic geological features such as rock textures and small faults. The fused edge image obtained by weighted fusion can clearly display the large-scale edge contours of the macroscopic geological structure without losing the fine edge details of the microscopic geological features, providing a richer and more reliable edge information basis for the subsequent accurate identification of geological resource areas, and helping to improve the accuracy and effectiveness of dynamic monitoring of geological resources.

[0090] In an optional embodiment, obtaining the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence includes the following steps:

[0091] The contrast of high-frequency edge blocks is calculated using the following formula:

[0092]

[0093] Among them, C is the contrast of high-frequency edge blocks, a j is the size of the jth grayscale mutation value, f j is the frequency of the jth grayscale mutation value, and m is the number of grayscale mutation values ​​in the grayscale mutation sequence.

[0094] Specifically, in the formula for calculating the contrast of high-frequency edge blocks In the equation, C represents the contrast of the high-frequency edge block, which is the final result to be calculated and is used to measure the drastic degree of grayscale change in the high-frequency edge block. j represents the size of the jth grayscale mutation value, that is, the absolute value of the grayscale value difference between two adjacent pixels (as calculated in the previous step), which reflects the steepness of the local edge. j is the frequency of the jth grayscale mutation value, that is, the proportion of the number of occurrences of this grayscale mutation value in the entire grayscale mutation sequence to the total number of occurrences. The frequency information can reflect the distribution of different grayscale mutation values ​​in the edge block. m is the number of grayscale mutation values ​​in the grayscale mutation sequence, which indicates how many different grayscale mutation values ​​appear in this high-frequency edge block.

[0095] Assume that in a high-frequency edge block, three different grayscale mutation values ​​are calculated, namely a1 = 3 (indicating that the grayscale value difference between some adjacent pixels is 3), a2 = 5, and a3 = 7. At the same time, the frequencies of these three grayscale mutation values ​​in the entire grayscale mutation sequence are statistically f1 = 0.2 (i.e., the proportion of the grayscale mutation value in the sequence is 20%), f2 = 0.3, and f3 = 0.5. Here m = 3, because there are three different grayscale mutation values.

[0096] For each grayscale mutation value a j , first calculate its contribution term (a j ) 2 f j Taking a1=3 and f1=0.2 as an example, the contribution term is (a1) 2 f1=(3) 2 × 0.2 = 1.8. Similarly, for a2 = 5 and f2 = 0.3, the contribution term is (a2) 2 f2=(5) 2 ×0.3=7.5; for a3=7 and f3=0.5, the contribution term is (a3) 2 f3=(7) 2 ×0.5=24.5.

[0097] Add up the contribution items of all grayscale mutation values ​​to get the contrast C of the high-frequency edge block. That is, C = 1.8 + 7.5 + 24.5 = 33.8. The calculation result 33.8 is the contrast value of the high-frequency edge block. The larger the contrast value, the more drastic the grayscale change in the high-frequency edge block, and the more obvious the edge; conversely, the smaller the contrast value, the relatively gentle grayscale change, and the less obvious the edge. Through this calculation method, the contrast characteristics of the high-frequency edge block can be accurately quantified.

[0098] In the dynamic monitoring of geological resources, the contrast of high-frequency edge blocks is a key indicator. A higher contrast means that the changes in geological characteristics within the edge block are more obvious, which may correspond to important areas such as the junction between different geological layers and the edge of geological structures. For example, when analyzing the boundary between rock layers and soil layers, due to the large differences in their physical and chemical properties, they often appear as high-contrast edges on RS images. By calculating the contrast, these key areas where geological resource changes may occur can be quickly located, providing clues for further detailed analysis.

[0099] Contrast also plays an important role in the identification and monitoring of geological resources such as mineral resources. Some specific mineral resources may have obvious contrast differences with the surrounding rocks or soil on RS images due to their unique physical properties (such as reflectivity, absorptivity, etc.). By monitoring the changes in the contrast of high-frequency edge blocks in RS images at different times, it is possible to determine whether the geological resources in the area have undergone dynamic changes such as mining and deposition, thereby achieving effective monitoring and management of geological resources. For example, if an area with originally low contrast in a certain area shows an increase in contrast in subsequent images, it may indicate that new geological activities or human intervention (such as mining operations) have led to changes in geological structure and composition.

[0100] In an optional embodiment, obtaining an adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block includes the following steps:

[0101] The following formula is used to obtain the adjustment parameter of the high-frequency edge block based on the contrast and representative grayscale of the high-frequency edge block;

[0102]

[0103] Among them, R is the adjustment parameter of the high-frequency edge block, C is the contrast of the high-frequency edge block, and M is the representative grayscale of the high-frequency edge block;

[0104] Use the following formula to calculate the local variance of the high-frequency edge block;

[0105]

[0106] Among them, σ 2 is the local variance of the high-frequency edge block, N is the number of pixels in the high-frequency edge block, G k is the gray value of the kth pixel, is the mean gray value of pixels in the high-frequency edge block;

[0107] The adjustment parameters are dynamically adjusted according to the local variance to obtain adaptive adjustment parameters.

[0108] Specifically, the calculation formula of the adjustment parameter R is In the method, the adjustment parameter is determined by combining the contrast C of the high-frequency edge block with the representative grayscale M. The contrast C reflects the severity of the grayscale change in the edge block, and the representative grayscale M reflects the overall grayscale level of the edge block. The two are divided and added to their own proportional relationship, aiming to generate an adjustment parameter that can adapt to its grayscale change according to the characteristics of the edge block itself. For example, when the contrast C is large, it means that the edge feature is obvious. At this time, the adjustment parameter R will be relatively large, so as to pay more attention to the edge part in subsequent processing; and when the representative grayscale M is large, it means that the overall grayscale is high, and the adjustment parameter R will be relatively small to balance the influence of the overall grayscale on the edge enhancement effect.

[0109] In calculating the local variance σ 2 Formula In G, N is the number of pixels in the high-frequency edge block, which determines the number of samples involved in the calculation. k is the grayscale value of the kth pixel, representing the grayscale information of each individual pixel. It is the mean grayscale value of the pixels in the high-frequency edge block, which is a measure of the overall grayscale level of the edge block. The square of the deviation between the gray value of each pixel and the mean is calculated. The local variance σ is obtained by summing the square of the deviations of all pixels and taking the average (dividing by N). 2 The local variance reflects the degree of dispersion of the grayscale values ​​of pixels in the high-frequency edge block relative to the mean. The larger the variance, the more dispersed the distribution of the grayscale values ​​of the pixels, the more uneven the grayscale changes in the edge block, and the more detailed information may be contained; the smaller the variance, the more concentrated the grayscale distribution, and the grayscale in the edge block is relatively uniform.

[0110] σ 2 The local variance provides a basis for the dynamic adjustment of the adjustment parameter R. When the local variance is large, it means that the grayscale changes in the high-frequency edge block are rich and uneven, and may contain more important details or complex edge structures. At this time, the adjustment parameter R needs to be appropriately increased to enhance the processing effect of these details and edges. For example, it can be achieved by multiplying by a coefficient greater than 1 (the coefficient is related to the local variance). On the contrary, when the local variance is small, the grayscale in the edge block is relatively uniform, and it may not be necessary to overemphasize edge enhancement. At this time, the adjustment parameter R should be appropriately reduced, such as multiplying by a coefficient less than 1. In this way, the adjustment parameters can be flexibly adjusted according to the local characteristics of the high-frequency edge block, making it more suitable for different image areas, thereby achieving a more optimized geological resource area recognition effect overall.

[0111] In an optional embodiment, dynamically adjusting the adjustment parameter according to the local variance to obtain the adaptive adjustment parameter includes the following steps:

[0112] Use the following formula to obtain the weight coefficient of contrast and the weight coefficient representing grayscale based on local variance;

[0113]

[0114] ω2=1-ω1;

[0115] Among them, ω1 is the weight coefficient of contrast, ω2 is the weight coefficient representing grayscale, μ and σ0 are adjustable parameters;

[0116] Use the following formula to obtain the adaptive adjustment parameter based on the weight coefficient of contrast and the weight coefficient representing grayscale;

[0117]

[0118] Among them, R ω is an adaptive adjustment parameter.

[0119] Specifically, for the contrast weight coefficient Here, a calculation method based on exponential function is adopted. μ and σ0 are adjustable parameters, which are used to control the weight coefficient with the local variance σ2 The rate of change and threshold characteristics. When the local variance σ 2 As the exponential part -μ(σ 2 -σ0) will change, thus affecting the denominator The value of , thus changing the contrast weight coefficient ω1. 2 When ω1 is larger, ω1 approaches 1, which means that in this case, contrast will occupy a larger proportion in calculating the adaptive adjustment parameters, because a larger local variance usually indicates that the area contains richer edge and detail information, and more attention should be paid to contrast.

[0120] The weight coefficient representing grayscale ω2=1-ω1, which complements the contrast weight coefficient ω1. That is, as ω1 increases, ω2 will decrease accordingly, and vice versa. This design ensures that the sum of the contrast weight coefficient and the representative grayscale weight coefficient is always 1, so that when calculating the adaptive adjustment parameters, the two can reasonably allocate weights and adjust together according to the local characteristics of the high-frequency edge blocks.

[0121] Adaptive adjustment parameters The calculation is based on the contrast weight coefficient ω1, the representative grayscale weight coefficient ω2, and the contrast C and representative grayscale M of the high-frequency edge block obtained above. This formula uses the contrast C multiplied by its weight coefficient ω1 as the numerator, and the sum of the contrast C multiplied by its weight coefficient ω1 and the representative grayscale M multiplied by its weight coefficient ω2 as the denominator to obtain an adjustment parameter that can be dynamically adjusted according to the local characteristics of the high-frequency edge block (influencing the weight coefficient through the local variance). This calculation method enables the adaptive adjustment parameter to reasonably balance the impact of contrast and representative grayscale on the final result under different local variance conditions, thereby more accurately adapting to the characteristics of high-frequency edge blocks and providing more suitable basic parameters for subsequent calculations of sharpening parameters for each pixel.

[0122] In an optional embodiment, obtaining a sharpening value of each pixel in each high-frequency edge block according to the representative grayscales of all high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block includes:

[0123] Use the following formula to calculate the sharpness value of each pixel in each high-frequency edge block;

[0124]

[0125] Among them, Q t,k is the sharpening value of the kth pixel in the tth high-frequency edge block, A t,k is the sharpening parameter of the kth pixel in the tth high-frequency edge block, M t is the representative grayscale of the tth high-frequency edge block, M′ t,lis the representative grayscale of the lth neighboring high-frequency edge block of the tth high-frequency edge block, X t is the number of high-frequency edge blocks in the neighborhood of the t-th high-frequency edge block.

[0126] Specifically, the formula for calculating the sharpening value of each pixel in the high-frequency edge block is In, Q t,k A is the sharpening value of the kth pixel in the tth high-frequency edge block to be finally calculated, which reflects the degree of grayscale enhancement of the pixel after sharpening. t,k M is the sharpening parameter of the kth pixel in the tth high-frequency edge block. This sharpening parameter is calculated based on the adaptive adjustment parameters and is used to preliminarily determine the sharpening adjustment amount of the pixel. t is the representative grayscale of the tth high-frequency edge block, which reflects the overall grayscale level of the edge block. t,l is the representative grayscale of the lth neighboring high-frequency edge block of the tth high-frequency edge block. By considering the representative grayscale of the neighboring high-frequency edge block, the grayscale change information of the local area can be introduced, thereby adjusting the sharpening value of the pixel more comprehensively. t is the number of high-frequency edge blocks in the neighborhood of the t-th high-frequency edge block, which determines the size of the neighborhood range involved in the calculation.

[0127] The sharpening value calculated in this way can effectively enhance the edge features of pixels in high-frequency edge blocks. In geological resource monitoring, clear edge features are critical for accurately identifying the boundaries of geological resource areas and the contours of geological structures. For example, when identifying the boundary between a mineral layer and surrounding rock, after sharpening, the sharpening value of the pixels at the boundary increases, making the boundary more obvious, which helps to improve the accuracy of geological resource area identification.

[0128] The representative grayscale information of the neighborhood high-frequency edge blocks is introduced into the formula, so that the sharpening value of the pixel point depends not only on the characteristics of the high-frequency edge block where it is located, but also on the grayscale changes in the surrounding area. This helps to better integrate the information of the local area while maintaining the edge details, and avoid problems such as edge discontinuity or noise amplification caused by over-sharpening. In geological resource areas, there is often a certain degree of continuity and correlation. This sharpening method that considers local area information can more realistically reflect the actual distribution of geological resources and provide a more reliable image data basis for subsequent dynamic monitoring of geological resources. For example, when monitoring the changes in sedimentary rock strata over a large area, by considering the sharpening of neighborhood information, the continuous change characteristics of the strata can be more accurately displayed, which helps to discover potential clues to changes in geological resources.

[0129] In an embodiment of the present application, a method for dynamic monitoring of geological resources based on RS image enhancement is provided, wherein the method collects RS images through satellites, obtains high-frequency edge images through graying, multi-scale edge detection and fusion (including weighted fusion of edge images of different scales), and high-pass filtering, and then obtains high-frequency edge blocks from the high-frequency edge images, calculates parameters related to its grayscale mutation sequence (such as contrast, representative grayscale) to obtain adaptive adjustment parameters, and then obtains pixel point sharpening parameters and sharpening values ​​(comprehensive high-frequency edge blocks themselves and neighborhood information), and finally forms an edge sharpening image superimposed on the original RS image for geological resource regional identification. Its technical effect is that it can comprehensively and accurately capture the edge information in the RS image, fully consider the characteristics of different scales and local areas, effectively enhance the edge contrast, clearly present the boundaries of geological resource regions, geological structure contours, etc., and provide more accurate and reliable image data for dynamic monitoring of geological resources, which helps to improve the accuracy of geological resource identification and better monitor the distribution and changes of geological resources.

[0130] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0131] Based on the same inventive concept, the embodiment of the present application also provides a system for implementing the above-mentioned method for dynamic monitoring of geological resources based on RS image enhancement. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the geological resource dynamic monitoring system based on RS image enhancement provided below can refer to the above limitations on the method for dynamic monitoring of geological resources based on RS image enhancement, which will not be repeated here.

[0132] In an exemplary embodiment, Figure 2 As shown, a geological resource dynamic monitoring system 200 based on RS image enhancement is provided, comprising:

[0133] The high-frequency edge block generation module 201 is used to obtain the RS image, and obtain the high-frequency edge image of the RS image based on the multi-scale edge detection and fusion method; based on the high-frequency edge image, obtain the high-frequency edge block of the high-frequency edge image.

[0134] The grayscale mutation sequence construction module 202 is used to obtain the grayscale values ​​of all pixels of each high-frequency edge block, calculate the absolute value of the grayscale value difference between two horizontally adjacent pixels, and use it as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence.

[0135] The high-frequency edge block parameter calculation module 203 is used to obtain the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence; and calculate the grayscale value mean of all pixels in the high-frequency edge block as the representative grayscale of the high-frequency edge block.

[0136] The pixel sharpening parameter acquisition module 204 is used to obtain an adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block; and to obtain a sharpening parameter for each pixel in the high-frequency edge block by multiplying the grayscale value of each pixel in the high-frequency edge block by the adaptive adjustment parameter.

[0137] The pixel sharpening value calculation module 205 is used to obtain the sharpening value of each pixel in each high-frequency edge block according to the representative grayscale of all high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block; and take the gradient amplitude of the pixel outside the high-frequency edge block as its sharpening value.

[0138] The geological resource area acquisition module 206 is used to use the sharpening value of each pixel in all high-frequency edge blocks and the sharpening values ​​of all background pixels to form an edge-sharpened image; superimpose the edge-sharpened image on the original RS image, identify the geological resource area on the superimposed RS image, and obtain the geological resource area.

[0139] Optionally, the high-frequency edge block generation module includes:

[0140] The RS image acquisition and preliminary processing unit is used to collect RS images through satellites and grayscale the RS images to obtain grayscale remote sensing images.

[0141] The multi-scale edge detection unit is used to adopt a multi-scale edge detection method and use multiple edge detection operators of different scales to process the grayscale remote sensing image respectively to obtain edge images of different scales.

[0142] The edge image fusion unit is used to fuse edge images of different scales to obtain a fused edge image.

[0143] The filtering processing unit is used to filter the fused edge image through a high-pass filter to obtain a high-frequency edge image.

[0144] Optionally, the edge image fusion unit is specifically used to perform the following operations:

[0145] Assign weights according to the reliability or importance of edges at different scales, and perform weighted fusion on edge images at different scales according to the weights to obtain a fused edge image;

[0146] The formula for weighted fusion is:

[0147]

[0148] Among them, E fusion is the fused edge image, E i is the edge image of the i-th scale, n is the number of scales, ω i is the edge image E of the i-th scale i The weight coefficient of .

[0149] Optionally, the high-frequency edge block parameter calculation module includes a contrast calculation submodule, which is used to obtain the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence. Specifically, the contrast calculation submodule is used to perform the following operations:

[0150] The contrast of high-frequency edge blocks is calculated using the following formula:

[0151]

[0152] Among them, C is the contrast of high-frequency edge blocks, a j is the size of the jth grayscale mutation value, f j is the frequency of the jth grayscale mutation value, and m is the number of grayscale mutation values ​​in the grayscale mutation sequence.

[0153] Optionally, the pixel sharpening parameter acquisition module includes an adaptive adjustment parameter acquisition submodule, which is used to obtain the adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block. The adaptive adjustment parameter acquisition submodule includes:

[0154] An initial adjustment parameter acquisition unit, used to obtain an adjustment parameter of the high-frequency edge block based on the contrast and representative grayscale of the high-frequency edge block using the following formula;

[0155]

[0156] Among them, R is the adjustment parameter of the high-frequency edge block, C is the contrast of the high-frequency edge block, and M is the representative grayscale of the high-frequency edge block.

[0157] A high-frequency edge block local variance calculation unit, used to calculate the local variance of the high-frequency edge block using the following formula;

[0158]

[0159] Among them, σ 2 is the local variance of the high-frequency edge block, N is the number of pixels in the high-frequency edge block, G k is the gray value of the kth pixel, It is the mean gray value of pixels in the high-frequency edge block.

[0160] The adaptive adjustment parameter acquisition unit is used to dynamically adjust the adjustment parameter according to the local variance to obtain the adaptive adjustment parameter.

[0161] Optionally, the adaptive adjustment parameter acquisition unit is specifically used to perform the following operations:

[0162] Use the following formula to obtain the weight coefficient of contrast and the weight coefficient representing grayscale based on local variance;

[0163]

[0164] ω2=1-ω1;

[0165] Among them, ω1 is the weight coefficient of contrast, ω2 is the weight coefficient representing grayscale, μ and σ0 are adjustable parameters;

[0166] Use the following formula to obtain the adaptive adjustment parameter based on the weight coefficient of contrast and the weight coefficient representing grayscale;

[0167]

[0168] Among them, R ω is an adaptive adjustment parameter.

[0169] Optionally, the pixel sharpening value calculation module includes a high-frequency edge block pixel sharpening value calculation unit, which is used to calculate the sharpening value of each pixel in each high-frequency edge block using the following formula;

[0170]

[0171] Among them, Q t,k is the sharpening value of the kth pixel in the tth high-frequency edge block, A t,k is the sharpening parameter of the kth pixel in the tth high-frequency edge block, M t is the representative grayscale of the tth high-frequency edge block, M′ t,l is the representative grayscale of the lth neighboring high-frequency edge block of the tth high-frequency edge block, X tis the number of high-frequency edge blocks in the neighborhood of the t-th high-frequency edge block.

[0172] An embodiment of the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0173] The embodiments of the present application further provide a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0174] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0175] The above-mentioned embodiments only express several implementation methods of the embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the embodiments of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the embodiments of the present application, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A method for dynamic monitoring of geological resources based on RS image enhancement, characterized in that: The method comprises: Acquire a RS image, and obtain a high-frequency edge image of the RS image based on a multi-scale edge detection and fusion method; obtain a high-frequency edge block of the high-frequency edge image based on the high-frequency edge image; Obtaining the grayscale values ​​of all pixels of each high-frequency edge block, calculating the absolute value of the grayscale value difference between two horizontally adjacent pixels as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence; According to the frequency of each grayscale mutation value in the grayscale mutation sequence, the contrast of the high-frequency edge block is obtained; the average grayscale value of all pixels in the high-frequency edge block is calculated as the representative grayscale of the high-frequency edge block; According to the contrast and representative grayscale of the high-frequency edge block, an adaptive adjustment parameter is obtained; and the grayscale value of each pixel in the high-frequency edge block is multiplied by the adaptive adjustment parameter to obtain a sharpening parameter for each pixel in the high-frequency edge block; According to the representative grayscales of all the high-frequency edge blocks and the sharpening parameters of each pixel in each of the high-frequency edge blocks, a sharpening value of each pixel in each of the high-frequency edge blocks is obtained; and the gradient amplitude of the pixel outside the high-frequency edge block is used as its sharpening value; An edge-sharpened image is constructed using the sharpening value of each pixel in all the high-frequency edge blocks and the sharpening values ​​of all the background pixels; the edge-sharpened image is superimposed on the original RS image, and the geological resource area is identified on the superimposed RS image to obtain the geological resource area.

2. The method according to claim 1, characterized in that The RS image is acquired, and a high-frequency edge image of the RS image is obtained based on a multi-scale edge detection and fusion method, including: Collect RS images through satellites, and grayscale the RS images to obtain grayscale remote sensing images; A multi-scale edge detection method is adopted to process the grayscale remote sensing image respectively using edge detection operators of multiple different scales to obtain edge images of different scales; Fusing the edge images of different scales to obtain a fused edge image; The fused edge image is filtered using a high-pass filter to obtain the high-frequency edge image.

3. The method according to claim 2, characterized in that The step of fusing the edge images of different scales to obtain a fused edge image includes: Assigning weights according to the reliability or importance of the edges of different scales, and weighted fusion of the edge images of different scales according to the weights to obtain the fused edge image; The formula of weighted fusion is: Among them, E fusion is the fused edge image, E i is the edge image of the i-th scale, n is the number of scales, ω i is the edge image E of the i-th scale i The weight coefficient of .

4. The method according to claim 1, characterized in that: The step of obtaining the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence includes: The contrast of the high-frequency edge block is calculated by the following formula: Wherein, C is the contrast of the high-frequency edge block, a j is the size of the j-th grayscale mutation value, f j is the frequency of the jth grayscale mutation value, and m is the number of types of grayscale mutation values ​​in the grayscale mutation sequence.

5. The method according to claim 4, characterized in that The step of obtaining the adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block includes: Using the following formula, based on the contrast and representative grayscale of the high-frequency edge block, an adjustment parameter of the high-frequency edge block is obtained; Wherein, R is the adjustment parameter of the high-frequency edge block, C is the contrast of the high-frequency edge block, and M is the representative grayscale of the high-frequency edge block; The local variance of the high-frequency edge block is calculated using the following formula; Among them, σ 2 is the local variance of the high-frequency edge block, N is the number of pixels in the high-frequency edge block, G k is the gray value of the kth pixel, is the mean grayscale value of the pixels in the high-frequency edge block; The adjustment parameter is dynamically adjusted according to the local variance to obtain the adaptive adjustment parameter.

6. The method according to claim 5, characterized in that The dynamically adjusting the adjustment parameter according to the local variance to obtain the adaptive adjustment parameter includes: Using the following formula, based on the local variance, a weight coefficient of the contrast and a weight coefficient of the representative grayscale are obtained; ω2=1-ω1; Wherein, ω1 is the weight coefficient of the contrast, ω2 is the weight coefficient of the representative grayscale, and μ and σ0 are adjustable parameters; The adaptive adjustment parameter is obtained according to the weight coefficient of the contrast and the weight coefficient representing the grayscale using the following formula; Among them, R ω is the adaptive adjustment parameter.

7. The method according to claim 6, characterized in that The step of obtaining a sharpening value of each pixel in each high-frequency edge block according to the representative grayscales of all the high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block comprises: The sharpening value of each pixel in each high-frequency edge block is calculated using the following formula; Among them, Q t，k is the sharpening value of the kth pixel in the tth high-frequency edge block, A t，k is the sharpening parameter of the kth pixel in the tth high-frequency edge block, M t is the representative grayscale of the t-th high-frequency edge block, M′ t，l is the representative grayscale of the lth neighboring high-frequency edge block of the tth high-frequency edge block, X t is the number of neighboring high-frequency edge blocks of the tth high-frequency edge block.

8. A geological resource dynamic monitoring system based on RS image enhancement, characterized in that: The system comprises: A high-frequency edge block generation module is used to obtain an RS image, and obtain a high-frequency edge image of the RS image based on a multi-scale edge detection and fusion method; based on the high-frequency edge image, obtain a high-frequency edge block of the high-frequency edge image; A grayscale mutation sequence construction module is used to obtain the grayscale values ​​of all pixels of each high-frequency edge block, calculate the absolute value of the grayscale value difference between two horizontally adjacent pixels, and use it as the grayscale mutation value of the corresponding high-frequency edge block, thereby constructing a grayscale mutation sequence; A high-frequency edge block parameter calculation module is used to obtain the contrast of the high-frequency edge block according to the frequency of each grayscale mutation value in the grayscale mutation sequence; calculate the grayscale value mean of all pixels in the high-frequency edge block as the representative grayscale of the high-frequency edge block; A pixel sharpening parameter acquisition module is used to obtain an adaptive adjustment parameter according to the contrast and representative grayscale of the high-frequency edge block; and to obtain a sharpening parameter for each pixel in the high-frequency edge block by multiplying the grayscale value of each pixel in the high-frequency edge block by the adaptive adjustment parameter; A pixel sharpening value calculation module is used to obtain the sharpening value of each pixel in each high-frequency edge block according to the representative grayscale of all the high-frequency edge blocks and the sharpening parameter of each pixel in each high-frequency edge block; and take the gradient amplitude of the pixel outside the high-frequency edge block as its sharpening value; The geological resource area acquisition module is used to use the sharpening value of each pixel in all the high-frequency edge blocks and the sharpening values ​​of all the background pixels to form an edge-sharpened image; superimpose the edge-sharpened image on the original RS image, identify the geological resource area of ​​the superimposed RS image, and obtain the geological resource area.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images