Geological disaster exploration image enhancement method

Abstract

The invention relates to the technical field of image data processing, in particular to a geological disaster exploration image enhancement method, which comprises the following steps of: acquiring values of pixel points in R, G and B channels in a current mountain image and gray values of the pixel points; for a target pixel point in the pixel points in the current mountain image, determining a noise factor of the target pixel point; determining a landslide factor of the target pixel point; determining the number of reference historical mountain images of the target pixel points; determining the weight of the dark channel value of the pixel point at the position corresponding to the target pixel point in each frame of reference historical mountain image during weighting; obtaining an optimized dark channel value of the target pixel point; and enhancing the current mountain image by using a dark channel defogging algorithm based on the optimized dark channel value. According to the method, more accurate input is provided for a dark channel defogging algorithm by accurately determining the optimized dark channel value of the pixel point, so that the definition and the contrast ratio of the mountain image are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method for enhancing geological disaster exploration images. Background Art

[0002] Nearby landslides not only pose a serious threat to human life and property safety, but may also lead to traffic disruption, infrastructure damage, and ecological environment deterioration. Therefore, timely identification of mountain images in potential landslide areas is of great significance for quickly responding to the negative impact of landslide disasters.

[0003] In the process of collecting mountain images, due to the interference of environmental factors such as haze and clouds, the image quality is often reduced and key geological features are covered, thus affecting the accuracy and real-time performance of the analysis results. In this context, the dark channel dehazing algorithm, as an advanced image enhancement technology, has shown significant advantages in improving image clarity and highlighting potential landslide features. The dark channel dehazing algorithm effectively removes haze interference in the image, making the image clearer and the geological features more prominent, thereby providing accurate data support for the timely response to landslide disasters.

[0004] The dark channel dehazing algorithm will enhance the image based on the dark channel value of each pixel when processing the image. The dark channel value comes from the minimum value of all pixels in the vicinity of the pixel in the three channels of R, G, and B. However, during the transmission of mountain images, due to signal interference or transmission errors, image data is lost or abnormal, thus forming noise pixels in the image. Noise pixels usually have extremely small dark channel values, so the presence of noise pixels will reduce the accuracy of the dark channel value of a single pixel, thereby affecting the overall performance of the dark channel dehazing algorithm. Summary of the invention

[0005] In order to solve the problem that during the transmission of mountain images, image data is lost or abnormal due to signal interference or transmission error, thereby forming noise pixels in the image. Noise pixels usually have extremely small dark channel values, so the existence of noise pixels will reduce the accuracy of the dark channel value of a single pixel, thereby affecting the overall performance of the dark channel defogging algorithm. The present invention proposes a method for enhancing geological disaster exploration images, which includes the following steps: Obtain the values ​​of the pixels in the current mountain image in the R, G, and B channels and the grayscale values ​​of the pixels; for the target pixel in the current mountain image, determine the noise factor of the target pixel according to the grayscale values ​​of the target pixel and the pixels in the neighborhood, and the values ​​of the target pixel and the pixels in the neighborhood in each channel; determine the landslide factor of the target pixel according to the mean value of the target pixel and the pixels in the surrounding area in the R channel, the mean value of the pixels in the current mountain image in the R channel, and the standard deviation of the gradient direction values ​​of the target pixel and all the pixels with gradient changes in the surrounding area. ; According to the noise factor, determine the number of reference historical mountain images of the target pixel point; according to the landslide factor, the landslide factor of the pixel point at the position corresponding to the target pixel point in each frame of the reference historical mountain image, and the noise factor, determine the weight of the dark channel value of the pixel point at the position corresponding to the target pixel point in each frame of the reference historical mountain image; according to the weight, perform weighted summation on the dark channel values ​​of the pixel points at the position corresponding to the target pixel point in the historical mountain image to obtain the optimized dark channel value of the target pixel point; based on the optimized dark channel value, use the dark channel defogging algorithm to enhance the current mountain image.

[0006] The present invention can effectively identify and suppress the influence of noise pixels on dark channel values ​​by analyzing the noise factors of target pixels, thereby improving the quality and accuracy of image data; dynamically adjust the weight of the reference image in the weighted operation according to the landslide factor and the noise conditions of the historical image, so that when performing dark channel defogging, the changes in environmental conditions can be better considered, thereby improving the robustness of the algorithm; by introducing historical mountain images and combining them with multi-frame images for processing, the richness of data can be effectively increased, and the changes in mountain images in time series can be tracked, thereby more accurately obtaining the dark channel value of the target pixel; by optimizing the dark channel value, the processing effect of the dark channel defogging algorithm can be significantly improved, making the restoration of clear visual information more effective, thereby providing more realistic and reliable images in actual exploration and analysis.

[0007] Furthermore, the values ​​in the R, G, and B channels are obtained by performing RGB separation processing on the current mountain image.

[0008] Furthermore, the grayscale value is obtained by graying the current mountain image.

[0009] Furthermore, the noise factor satisfies: ; In the formula, is the pixel point in the current mountain image The noise factor, is the pixel point in the current mountain image The number of pixels in the neighborhood of is the pixel point in the current mountain image The gray value of is the pixel point in the current mountain image In the neighborhood of The gray value of a pixel, is the pixel point in the current mountain image The number of channels of R, G, and B channels, is the pixel point in the current mountain image In the R, G, B channels The values ​​in the channels, is the pixel point in the current mountain image In the neighborhood of The pixel point in the R, G, B channels The values ​​in the channels, is the standard normalization function.

[0010] The present invention can more accurately identify noise pixels by comprehensively considering the differences in grayscale values ​​and channel values ​​between the target pixel and all pixels in its neighborhood, thereby reducing misjudgments and improving the rigor of noise detection. The noise factor not only takes into account the differences in grayscale values, but also takes into account the data in the RGB channels, making noise identification more comprehensive and reducing the judgment errors caused by a single viewing angle, thereby better adapting to complex image environments.

[0011] Furthermore, the gradient direction value is obtained using a Sobel operator.

[0012] Furthermore, the landslide factor satisfies: ; In the formula, is the pixel point in the current mountain image The landslide factor, is the pixel point in the current mountain image The mean value of the pixel points in the surrounding area in the R channel, is the mean value of the pixel points in the current mountain image in the R channel, is the pixel point in the current mountain image The standard deviation of the gradient direction values ​​of all pixels with gradient changes in the surrounding area, is the first hyperparameter, is the standard normalization function.

[0013] The present invention can dynamically monitor the possible landslide characteristics of pixel points by comparing the R channel value of the target pixel point with the mean value of its surrounding area, which is convenient for quickly identifying key areas in geological disaster research; the calculation of the landslide factor utilizes the pixel gradient change, so that the algorithm can respond to instantaneous changes in real time, which helps to provide more accurate image information in a changing environment; by jointly considering the gray value mean and the standard deviation of the gradient direction, the landslide factor can more comprehensively reflect the structural characteristics in the area and improve the accuracy of feature extraction.

[0014] Furthermore, the number of the reference historical mountain images satisfies: ; In the formula, is the pixel point in the current mountain image The number of reference historical mountain images, is the preset initial quantity, is the pixel point in the current mountain image The noise factor, Is the rounding function.

[0015] The present invention incorporates the noise factor into the calculation of the number of reference historical mountain images, so that the number of reference images is dynamically adjusted according to the noise level; when the noise level is high, increasing the number of reference historical images helps to provide more supplementary information to compensate for the information that may be lost in the current image, and more historical data can make the restored image closer to the real situation and reduce distortion; through the product adjustment of the initial number and the noise factor, it is possible to ensure the image restoration quality while avoiding the computational burden caused by over-reliance on historical images.

[0016] Furthermore, the weights satisfy: ; In the formula, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, is the pixel point in the current mountain image The landslide factor, For the Frame reference historical mountain image and current mountain image pixel points The landslide factor of the pixel point at the corresponding position, For the Frame reference historical mountain image and current mountain image pixel points The noise factor of the pixel at the corresponding position, is the first hyperparameter, is the second hyperparameter, is the standard normalization function, is the absolute value symbol.

[0017] The weight calculation of the present invention comprehensively considers the difference between the landslide factor in the current mountain image and the landslide factor at the corresponding position in the historical image, as well as the noise factor, so that the weight can better reflect the similarity between the image and the historical data, so that the information similar to the current image can be more effectively highlighted in the weighting process; by introducing the tolerance of the landslide factor and the noise factor, the flexibility of the weighting process is enhanced, so that the reference historical information can be more accurately selected during the defogging process, thereby effectively improving the clarity and detail retention of the processed image; by incorporating the influence of the noise factor into the weight calculation, the erroneous weight distribution caused by noise can be effectively suppressed, ensuring that the noise influence is reduced during image enhancement, and making the image restoration more natural.

[0018] Furthermore, the dark channel value is the minimum value of the pixel point and all the pixel points in the adjacent range in the R, G, and B channels.

[0019] Furthermore, the optimized dark channel value satisfies: ; In the formula, is the pixel point in the current mountain image The optimized dark channel value of is the pixel point in the current mountain image The number of reference historical mountain images, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, For the Frame reference historical mountain image and current mountain image pixel points The dark channel value of the pixel at the corresponding position.

[0020] The optimized dark channel value of the present invention can effectively fuse image information from multiple times and different environments by weighted averaging corresponding pixel points in historical mountain images, reflect local features and global trends, and make the final image enhancement effect more accurate; the weighted average method of weights is used, so that more relevant and important information in historical images has a greater impact on the final result, improves the adaptability to influencing factors, and enhances the processing ability in complex scenes; because the noise factor is taken into account in the setting of weights, the sensitivity of the optimized dark channel value to noise is reduced, so that useful information can be better isolated in scenes with more noise, and the overall image quality is improved; combined with the information of multiple frames of images, in the dark channel defogging process, the details and colors in the image can be better restored, so that the final image is more vivid and realistic, which is particularly important in geological risk assessment; the optimized dark channel value provides better basic data support for the defogging algorithm, and its high accuracy and reliability lay a solid foundation for subsequent processing steps such as reverse color restoration and contrast enhancement.

[0021] The present invention has the following beneficial effects: (1) By determining the noise factor based on the grayscale values ​​and channel values ​​of the target pixel and the pixels in the neighborhood, the noise pixels in the image can be accurately identified. The number of reference historical mountain images is determined based on the noise factor, and then the dark channel values ​​of the pixels at the corresponding positions in the historical images are weighted. This can effectively reduce the influence of noise pixels on the accuracy of the dark channel value of a single pixel, avoid the interference of extremely small dark channel values ​​of noise pixels on the overall image, improve the reliability of dark channel value calculation, and enable the dark channel dehazing algorithm to run more stably in the presence of noise.

[0022] (2) Accurately determine the optimized dark channel value of the target pixel, providing a more accurate input for the dark channel dehazing algorithm. Since the adverse effects of noise pixels are removed, the dark channel dehazing algorithm can more accurately estimate the atmospheric light value and transmittance, thereby more effectively removing fog from mountain images and improving image clarity and contrast, making the enhanced mountain images more conducive to the exploration and analysis of geological disasters.

[0023] (3) The concept of landslide factor is introduced. The landslide factor is determined by comprehensively considering the mean value of the target pixel and the pixels in the surrounding area in the R channel, the mean value of the pixels in the current mountain image in the R channel, and the standard deviation of the gradient direction values ​​of the target pixel and all the pixels with gradient changes in the surrounding area. This not only helps to determine the characteristics of the area where the pixel is located, but also takes into account factors related to geological disasters such as landslides when weighting the dark channel values ​​of the pixels at the corresponding positions in the historical images, so that the enhanced image is more in line with the actual needs of geological disaster exploration and provides more valuable information for the analysis of geological disasters.

[0024] (4) The number of reference historical mountain images is determined according to the noise factor, and the weight of the dark channel value of the pixel at the corresponding position of each historical image is determined in combination with the landslide factor and the noise factor. This makes full use of the information of the historical images and can integrate the advantages of multiple frames of images to further improve the accuracy and stability of the dark channel value. At the same time, it reduces the impact caused by the loss or abnormality of single-frame image data and enhances the robustness of the image enhancement method.

[0025] (5) It can adapt to data loss or abnormal situations caused by signal interference or transmission errors during the transmission of mountain images, and can still effectively enhance the images in complex environments. Whether in the presence of noise interference or in the case of data anomalies, it can improve image quality through reasonable algorithm steps, provide reliable image data support for geological disaster exploration, and improve practicality and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The present invention is a flowchart of the steps of a method for enhancing geological disaster exploration images according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present invention are described clearly and completely below, and the described embodiments are part of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0029] See also Figure 1 , which shows a flowchart of a method for enhancing geological disaster exploration images provided by an embodiment of the present invention, the method comprising the following steps: S1: Get the values ​​of the pixels in the current mountain image in the R, G, B channels and the grayscale values ​​of the pixels.

[0030] It should be noted that the present invention uses a drone to continuously collect mountain images. Since the subsequent steps will use historical mountain images of the current mountain image, the first collected continuous multi-frame mountain images can be used only for reference without enhancement processing.

[0031] Implementers can set the number of consecutive multi-frame mountain images, for example, 10, according to specific implementation circumstances.

[0032] Specifically, the values ​​in the R, G, and B channels are obtained by performing RGB separation processing on the current mountain image.

[0033] Specifically, the grayscale value is obtained by graying the current mountain image.

[0034] S2: For a target pixel point among the pixels in the current mountain image, determine the noise factor of the target pixel point.

[0035] It should be noted that the noise factor of each pixel is obtained by analyzing the gray value performance of each pixel in the current mountain image. When analyzing the noise factor of each pixel in the current mountain image, the difference between the gray value of each pixel and the pixels in its neighborhood, the greater the difference, the greater the noise factor. However, when processing color images, the gray value analysis may not be able to fully reflect the noise situation. Therefore, continue to analyze the difference in the values ​​of each pixel in the current mountain image and the pixels in its neighborhood in the R, G, and B channels. The greater the difference, the greater the noise factor.

[0036] The noise factor of the target pixel is determined according to the grayscale values ​​of the target pixel and the pixels in the neighborhood, as well as the values ​​of the target pixel and the pixels in the neighborhood in each channel.

[0037] Specifically, the noise factor satisfies: ; In the formula, is the pixel point in the current mountain image The noise factor, is the pixel point in the current mountain image The number of pixels in the neighborhood of is the pixel point in the current mountain image The gray value of is the pixel point in the current mountain image In the neighborhood of The gray value of a pixel, is the pixel point in the current mountain image The number of channels of R, G, and B channels, is the pixel point in the current mountain image In the R, G, B channels The values ​​in the channels, is the pixel point in the current mountain image In the neighborhood of The pixel point in the R, G, B channels The values ​​in the channels, is the standard normalization function.

[0038] Implementers can set the range of the neighborhood according to specific implementation conditions, for example, 8 neighborhoods.

[0039] in, The larger the value is, the more pixels there are in the current mountain image. The greater the difference in grayscale value between the pixels in its 8-neighborhood, the greater the grayscale value of the pixels in the current mountain image. The more likely it is a noise pixel, the larger the noise factor; The larger the value is, the more pixels there are in the current mountain image. The greater the difference between the values ​​of the pixels in the R, G, and B channels in its 8-neighborhood, the greater the difference between the pixels in the current mountain image and the The greater the difference in grayscale value between the pixels in its 8-neighborhood, the greater the difference in grayscale value between the pixels in the current mountain image. The more likely a pixel is a noise pixel, the larger the noise factor.

[0040] S3: Determine the landslide factor of the target pixel.

[0041] It should be noted that, since the soil in the landslide area usually presents a redder tone, the pixel points in such areas will have larger values ​​in the R channel; therefore, the larger the value of the pixel points in the area around a pixel point in the R channel, the greater the possibility that the area around this pixel point belongs to the landslide area, and the greater the landslide factor of this pixel point; however, the pixel points of the normal land area exposed between the vegetation will also show the characteristics of large values ​​in the R channel, and the existence of this area will affect the judgment of the possibility that the area around a pixel point belongs to the landslide area; then, after further scene investigation, it was found that landslides are geological disasters caused by gravity and the decrease in the stability of soil or rock. In the area where landslides occur, the movement and change of soil or rock often proceeds in a specific direction, and the flow or displacement of this material will cause a specific texture extension direction to appear in the image; therefore, if the gradient direction of all pixels with gradient changes in the area around a pixel point shows a stronger consistency, it means that the area around this pixel point is more likely to belong to the landslide area, and the landslide factor of this pixel point will also be greater.

[0042] The landslide factor of the target pixel is determined based on the mean value of the target pixel and the pixels in the surrounding area in the R channel, the mean value of the pixels in the current mountain image in the R channel, and the standard deviation of the gradient direction values ​​of the target pixel and all pixels with gradient changes in the surrounding area.

[0043] Specifically, the gradient direction value is obtained using the Sobel operator.

[0044] Specifically, the landslide factor satisfies: ; In the formula, is the pixel point in the current mountain image The landslide factor, is the pixel point in the current mountain image The mean value of the pixel points in the surrounding area in the R channel, is the mean value of the pixel points in the current mountain image in the R channel, is the pixel point in the current mountain image The standard deviation of the gradient direction values ​​of all pixels with gradient changes in the surrounding area, is the first hyperparameter, is the standard normalization function.

[0045] The implementer can set the size of the surrounding area and the first hyperparameter according to the specific implementation situation. For example, the surrounding area is , the first hyperparameter is 0.001; the first hyperparameter exists to prevent When it is 0, the calculation result becomes meaningless.

[0046] in, The larger the value is, the more pixels there are in the current mountain image. The larger the relative value of all pixels in the R channel in the surrounding area of ​​the image, the larger the relative value of all pixels in the R channel in the image, the larger the relative value of all pixels in the current mountain image. The greater the possibility that the surrounding area belongs to the landslide area, the greater the pixel point in the current mountain image The landslide factor will also be larger; The smaller it is, the more pixels in the current mountain image The closer the gradient direction value is to all the pixels with gradient changes in the surrounding area, the better the pixel value in the current mountain image is. The stronger the consistency of the gradient direction with all the pixels with gradient changes in the surrounding area, the better the pixel in the current mountain image is. The greater the possibility that the surrounding area belongs to the landslide area, the greater the pixel point in the current mountain image The landslide factor will also be greater.

[0047] S4: Determine the number of reference historical mountain images of the target pixel point.

[0048] It should be noted that after obtaining the noise factor of each pixel in the current mountain image, the number of historical mountain images that need to be referenced when calculating the dark channel value of each pixel in the current mountain image is determined according to the noise factor.

[0049] The number of reference historical mountain images of the target pixel point is determined according to the noise factor.

[0050] Specifically, the number of the reference historical mountain images satisfies: ; In the formula, is the pixel point in the current mountain image The number of reference historical mountain images, is the preset initial quantity, is the pixel point in the current mountain image The noise factor, Is the rounding function.

[0051] Implementers can set the initial number, for example, 10, based on specific implementation circumstances.

[0052] in, The larger it is, the worse the accuracy of the dark channel value at the pixel point is, the greater the demand for historical mountain images for this pixel point will be, and the greater the number of reference historical mountain images for this pixel point will be.

[0053] S5: Determine the weight of the dark channel value of the pixel point at the position corresponding to the target pixel point in each frame of the reference historical mountain image.

[0054] It should be noted that the landslide factor reflects the possibility that a pixel point and the surrounding area belong to the landslide area. Therefore, the landslide factor can be used as the basis for weight calculation to a certain extent.

[0055] The weight of the dark channel value of the pixel point corresponding to the target pixel point in each frame of the reference historical mountain image is determined when weighting according to the landslide factor, the landslide factor of the pixel point corresponding to the target pixel point in each frame of the reference historical mountain image, and the noise factor.

[0056] Specifically, the weights satisfy: ; In the formula, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, is the pixel point in the current mountain image The landslide factor, For the Frame reference historical mountain image and current mountain image pixel points The landslide factor of the pixel point at the corresponding position, For the Frame reference historical mountain image and current mountain image pixel points The noise factor of the pixel at the corresponding position, is the first hyperparameter, is the second hyperparameter, is the standard normalization function, is the absolute value symbol.

[0057] The implementer can set the second hyperparameter according to the specific implementation situation, for example, 0.001; the first hyperparameter exists to prevent When it is 0, the calculation result becomes meaningless. The second hyperparameter exists to prevent When it is 0, the calculation result becomes meaningless.

[0058] in, The smaller it is, the more likely it is that the two pixels are in the same area of ​​the actual mountain. Then the reference value of the dark channel value of the pixel at the corresponding position in the historical mountain image of this frame is greater, and the weight of the dark channel value should be greater when weighting. The smaller it is, the more real the pixel point at the corresponding position in this frame of reference historical mountain image is, and the greater the weight of the dark channel value of the pixel point at the corresponding position in this frame of reference historical mountain image will be when weighted.

[0059] S6: Obtain the optimized dark channel value of the target pixel.

[0060] It should be noted that the optimized dark channel value of each pixel in the current mountain image is obtained by weighted calculation based on the difference in landslide factors between the pixels at corresponding positions in the current mountain image and all the reference historical mountain images, combined with the noise factors of the pixels at corresponding positions in all the reference historical mountain images.

[0061] The dark channel values ​​of the pixels at positions corresponding to the target pixel points in the historical mountain image are weighted and summed according to the weights to obtain the optimized dark channel value of the target pixel point.

[0062] Implementers can set the size of the neighborhood based on their specific implementation, for example, .

[0063] Specifically, the dark channel value is the minimum value of the pixel point and all the pixel points in the adjacent range in the R, G, and B channels.

[0064] Specifically, the optimized dark channel value satisfies: ; In the formula, is the pixel point in the current mountain image The optimized dark channel value of is the pixel point in the current mountain image The number of reference historical mountain images, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, For the Frame reference historical mountain image and current mountain image pixel points The dark channel value of the pixel at the corresponding position.

[0065] S7: Based on the optimized dark channel value, the current mountain image is enhanced using a dark channel defogging algorithm.

[0066] Implementers can set the constant in the calculation formula for estimating the transmittance map of each pixel in the dark channel dehazing algorithm according to specific implementation conditions, for example, 0.95.

[0067] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for enhancing geological disaster exploration images, characterized in that: include: Get the values ​​of the pixels in the current mountain image in the R, G, B channels and the grayscale values ​​of the pixels; For a target pixel in the current mountain image, the noise factor of the target pixel is determined according to the grayscale values ​​of the target pixel and the pixels in the neighborhood, and the values ​​of the target pixel and the pixels in the neighborhood in each channel; Determine the landslide factor of the target pixel point according to the mean value of the target pixel point and the pixels in the surrounding area in the R channel, the mean value of the pixels in the current mountain image in the R channel, and the standard deviation of the gradient direction values ​​of the target pixel point and all the pixels with gradient changes in the surrounding area; Determining the number of reference historical mountain images of target pixel points according to the noise factor; Determine the weight of the dark channel value of the pixel point at the position corresponding to the target pixel point in each frame of the reference historical mountain image according to the landslide factor, the landslide factor of the pixel point at the position corresponding to the target pixel point in each frame of the reference historical mountain image, and the noise factor when weighting; According to the weight, the dark channel values ​​of the pixel points at the positions corresponding to the target pixel points in the historical mountain image are weighted and summed to obtain the optimized dark channel value of the target pixel points; Based on the optimized dark channel value, the current mountain image is enhanced using a dark channel defogging algorithm.

2. A method for enhancing geological disaster exploration images according to claim 1, characterized in that: The values ​​in the R, G, and B channels are obtained by performing RGB separation processing on the current mountain image.

3. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The grayscale value is obtained by graying the current mountain image.

4. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The noise factor satisfies: ； In the formula, is the pixel point in the current mountain image The noise factor, is the pixel point in the current mountain image The number of pixels in the neighborhood of is the pixel point in the current mountain image The gray value of is the pixel point in the current mountain image In the neighborhood of The gray value of a pixel, is the pixel point in the current mountain image The number of channels of R, G, and B channels, is the pixel point in the current mountain image In the R, G, B channels The values ​​in the channels, is the pixel point in the current mountain image In the neighborhood of The pixel point in the R, G, B channels The values ​​in the channels, is the standard normalization function.

5. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The gradient direction value is obtained using the Sobel operator.

6. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The landslide factor satisfies: ； In the formula, is the pixel point in the current mountain image The landslide factor, is the pixel point in the current mountain image The mean value of the pixel values ​​in the surrounding area in the R channel, is the mean value of the pixel points in the current mountain image in the R channel, is the pixel point in the current mountain image The standard deviation of the gradient direction values ​​of all pixels with gradient changes in the surrounding area, is the first hyperparameter, is the standard normalization function.

7. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The number of reference historical mountain images satisfies: ； In the formula, is the pixel point in the current mountain image The number of reference historical mountain images, is the preset initial quantity, is the pixel point in the current mountain image The noise factor, Is the rounding function.

8. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The weights satisfy: ； In the formula, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, is the pixel point in the current mountain image The landslide factor, For the Frame reference historical mountain image and current mountain image pixel points The landslide factor of the pixel point at the corresponding position, For the Frame reference historical mountain image and current mountain image pixel points The noise factor of the pixel at the corresponding position, is the first hyperparameter, is the second hyperparameter, is the standard normalization function, is the absolute value symbol.

9. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The dark channel value is the minimum value of the pixel point and all the pixels in the adjacent range in the R, G, and B channels.

10. The method for enhancing geological disaster exploration images according to claim 1, characterized in that: The optimized dark channel value satisfies: ； In the formula, is the pixel point in the current mountain image The optimized dark channel value of is the pixel point in the current mountain image The number of reference historical mountain images, For the Frame reference historical mountain image and current mountain image pixel points The weight of the dark channel value of the pixel at the corresponding position during weighting, For the Frame reference historical mountain image and current mountain image pixel points The dark channel value of the pixel at the corresponding position.

Images