Underground engineering cavern surrounding rock fissure and blasthole identification method, system and device based on adaptive light compensation

Abstract

The application discloses a method, system and device for identifying cracks and blast holes of surrounding rock of underground engineering cavern based on adaptive light compensation, and relates to the technical field of computer vision identification. The technical scheme is as follows: a first image of the surrounding rock of the underground engineering cavern is acquired; the first image is clustered to obtain a clustering result of bright and dark areas of the first image; the bright and dark areas of the first image are subjected to light balance processing according to the clustering result to obtain a second image subjected to light balance; the second image is segmented by using an improved yolov10 model to obtain a probability graph of identification of cracks and blast holes of the surrounding rock of the underground engineering cavern; wherein the improved yolov10 model is specifically as follows: a Conv module layer of a backbone network of the yolov10 model is replaced by an SPDConv module layer; wherein a first C2F feature layer of the backbone network is connected with the SPDConv module layer, and features output by the SPDConv module layer are spliced with features output by a second C2F feature layer.

Description

Technical Field

[0001] This invention relates to the field of computer vision recognition technology for underground engineering, and more specifically, to a method, system, and device for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation. Background Technology

[0002] In the field of engineering geology, the detection of joints and fissures in the surrounding rock of underground caverns mainly relies on on-site identification by geological engineers. This method suffers from problems such as strong subjectivity, large errors, low efficiency, and difficulty in digitizing the data.

[0003] While computer vision and deep learning technologies have been introduced into the field of gap recognition, the following key issues still exist:

[0004] 1. Due to the complex environment of the cave, the rock wall images often show local overexposure caused by strong light or underexposure caused by shadows.

[0005] 2. After blasting, numerous blast holes, typically 5-10 cm in diameter and resembling actual fissures, remain on the rock face. Existing identification technologies are not optimized to handle blast hole interference, making them prone to being mistaken for fissures. Summary of the Invention

[0006] The purpose of this invention is to provide a method, system, and device for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation. This solves the problem that existing technologies are not accurate enough in identifying fissures in the surrounding rock walls of underground caverns, and easily misidentify blast holes as fissures.

[0007] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0008] In a first aspect, the present invention provides a method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, the method comprising:

[0009] Acquire the first image of the rock wall of the underground cavern;

[0010] Cluster the first image to obtain the clustering results of the bright and dark regions of the first image;

[0011] Based on the clustering results, the first image is subjected to illumination equalization processing on the bright and dark areas to obtain the second image after illumination equalization.

[0012] The improved YOLOv10 model is used to segment the second image to obtain a probability map for identifying rock fissures and blast holes in underground engineering caverns. Specifically, the improved YOLOv10 model involves replacing the Conv module layer of the YOLOv10 model backbone network with an SPDConv module layer. The first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

[0013] A second aspect of the present invention provides a system for identifying rock fissures and blast holes in underground engineering caverns based on adaptive illumination compensation, the system comprising:

[0014] The image acquisition module is used to acquire the first image of the surrounding rock of the underground engineering cavern;

[0015] The image clustering module is used to cluster the first image to obtain the clustering results of the bright and dark regions of the first image;

[0016] The image processing module is used to perform illumination equalization processing on the bright and dark areas of the first image based on the clustering results to obtain the second image after illumination equalization.

[0017] A fracture identification module is used to segment the second image using an improved YOLOv10 model to obtain a probability map of fractures and blast holes in the surrounding rock of underground engineering caverns. Specifically, the improved YOLOv10 model replaces the Conv module layer of the backbone network of the YOLOv10 model with an SPDConv module layer. The first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

[0018] A third aspect of the present invention provides an electronic device, including a memory and a processor;

[0019] A memory for storing computer programs, the computer programs including program instructions;

[0020] A processor is configured to execute the program instructions to cause the electronic device to perform the steps of a method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, as provided in the first aspect of the invention.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] This invention first performs clustering processing on the first image to divide the bright and dark regions in the first image. Based on this clustering, adaptive illumination compensation is performed for the partitions, dynamically adjusting the enhancement parameters for different brightness regions to eliminate the interference of highlight or shadow regions on the crack features and ensure the stability of crack and borehole identification. Second, this invention improves the original YOLOv10 backbone network and detection head: the first C2F feature layer is passed through the SPDConv module layer to obtain small target information features, which are then concatenated and fused with the features output by the second C3F feature layer. The convolution module of the improved YOLOv10 model is the SPDConv module layer. By replacing the stride convolution and pooling sampling layers of the original YOLOv10, the SPDConv module layer reduces information loss and enhances the extraction of important details, thereby improving the recognition performance of the recognition model when processing small targets on rock walls and low-resolution images. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0024] Figure 1 A flowchart illustrating the method for identifying rock fissures and blast holes in underground engineering caverns based on adaptive illumination compensation, provided in an embodiment of the present invention.

[0025] Figure 2 The network structure diagram of the improved YOLOv10 model provided in the embodiments of the present invention;

[0026] Figure 3 A schematic diagram illustrating the clustering results and illumination equalization processing of the first image provided in an embodiment of the present invention;

[0027] Figure 4 The first image provided for the embodiment of the present invention is a schematic diagram of the division of the surrounding rock fissures (0) and blast holes (1) in the underground engineering cavern. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0029] It should be noted that the terms "comprising" or "may include" used in the various embodiments of this application indicate the presence of the claimed function, operation, or element, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in the various embodiments of this application, the terms "comprising," "having," and their cognates are intended only to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or adding one or more combinations of the foregoing.

[0030] It should be understood that terms such as "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0031] like Figure 1 As shown, this embodiment of the invention provides a method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation. The method includes:

[0032] S101, acquire the first image of the surrounding rock of the underground engineering cavern.

[0033] Specifically, the first image is a visible light image, which can be acquired by camera components carried by shooting equipment, monitoring equipment, etc., deployed on the ground of the underground cavern. This is common knowledge in the technical field, and this embodiment will not describe it in detail.

[0034] S102, cluster the first image to obtain the clustering results of the bright and dark regions of the first image.

[0035] In this embodiment, the clustering process is as follows: First, the first image is converted into the LAB color space, and the luminance channel is separated from the LAB color space.

[0036] It should be noted that only the luminance channel of the first image is retained to avoid the color information of the color channels interfering with the luminance analysis.

[0037] Then, K-means clustering is performed on the pixel values ​​of the brightness channel based on the pixel brightness value and pixel coordinates to obtain the clustering results of the bright and dark regions of the first image; wherein, the clustering results include overexposed areas, normal areas and underexposed areas.

[0038] Specifically, the clustering features of K-means clustering include pixel brightness value and pixel coordinates; therefore, spatially constrained clustering (k=3) is performed on the L channel pixel values ​​based on the clustering features to obtain three clustering results: overexposed area (high brightness), normal area (medium brightness), and underexposed area (low brightness).

[0039] The objective function for K-means clustering is set as follows: ,in, For pixels The luminance values ​​(L channel) are μi and ci, respectively, representing the luminance center and spatial center of the i-th class.

[0040] The execution flow of the K-means clustering algorithm is as follows:

[0041] Randomly select k initial cluster centers.

[0042] For each pixel x, calculate its distance to each cluster center. Assign it to the class with the smallest distance. Recalculate the center of each class: The iteration ends when the center point moves less than the threshold ϵ = 0.01 or the maximum number of iterations (100 times) is reached.

[0043] Finally, the segmentation results of the bright and dark regions of the first image after clustering are as follows: Figure 3 As shown.

[0044] S103, perform illumination equalization processing on the bright and dark areas of the first image based on the clustering results to obtain the second image after illumination equalization.

[0045] In this embodiment, if the clustering result is an overexposed area, a highlight suppression algorithm is applied to the overexposed area to reduce local brightness through diffusion of neighboring pixels. The algorithm steps are as follows:

[0046] The attenuation coefficient α of the exposure area is set to control the repair intensity. If the overexposed pixel is extremely different from the surrounding normal pixels, it is an extreme highlight and requires strong repair (α→1); if the difference is small (such as slight overexposure), then weak repair is required (α→0, retaining some of the original brightness).

[0047] The relative exposure deviation δ(x,y) is determined based on the mean and standard deviation of the normal zone: ,in, This represents the mean of the normal range. The standard deviation of the normal region. To prevent division by zero.

[0048] A dynamic mapping function for the attenuation coefficient is constructed based on the relative exposure deviation: ,in, To control the slope, k=2.0 was used for the images in the cavern environment. For the offset, take the image images in the cavern environment. =2.0

[0049] By combining the mean value of the neighborhood range and the brightness value of the overexposed area using a dynamic mapping function, the pixel value of the overexposed pixels is adjusted to complete the illumination equalization processing of the overexposed area. For example, for each pixel in the overexposed area... In its Within the neighborhood of the pixel, collect the brightness values ​​{I_neighbors} of the non-overexposed pixels, and calculate the neighborhood mean or median. Adjust the pixel values ​​of the overexposed areas: ,in, express.

[0050] If the clustering result indicates an underexposed region, then the luminance component I (Lab's L channel) is extracted from the underexposed region. A multi-scale Gaussian filter is used to estimate the illumination components. : The recommended Gaussian kernel size for multi-scale Gaussian filters is: small=15, medium=30, large=60.

[0051] Calculation of reflection components:

[0052] The weights wk are distributed according to the scale (e.g., [0.3, 0.5, 0.2]), and ε = 1e-6 to prevent division by zero.

[0053] Adaptively adjust the Gamma value based on local contrast:

[0054] , The standard deviation is the value within a 5x5 window.

[0055] Applying Gamma correction: The enhanced luminance channel is merged with the original chroma channel and converted back to RGB space.

[0056] The original texture is preserved in the normal area, with only slight contrast optimization:

[0057] First, calculate a local histogram (8×8 blocks) for the normal region, crop the top 1% of pixel values ​​from the histogram, and redistribute them to other regions.

[0058] Then, local contrast stretching is performed, which involves calculating the minimum / maximum value for each pixel's neighborhood (3×3):

[0059] And limit the tensile strength of contrast stretching: .

[0060] Based on the illumination compensation processing for overexposed, normal, and underexposed areas described in this embodiment, the resulting second image is as follows: Figure 3 The right-hand area is shown.

[0061] S104, the improved YOLOv10 model is used to segment the second image to obtain the identification probability map of the surrounding rock fissures and blast holes in the underground engineering cavern; wherein, the improved YOLOv10 model is specifically: the Conv module layer of the backbone network of the YOLOv10 model is replaced with the SPDConv module layer; wherein, the first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

[0062] In this embodiment, before segmenting the second image using the improved YOLOv10 model, the method further includes: using a nonlocal means denoising algorithm to denoise the second image, reducing the impact of dust and water stains on the rock wall surface on subsequent crack and borehole identification. The specific steps are as follows:

[0063] Obtain the three channels of the second image, and normalize the pixel values ​​of each channel to the range [0,1].

[0064] Search window radius (search_radius, 10-20 pixels), neighborhood patch radius (patch_radius, 3-5 pixels), and filter strength parameter (h, which controls the smoothness and is generally 0.1-0.3 times the pixel value range).

[0065] For each pixel in the image Within the radius of the window centered on it, traverse all possible neighboring pixels. .

[0066] Calculate two neighboring blocks (with and Centered on, with a size of ( Similarity weights for )

[0067] ,in Therefore The pixel value vector of the neighborhood block centered on the center.

[0068] Re-normalization of ownership: ;

[0069] The weighted average yields the denoised pixel values: ;

[0070] The denoised image is scaled back to the original pixel range (e.g., [0, 255]). Figure 4 As shown, this invention uses YOLOv11 to segment the image, labeling boreholes and fissures with different label values, for example, a label value of 0 represents a borehole and a label value of 1 represents a fissure.

[0071] To address the small features of structural joints, cracks, and boreholes, the original YOLOv10 backbone network and detection head are improved: the first C2F feature layer is passed through an SPDConv module layer to obtain small target information features, which are then concatenated and fused with the features output from the second C3F feature layer. This effectively learns feature representations from global to local levels, ultimately improving the detection performance of small targets. The improved YOLOv10 model is shown below. Figure 2 As shown, compared with the initial YOLOv10, the improved YOLOv10 model uses the SPDConv module layer for convolution. By replacing the traditional stride convolution and pooling sampling methods, SPDConv reduces information loss and enhances the extraction of important details, thereby improving the model's performance when processing small targets and low-resolution images.

[0072] Loss function: CIoU Loss is used to optimize bounding box regression, which considers the overlap between the predicted and ground truth boxes, the distance between their centers, and the consistency of their aspect ratios. Therefore, the loss function of the improved YOLOv10 model is: ;in,( The coordinates () represent the center points of the predicted bounding box and the ground truth bounding box. IoU measures the degree of overlap between the predicted and ground truth bounding boxes. It is the Euclidean distance, where C is the diagonal length of the smallest bounding box. A coefficient used to measure the consistency of aspect ratio. It is the weighting coefficient.

[0073] Since the blast holes are manually excavated, they exhibit a clear distribution pattern compared to the fissures in the underground rock wall structure. A single blast hole typically displays a locally parallel double-line structure, and the blast holes are also relatively parallel to each other. Utilizing the visual distribution pattern of the blast holes, the identification of fissures can be further corrected. Therefore, this embodiment also provides a method for verifying the geometric patterns of blast holes and fissures. Specifically, it verifies each marked area in the identification probability map according to preset feature verification rules for both blast holes and fissures, determining whether a blast hole is identified as a fissure or a fissure is identified as a blast hole.

[0074] The steps are as follows: Perform a closing operation on the identified boreholes and fissures, first expanding and then corroding:

[0075] ,in, For inputting a binary mask for boreholes or fissures, B is a kernel element. and These represent expansion and corrosion, respectively.

[0076] Geometric Feature Analysis of Connected Components: Defining Aspect Ratio: Because the joints, fissures, and blast holes are all elongated, the screening... The output characteristics are joints, fissures, or boreholes.

[0077] Define compactness: Where A is the area of ​​the connected region, and since the edges of joints and fissures are irregular, the necessary condition for determining a fissure is: .

[0078] Hough line detection: The Hough transform maps the image space (x,y) to the parameter space (ρ,θ). The necessary condition for detecting and identifying a borehole is that it is located on two straight lines. and .

[0079] Directional consistency check: For the output borehole features, extract the coordinate set of pixels in each segmented region:

[0080] ,in Let be the coordinates of the i-th pixel;

[0081] Calculate the mean vector and center the data: ; as well as

[0082] The covariance matrix can be further obtained by combining the centralized data:

[0083] ,in: ; .

[0084] Find the eigenvalues ​​and eigenvectors of the covariance matrix: The eigenvector corresponding to the largest eigenvalue This refers to the main direction of the blast hole.

[0085] Calculation of principal direction angle: For the principal direction angle of all boreholes Perform cluster analysis and calculate the mean of the main borehole direction for each category: For the a-th borehole of this category, if the following conditions are met: If the area is misclassified, it is determined to be a fissure and needs to be reclassified.

[0086] This invention also provides a system for identifying rock fissures and blast holes in underground engineering caverns based on adaptive illumination compensation. The system includes:

[0087] The image acquisition module is used to acquire the first image of the surrounding rock of the underground engineering cavern;

[0088] The image clustering module is used to cluster the first image to obtain the clustering results of the bright and dark regions of the first image;

[0089] The image processing module is used to perform illumination equalization processing on the bright and dark areas of the first image based on the clustering results to obtain the second image after illumination equalization.

[0090] A fracture identification module is used to segment the second image using an improved YOLOv10 model to obtain a probability map of fractures and blast holes in the surrounding rock of underground engineering caverns. Specifically, the improved YOLOv10 model replaces the Conv module layer of the backbone network of the YOLOv10 model with an SPDConv module layer. The first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

[0091] This invention provides a system for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, which is similar to the above-mentioned... Figure 1 The method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, as shown above, is a technical solution based on the same inventive concept. Through the detailed description of the method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation provided in the above embodiments, those skilled in the art can clearly understand the implementation process of each module included in the system for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0092] Accordingly, the recognition system provided in this embodiment first performs clustering processing on the first image to divide the bright and dark regions in the first image. Based on this clustering division, adaptive illumination compensation is performed for the partitions, and enhancement parameters are dynamically adjusted for different brightness regions to eliminate the interference of highlight or shadow regions on the crack features and ensure the stability of crack recognition. Secondly, the present invention improves the original YOLOv10 backbone network and detection head: the first C2F feature layer is passed through the SPDConv module layer to obtain small target information features, and then concatenated and fused with the features output by the second C3F feature layer. The convolution module of the improved YOLOv10 model is the SPDConv module layer. By replacing the stride convolution and pooling sampling layers of the original YOLOv10, the SPDConv module layer reduces information loss and enhances the extraction of important details, thereby improving the recognition performance of the recognition model when processing small targets on rock walls and low-resolution images.

[0093] This invention also provides an electronic device. The electronic device includes a processor, a memory, a communication interface, and at least one communication bus for connecting the processor, the memory, and the communication interface. The memory includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (PROM), or portable read-only memory (CD-ROM), and is used for related instructions and data.

[0094] The communication interface is used to receive and send data. The processor can be one or more CPUs; if the processor is a single CPU, it can be a single-core CPU or a multi-core CPU. The processor in the electronic device is used to read one or more programs stored in the memory and perform the following operations: acquire a first image of the surrounding rock of an underground engineering cavern; cluster the first image to obtain clustering results of the bright and dark regions of the first image; perform illumination equalization processing on the bright and dark regions of the first image based on the clustering results to obtain a second image after illumination equalization; segment the second image using an improved YOLOv10 model to obtain a probability map for identifying fissures and blast holes in the surrounding rock of the underground engineering cavern; wherein, the improved YOLOv10 model specifically involves replacing the Conv module layer of the YOLOv10 model backbone network with an SPDConv module layer; wherein, the first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

[0095] It should be noted that the specific implementation of each operation can be described above. Figure 1 The corresponding description of the method embodiments shown indicates that the electronic device can be used to execute the method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, as described in the above-described method embodiments of this application, and will not be elaborated further here.

[0096] This invention also provides a computer-readable storage medium, which is a memory device in a computer device for storing programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the operating system of the terminal. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the method for identifying surrounding rock fissures and blast holes in underground engineering caverns based on adaptive illumination compensation in the above embodiments. Those skilled in the art should understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0097] This invention also provides a computer program product containing program instructions. The computer program product can be software or program products containing program instructions, capable of running on a computing device or stored on any usable medium. When the computer program product runs on at least one electronic device, it causes the at least one electronic device to perform a method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation.

[0098] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, characterized in that, The methods include: Obtain the first image of the surrounding rock of the underground engineering cavern; Cluster the first image to obtain the clustering results of the bright and dark regions of the first image; Based on the clustering results, the first image is subjected to illumination equalization processing on the bright and dark areas to obtain the second image after illumination equalization. The improved YOLOv10 model is used to segment the second image to obtain a probability map for identifying rock fissures and blast holes in underground engineering caverns. Specifically, the improved YOLOv10 model involves replacing the Conv module layer of the YOLOv10 model backbone network with an SPDConv module layer. The first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

2. The method according to claim 1, characterized in that, Clustering is performed on the first image to obtain the clustering results of the bright and dark regions of the first image, including: The first image is converted to the LAB color space, and the luminance channel is separated from the LAB color space; K-means clustering is performed on the pixel values ​​of the brightness channel based on the pixel brightness value and pixel coordinates to obtain the clustering results of the bright and dark regions of the first image; wherein, the clustering results include overexposed areas, normal areas and underexposed areas.

3. The method according to claim 2, characterized in that, If the clustering result is an overexposed area, then the illumination equalization processing of the bright and dark areas of the first image based on the clustering result specifically includes: The relative exposure deviation is determined based on the mean and standard deviation of the normal zone; A dynamic mapping function for the attenuation coefficient is constructed based on the relative exposure deviation. Within the neighborhood of each pixel in the overexposed area, the brightness values ​​of non-overexposed pixels are collected, and the average value of the neighborhood is calculated based on the brightness values ​​of the non-overexposed pixels. By combining the mean value of the neighborhood range and the brightness value of the overexposed area using a dynamic mapping function, the pixel value of the overexposed pixel is adjusted to complete the illumination equalization processing of the overexposed area.

4. The method according to claim 2, characterized in that, If the clustering result is a normal region, then the illumination equalization processing of the bright and dark regions of the first image based on the clustering result specifically includes: Calculate the local histogram of the normal area and crop the pixel values ​​of the top region of the local histogram to complete the illumination equalization processing of the normal area.

5. The method according to claim 2, characterized in that, If the clustering result is an underexposed area, then the illumination equalization processing of the bright and dark areas of the first image based on the clustering result specifically includes: Extract the brightness channel components of the underexposed areas, and estimate the illumination components of the brightness channel components using a multi-scale Gaussian filter. The reflection component is calculated based on the illumination component; Combined with local contrast correction of the reflection component; The corrected reflection component is superimposed on the luminance channel component to complete the illumination equalization process for the underexposed area.

6. The method according to claim 1, characterized in that, Before segmenting the second image using the improved YOLOv10 model, the method further includes: denoising the second image using a nonlocal mean denoising algorithm; Different label values ​​are set for the cracks and boreholes in the denoised second image.

7. The method according to claim 1, characterized in that, The loss function of the improved YOLOv10 model is: ;in,( () represents the center coordinates of the predicted bounding box and the ground truth bounding box. IoU measures the degree of overlap between the predicted bounding box and the ground truth bounding box. It is the Euclidean distance, where C is the diagonal length of the smallest bounding box. A coefficient used to measure the consistency of aspect ratio. It is the weighting coefficient.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: The identification probability map is verified according to the preset feature verification rules for each type of borehole and fracture to determine whether a borehole is identified as a fracture or a fracture is identified as a borehole.

9. A system for identifying fissures and blast holes in the surrounding rock of underground engineering caverns based on adaptive illumination compensation, characterized in that, The system includes: The image acquisition module is used to acquire the first image of the surrounding rock of the underground engineering cavern; The image clustering module is used to cluster the first image to obtain the clustering results of the bright and dark regions of the first image; The image processing module is used to perform illumination equalization processing on the bright and dark areas of the first image based on the clustering results to obtain the second image after illumination equalization. A fracture identification module is used to segment the second image using an improved YOLOv10 model to obtain a probability map of fractures and blast holes in the surrounding rock of underground engineering caverns. Specifically, the improved YOLOv10 model replaces the Conv module layer of the backbone network of the YOLOv10 model with an SPDConv module layer. The first C2F feature layer of the backbone network is connected to the SPDConv module layer, and the features output by the SPDConv module layer are concatenated with the features output by the second C2F feature layer.

10. An electronic device, characterized in that, Including memory and processor; A memory for storing computer programs, the computer programs including program instructions; A processor is configured to execute the program instructions to cause the electronic device to perform the steps of a method for identifying surrounding rock fissures and blast holes in underground engineering caverns based on adaptive illumination compensation, as described in any one of claims 1 to 8.

Images