A tunnel safety monitoring system and method based on edge computing

Abstract

The application provides a tunnel safety monitoring system and method based on edge computing, which is used for tunnel safety monitoring of the tunnel safety monitoring system, and the tunnel safety monitoring system comprises sensors, edge computing nodes and a cloud platform, the sensors comprise displacement sensors based on Beidou technology and image acquisition devices based on radar vision technology, the image acquisition devices based on radar vision technology are used for continuously collecting tunnel inner wall images of a target tunnel in a safety monitoring process, and structural edge features of a target tunnel inner wall region are extracted from the tunnel inner wall images according to edge convolution kernels in each angle direction; the tunnel inner wall images are fused according to the structural edge features and gray projection features to obtain a recognition feature map; and safety hazards of the target tunnel are analyzed according to the recognition feature map. The above scheme can identify the detailed features of the tunnel inner wall images under the influence of the complex distribution of the surface defects of the tunnel inner wall and the uneven distribution of light, so as to timely perform tunnel safety alarm.

Description

Technical Field

[0001] This application relates to the field of tunnel safety monitoring technology, and more specifically, to a tunnel safety monitoring system and method based on edge computing. Background Technology

[0002] With the rapid development of the Internet of Things (IoT) and the Industrial Internet, the amount of data is growing explosively. Traditional cloud computing architectures can no longer meet the demands for real-time performance, bandwidth, and privacy protection. Edge computing has emerged to address this need. It deploys computing, storage, and network resources closer to the data source or user terminal, processing local data at edge nodes to reduce dependence on the cloud and significantly reduce data transmission latency and bandwidth consumption. At the same time, edge computing can enhance data privacy protection, avoiding the risks of storing and processing sensitive data centrally in the cloud. In addition, edge computing supports distributed intelligent decision-making, improving system reliability and autonomy, and providing key support for achieving efficient, secure, and intelligent digital transformation.

[0003] In existing edge computing, when data is generated, edge nodes first preprocess the data, such as data cleaning, compression, and feature extraction, to reduce transmission volume and improve efficiency. Subsequently, locally deployed algorithms (such as machine learning models or rule engines) perform real-time analysis and decision-making on the data, and upload key results or a small amount of data to the cloud. However, in tunnel safety monitoring, during the preprocessing of tunnel inner wall images by edge computing, the complex distribution of surface defects and uneven illumination make it difficult for traditional image processing methods to accurately extract detailed features (such as texture and edges) of the tunnel inner wall images, resulting in a decrease in the accuracy of target tunnel safety assessment. Therefore, how to identify detailed features of tunnel inner wall images under the influence of complex surface defects and uneven illumination to enable timely tunnel safety alarms has become a challenge for the industry. Summary of the Invention

[0004] This application provides a tunnel safety monitoring system and method based on edge computing, which can identify detailed features of tunnel inner wall images under the influence of complex distribution of defects on the tunnel inner wall surface and uneven illumination distribution, so as to promptly issue tunnel safety alarms.

[0005] In a first aspect, this application provides a tunnel safety monitoring method based on edge computing, used by a tunnel safety monitoring system for tunnel safety monitoring. The tunnel safety monitoring system includes sensors, edge computing nodes, and a cloud platform. The sensors are deployed at various monitoring points in the target tunnel. The method includes the following steps:

[0006] Images of the tunnel interior walls are collected by sensors deployed at various monitoring points in the target tunnel, and the tunnel interior wall images are transmitted to the edge computing node in real time.

[0007] The edge convolution kernels at different angles are determined when the edge computing node extracts the edges of the tunnel inner wall image. Based on the edge convolution kernels at each angle, the inner wall image of the tunnel inner wall is identified in multiple directions to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state.

[0008] Determine the grayscale compression coefficient of the tunnel inner wall image in the luminance channel, and perform projection compression on the tunnel inner wall image based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions.

[0009] Heterogeneous feature fusion is performed on the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain the identification feature map of the inner wall region when the target tunnel is identified for security purposes.

[0010] Based on the identified feature map, the safety hazards of the inner wall of the target tunnel are analyzed, and the edge computing node outputs the safety hazard information of the inner wall of the target tunnel to the cloud platform.

[0011] In some embodiments, determining the edge convolution kernels at different angles when the edge computing node extracts edges from the tunnel wall image specifically includes:

[0012] Acquire historical crack monitoring data of the target tunnel;

[0013] Based on the historical crack monitoring data, the distribution of crack directions in the target tunnel is predicted to obtain the predicted distribution probability of different crack directions in the target tunnel.

[0014] The predicted distribution entropy of the target tunnel cracks is calculated by using the predicted distribution probability of each crack direction.

[0015] The edge computing node calibrates the angle direction when performing edge extraction on the tunnel inner wall image based on the predicted distribution entropy, and obtains multiple different angle directions;

[0016] By setting edge convolution kernels in different angle directions based on the edge detection operator, edge convolution kernels in different angle directions can be obtained when extracting edges from the tunnel inner wall image.

[0017] In some embodiments, multi-directional inner wall edge recognition is performed on the tunnel inner wall image based on edge convolution kernels in various angular directions to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state, specifically including:

[0018] Extract the convolutional blocks corresponding to each pixel in the tunnel inner wall image;

[0019] Select a pixel as the selected pixel, and convolve the convolution block of the selected pixel with the edge convolution kernels in each angular direction to obtain the convolution pixel of the selected pixel;

[0020] Continue to determine the convolution pixels for the remaining pixels;

[0021] An edge detail map corresponding to the tunnel inner wall image is constructed based on all convolutional pixels;

[0022] The structural edge features of the target tunnel inner wall region under the current monitoring state are identified from the edge detail map.

[0023] In some embodiments, determining the grayscale compression coefficient of the tunnel inner wall image in the luminance channel specifically includes:

[0024] Convert the tunnel inner wall image into a grayscale inner wall image;

[0025] Calculate the brightness change rate sequence corresponding to the grayscale inner wall image;

[0026] The grayscale compression coefficient of the tunnel inner wall image in the brightness channel is determined by the brightness change rate sequence.

[0027] In some embodiments, the projection compression of the tunnel inner wall image based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions specifically includes:

[0028] Obtain the grayscale data of the tunnel inner wall image in the luminance channel;

[0029] The grayscale statistical parameters of the brightness channel are determined based on the grayscale data;

[0030] An adaptive adjustment factor is set based on the grayscale statistical parameters when projecting and compressing the tunnel inner wall image.

[0031] The grayscale of each pixel in the brightness channel of the tunnel inner wall image is projected and compressed using the adaptive adjustment factor and the grayscale compression coefficient to obtain the brightness projection compression map corresponding to the tunnel inner wall image.

[0032] The grayscale projection features of the target tunnel inner wall region under the current lighting conditions are extracted from the brightness projection compression map.

[0033] In some embodiments, the sensor includes: a displacement sensor based on BeiDou technology and an image acquisition device based on Ravage technology.

[0034] In some embodiments, the displacement sensor based on BeiDou technology is deployed at key parts of the tunnel to acquire high-precision displacement data. The key parts of the tunnel include: the arch, the arch waist, and the sidewalls. The image acquisition device based on radar vision technology is installed at various monitoring points on the inner wall of the tunnel. The image acquisition device based on radar vision technology continuously acquires images of the inner wall of the target tunnel during the safety monitoring process.

[0035] Secondly, this application provides a tunnel safety monitoring system based on edge computing. The system includes sensors, edge computing nodes, a cloud platform, and monitoring units. The sensors are deployed at various monitoring points in the target tunnel, and the monitoring units include:

[0036] The acquisition module is used to acquire images of the inner wall of the target tunnel through sensors deployed at various monitoring points in the target tunnel, and transmit the images of the inner wall of the tunnel to the edge computing node in real time.

[0037] The processing module is used to determine the edge convolution kernels in different angle directions when the edge computing node extracts the edge of the tunnel inner wall image, and to perform multi-directional inner wall edge recognition on the tunnel inner wall image based on the edge convolution kernels in each angle direction, so as to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state.

[0038] The processing module is further configured to determine the grayscale compression coefficient of the tunnel inner wall image in the luminance channel, and to perform projection compression on the tunnel inner wall image based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel, so as to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions.

[0039] The processing module is further configured to perform heterogeneous feature fusion on the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain an identification feature map of the inner wall region when the target tunnel is identified for security purposes;

[0040] The execution module is used to analyze the safety hazards of the inner wall of the target tunnel based on the identification feature map, and the edge computing node outputs the safety hazard information of the inner wall of the target tunnel to the cloud platform.

[0041] Thirdly, this application provides a computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described edge computing-based tunnel security monitoring method.

[0042] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned edge computing-based tunnel security monitoring method.

[0043] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0044] The tunnel safety monitoring system and method based on edge computing provided in this application firstly acquires images of the tunnel's inner wall using sensors deployed at various monitoring points in the target tunnel, and transmits these images to an edge computing node in real time. Secondly, the edge computing node determines the edge convolution kernels at different angles when extracting edges from the tunnel's inner wall images, and performs multi-directional inner wall edge recognition based on these kernels to obtain the structural edge features of the target tunnel's inner wall region under the current monitoring state. Further, the system determines the tunnel's inner wall image... The image of the tunnel inner wall is compressed by a grayscale compression coefficient in the luminance channel and then projected onto the image based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions. Then, heterogeneous feature fusion is performed on the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain the identification feature map of the inner wall region when the target tunnel is identified for safety purposes. Finally, the safety hazards of the target tunnel inner wall are analyzed based on the identification feature map, and the safety hazard information of the target tunnel inner wall is output to the cloud platform by the edge computing node.

[0045] Therefore, this application can identify detailed features of tunnel inner wall images under the influence of complex distribution of surface defects and uneven illumination. First, the edge computing node sets edge convolution kernels at different angles when extracting edges from the tunnel inner wall image based on the predicted distribution probability of the crack direction in the target tunnel. This provides conditions for highlighting grayscale gradient changes in the tunnel inner wall image, thereby improving the recognition accuracy of the tunnel inner wall image. Second, based on the edge convolution kernels at various angles, multi-directional inner wall edge recognition is performed on the tunnel inner wall image to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state, providing edge information of the tunnel structure. This improves the accuracy and stability of automated tunnel safety monitoring, and avoids the loss of some recognition information caused by the complex directional distribution of surface defects on the tunnel inner wall. Furthermore, the tunnel inner wall image is projected and compressed based on the grayscale compression coefficient combined with the grayscale statistical parameters of the brightness channel to obtain the target tunnel... The grayscale projection features of the inner wall region under the current lighting conditions are used to effectively identify possible abnormal structures in the tunnel inner wall image, avoiding recognition interference caused by uneven lighting inside the tunnel and effectively reducing the impact of local lighting changes. Then, heterogeneous feature fusion is performed on the tunnel inner wall image based on structural edge features and grayscale projection features to obtain an identification feature map of the inner wall region when the target tunnel is being identified for safety purposes. This simultaneously preserves the edge information and brightness feature information of the tunnel structure, enabling the identification feature map to adapt to tunnel environments with different lighting, materials, and damage types, thereby effectively enhancing the ability to identify tunnel defects. Finally, the safety hazards of the target tunnel are analyzed based on the identification feature map, and the safety hazard information of the target tunnel is output to the cloud platform by the edge computing node. In summary, the technical solution provided in this application can identify the detailed features of the tunnel inner wall image under the influence of complex distribution of defects on the tunnel inner wall surface and uneven lighting distribution, so as to promptly issue tunnel safety alarms. Attached Figure Description

[0046] Figure 1 This is an exemplary flowchart of a tunnel safety monitoring method based on edge computing, as shown in some embodiments of this application;

[0047] Figure 2 This is an exemplary flowchart illustrating the determination of structural edge features according to some embodiments of this application;

[0048] Figure 3 This is an exemplary flowchart illustrating the determination of grayscale compression coefficients according to some embodiments of this application;

[0049] Figure 4 This is a schematic diagram of the structure of the monitoring unit shown in some embodiments of this application;

[0050] Figure 5This is a schematic diagram of the structure of a computer device implementing an edge computing-based tunnel safety monitoring method according to some embodiments of this application. Detailed Implementation

[0051] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] refer to Figure 1 The figure is an exemplary flowchart of an edge computing-based tunnel safety monitoring method according to some embodiments of this application. The edge computing-based tunnel safety monitoring method 100 mainly includes the following steps:

[0053] In step 101, the tunnel inner wall images of the target tunnel are collected by sensors deployed at various monitoring points in the target tunnel, and the tunnel inner wall images are transmitted to the edge computing node in real time.

[0054] In specific implementation, images of the tunnel interior walls are collected by sensors deployed at various monitoring points in the target tunnel. These sensors include a displacement sensor based on BeiDou technology and an image acquisition device based on radar-based technology. The BeiDou-based displacement sensor is deployed at key parts of the tunnel (such as the arch, arch waist, and sidewalls) to acquire high-precision displacement data. The radar-based image acquisition device is installed at various monitoring points on the tunnel interior walls, where each monitoring point is a pre-defined location. The radar-based image acquisition device can continuously collect images of the tunnel interior walls during safety monitoring. Furthermore, this application uses 5G communication to transmit the tunnel interior wall images to edge computing nodes in real time. It should be noted that the tunnel interior wall images in this application are digital images of the tunnel interior wall structure, which can be used to analyze the surface condition of the tunnel interior walls and detect defects that may affect tunnel safety.

[0055] It should also be noted that, in this application, edge computing nodes refer to computing devices deployed close to the data source (such as tunnel monitoring equipment) for localized data processing, reducing cloud computing pressure, and improving real-time performance. In the tunnel safety monitoring system, edge computing nodes are responsible for receiving, processing, and analyzing images of the tunnel wall and outputting safety hazard information without transmitting all raw data to a remote server or cloud.

[0056] In step 102, the edge convolution kernels of the edge computing node at different angles are determined when the edge is extracted from the tunnel inner wall image. Based on the edge convolution kernels at each angle, the inner wall image of the tunnel is identified in multiple directions to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state.

[0057] In some embodiments, determining the edge convolution kernels at different angles when the edge computing node extracts edges from the tunnel inner wall image can be achieved using the following steps:

[0058] Acquire historical crack monitoring data of the target tunnel;

[0059] Based on the historical crack monitoring data, the distribution of crack directions in the target tunnel is predicted to obtain the predicted distribution probability of different crack directions in the target tunnel.

[0060] The predicted distribution entropy of the target tunnel cracks is calculated by using the predicted distribution probability of each crack direction.

[0061] The edge computing node calibrates the angle direction when performing edge extraction on the tunnel inner wall image based on the predicted distribution entropy, and obtains multiple different angle directions;

[0062] By setting edge convolution kernels in different angle directions based on the edge detection operator, edge convolution kernels in different angle directions can be obtained when extracting edges from the tunnel inner wall image.

[0063] In specific implementation, firstly, historical crack monitoring data of the target tunnel is acquired. This historical crack monitoring data can be obtained by collecting a large number of crack monitoring images through cameras in the tunnel safety monitoring system, which will not be elaborated here. The historical crack monitoring data includes crack monitoring images from multiple different time points. Secondly, the distribution probability of crack directions in the target tunnel is predicted based on the historical crack monitoring data. Specifically, the crack direction distribution of each crack monitoring image in the historical crack data is extracted using Hough transform, and the probability distribution of different crack directions is calculated using histogram statistics to obtain the predicted distribution probability of different crack directions in the target tunnel. Transform (HT) is a feature extraction method for detecting straight lines, arcs, or specific shapes in images, particularly suitable for applications such as crack detection, road boundary detection, and medical image analysis. Further, the predicted distribution entropy of the target tunnel crack is calculated using the predicted distribution probabilities of different crack directions. That is, the predicted probability distributions of different crack directions are input as input parameters to a preset entropy model, and the entropy model outputs the predicted distribution entropy of the target tunnel crack. The preset entropy model can be the Shannon entropy model; other entropy models can also be used in other implementation examples, without limitation. Then, the edge computing node calibrates the angle direction when extracting edges from the tunnel inner wall image based on the predicted distribution entropy, obtaining multiple different angle directions. That is, the edge computing node compares the predicted distribution entropy with a distribution entropy threshold; when the predicted distribution entropy is less than or equal to the distribution entropy threshold, the edge is extracted. When performing edge extraction on the tunnel wall image, the angular direction is calibrated as horizontal and vertical. When the predicted distribution entropy is greater than the distribution entropy threshold, the angular direction for edge extraction on the tunnel wall image is calibrated as horizontal, vertical, and oblique symmetry directions. The oblique symmetry directions are 45° and 135°. The distribution entropy threshold is a pre-set standard predicted distribution entropy, which can be set according to actual conditions and is not limited here. Finally, edge convolution kernels in different angular directions are set based on the edge detection operator to obtain edge convolution kernels in different angular directions when extracting edges from the tunnel wall image. That is, the horizontal and vertical edge convolution kernels are generated by the Sobel directional gradient operator, and the convolution kernels in the main diagonal direction and the convolution kernels in the secondary diagonal direction generated by the Roberts cross operator are used as edge convolution kernels in the oblique symmetry directions of 45° and 135°, respectively. This will not be elaborated further here.

[0064] It should be noted that, in this embodiment, the historical crack monitoring data represents a set containing multiple historical crack monitoring images, which are images containing cracks in the tunnel inner wall; the predicted distribution probability in this application represents the predicted probability distribution of different crack directions in the target tunnel, and the predicted distribution probability is used to describe the probability of cracks appearing in different directions; the predicted distribution entropy in this embodiment represents a quantitative index for measuring the uncertainty of crack direction, used to describe the degree of dispersion of the crack direction distribution in the target tunnel, that is, whether the crack distribution in different directions is uniform, or whether it is concentrated in a specific direction; the angular direction in this embodiment represents the direction in which the convolution kernel acts in the image; the edge convolution kernel in this application represents a convolution operator used to detect edge information in the image, used to highlight the gradient change area of ​​the image. In the edge recognition of the tunnel inner wall image, edge detection in a single direction may lead to the loss of some crack information, affecting subsequent crack recognition. Therefore, by determining the edge convolution kernel in different directions, conditions can be provided for highlighting the gray-level gradient changes in the tunnel inner wall image, thereby improving the recognition accuracy of the tunnel inner wall image.

[0065] In some embodiments, reference Figure 2 As shown in the figure, this is an exemplary flowchart of determining structural edge features according to some embodiments of this application. In this embodiment, the inner wall edge recognition of the tunnel inner wall image is performed in multiple directions based on edge convolution kernels in various angular directions to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state. This can be achieved by the following steps:

[0066] First, in step 1021, the convolutional blocks corresponding to each pixel in the tunnel inner wall image are extracted;

[0067] Secondly, in step 1022, a pixel is selected as the selected pixel, and the convolution block of the selected pixel is convolved with the edge convolution kernels in each angular direction to obtain the convolution pixel of the selected pixel.

[0068] Furthermore, in step 1023, the convolutional pixels of the remaining pixels are determined;

[0069] Then, in step 1024, an edge detail map corresponding to the tunnel inner wall image is constructed based on all the convolutional pixels;

[0070] Finally, in step 1025, the structural edge features of the target tunnel inner wall region under the current monitoring state are identified from the edge detail map.

[0071] In specific implementation, firstly, the convolution blocks corresponding to each pixel in the tunnel inner wall image are extracted using the image processing tool OpenCV. That is, the convolution block corresponding to each pixel in the tunnel inner wall image is obtained. The size of the convolution block is usually determined by the size of the edge convolution kernels in each angular direction, typically 3×3. Secondly, a pixel is selected as the selected pixel, and the convolution block of the selected pixel is convolved with the edge convolution kernels in each angular direction to obtain the convolution pixel of the selected pixel. In other words, a pixel is selected as the selected pixel. For each pixel, the convolutional block of the selected pixel is multiplied element-wise by edge convolutional kernels in each angular direction, and the sum is obtained to get the convolutional sub-pixels of the selected pixel in each angular direction. The sum of squares of all convolutional sub-pixels is calculated, and the square root of the result of the sum of squares and the value 2 is taken to obtain the convolutional pixel of the selected pixel. The convolutional sub-pixels represent the convolutional pixels in different angular directions. Further, the method of determining the convolutional pixels of the selected pixel is to "convolve the convolutional block of the selected pixel with edge convolutional kernels in each angular direction". Continue to determine the convolutional pixels of the remaining pixels; then, construct the edge detail map corresponding to the tunnel inner wall image based on all the convolutional pixels, that is: combine the convolutional pixels corresponding to each pixel to obtain the edge detail map corresponding to the tunnel inner wall image; finally, identify the structural edge features of the target tunnel inner wall region in the current monitoring state from the edge detail map, that is: set the edge judgment value of the edge detail map, compare the edge judgment value with the pixel value of each pixel in the edge detail map, and take the pixel corresponding to the edge judgment value as the edge pixel, and then extract the edge pixel statistics, the edge pixel statistics represent the total number of all edge pixels, and take the quotient of the edge pixel statistics and the total number of pixels in the edge detail image as the structural edge features of the target tunnel inner wall region in the current monitoring state. The edge judgment value can be set according to actual needs. For example, the average pixel value of edge pixels in historical tunnel inner wall images can be used as the edge judgment value. In addition, in other embodiments, other calculation methods can be used to calculate the structural edge features of the target tunnel, which is not limited here.

[0072] It should be noted that, in this embodiment, the convolution block represents a matrix region formed by a selected pixel and its surrounding neighborhood in the image; the convolution pixel represents the new pixel value obtained after the convolution operation; the edge detail map represents a tunnel inner wall image that highlights edge details. The edge detail map contains rich edge information, making the edge features of the tunnel inner wall image clearer and more complete, thereby providing effective analytical data for tunnel safety monitoring; in this application, structural edge features represent structural edge information in the tunnel inner wall image. Determining structural edge features is a key step in accurately identifying structural problems such as cracks, spalling, and lining joints in tunnel safety monitoring. Structural edge features effectively provide edge information of the tunnel structure, which can improve the accuracy and stability of automated tunnel safety monitoring, thereby avoiding the loss of some identification information caused by the complex directional distribution of defects on the tunnel inner wall surface.

[0073] It should also be noted that the multi-directional inner wall edge recognition in this application refers to the process of recognizing the inner wall edges in an inner wall image from multiple directions. Specifically, the multi-directional inner wall edge recognition of the tunnel inner wall image is performed based on edge convolution kernels in various angular directions. This involves: extracting the convolution blocks corresponding to each pixel in the tunnel inner wall image; selecting a pixel as the selected pixel, convolving the selected pixel's convolution block with edge convolution kernels in various angular directions to obtain the selected pixel's convolution pixels; further determining the convolution pixels of the remaining pixels; constructing an edge detail map corresponding to the tunnel inner wall image based on all the convolution pixels; and identifying the structural edge features of the target tunnel inner wall region under the current monitoring state from the edge detail map, thus completing the multi-directional inner wall edge recognition of the tunnel inner wall image.

[0074] In step 103, the grayscale compression coefficient of the tunnel inner wall image in the luminance channel is determined, and the tunnel inner wall image is projected and compressed based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions.

[0075] In some embodiments, reference Figure 3 As shown, this figure is an exemplary flowchart of determining the grayscale compression coefficient according to some embodiments of this application. In this embodiment, the grayscale compression coefficient of the tunnel inner wall image in the luminance channel can be determined by the following steps:

[0076] First, in step 1031, the tunnel inner wall image is converted into a grayscale inner wall image;

[0077] Then, in step 1032, the brightness change rate sequence corresponding to the grayscale inner wall image is calculated;

[0078] Finally, in step 1033, the grayscale compression coefficient of the tunnel inner wall image in the brightness channel is determined by the brightness change rate sequence.

[0079] In specific implementation, firstly, the tunnel inner wall image is converted into a grayscale image using RGB weighted averaging. Alternatively, in other embodiments, other methods can be used to convert the tunnel inner wall image into a grayscale inner wall image; this is not limited here. Secondly, the brightness change rate sequence corresponding to the grayscale inner wall image is calculated. That is, the brightness change rate of each pixel in the grayscale inner wall image is calculated using the Sobel operator, and all brightness change rates are combined sequentially to obtain the brightness change rate sequence corresponding to the grayscale inner wall image; this will not be elaborated further here. Finally, the tunnel inner wall is determined using the brightness change rate sequence. The grayscale compression coefficient of the image in the luminance channel is calculated by extracting the maximum and minimum luminance change rates from the luminance change rate sequence, obtaining the contrast scale of the tunnel inner wall image, calculating the difference between the maximum and minimum luminance change rates, and using the contrast scale and the result of the difference calculation as the grayscale compression coefficient of the tunnel inner wall image in the luminance channel. The contrast scale of the tunnel inner wall image is set to 255. In addition, in other embodiments, other calculation methods can be used to calculate the grayscale compression coefficient of the tunnel inner wall image in the luminance channel, which is not limited here.

[0080] It should be noted that, in this embodiment, the grayscale inner wall image represents a grayscale image converted from the original color image. The grayscale inner wall image only contains brightness information and not color information. In tunnel monitoring, the grayscale inner wall image can display the structural features and possible defects of the tunnel's inner surface. In this embodiment, the brightness change rate sequence represents a combination of multiple brightness change rates, whereby the brightness change rate represents the degree of change of a pixel's brightness value relative to its neighboring pixels. In this application, the grayscale compression coefficient represents a parameter used to adjust the distribution of brightness values ​​in an image. Specifically, the grayscale compression coefficient is used to control the degree of compression of the image's brightness (grayscale) values, making the image's brightness range more suitable for subsequent processing or analysis. By determining the grayscale compression coefficient, the details in the tunnel inner wall image can be effectively enhanced, highlighting areas with large brightness changes, such as cracks and defects.

[0081] In some embodiments, the projection compression of the tunnel inner wall image based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel to obtain the grayscale projection features of the target tunnel inner wall region under the current illumination conditions can be achieved by the following steps:

[0082] Obtain the grayscale data of the tunnel inner wall image in the luminance channel;

[0083] The grayscale statistical parameters of the brightness channel are determined based on the grayscale data;

[0084] An adaptive adjustment factor is set based on the grayscale statistical parameters when projecting and compressing the tunnel inner wall image.

[0085] The grayscale of each pixel in the brightness channel of the tunnel inner wall image is projected and compressed using the adaptive adjustment factor and the grayscale compression coefficient to obtain the brightness projection compression map corresponding to the tunnel inner wall image.

[0086] The grayscale projection features of the target tunnel inner wall region under the current lighting conditions are extracted from the brightness projection compression map.

[0087] In specific implementation, firstly, grayscale data of the tunnel inner wall image in the luminance channel is obtained using the image processing tool OpenCV. This grayscale data contains multiple grayscale values, representing the grayscale values ​​of each pixel in the tunnel inner wall image within the luminance channel. Secondly, grayscale statistical parameters of the luminance channel are determined based on the grayscale data, specifically by calculating the average grayscale value and standard deviation of the grayscale data, and combining these values ​​as the grayscale statistical parameters of the luminance channel. Further, an adaptive adjustment factor is set based on the grayscale statistical parameters for projection compression of the tunnel inner wall image, specifically by adjusting the average grayscale value in the grayscale statistical parameters. The mean value is summed with 0.2 times the grayscale standard deviation, and the quotient of the summation result and the average grayscale value is used as the independent variable value of an exponential function with base e to obtain the adaptive adjustment factor for projection compression of the tunnel inner wall image. The average grayscale value is always not equal to 0. In addition, in other embodiments, other calculation methods can be used to calculate the adaptive adjustment factor for projection compression of the tunnel inner wall image, which is not limited here. Then, the grayscale values ​​of each pixel in the luminance channel of the tunnel inner wall image are projected and compressed using the adaptive adjustment factor and the grayscale compression coefficient to obtain the luminance projection compression map corresponding to the tunnel inner wall image, that is: using a power law transformation function Using the grayscale compression coefficient as the product coefficient of the power-law transform function and the adaptive adjustment factor as the exponent of the power-law transform function, the grayscale values ​​of each pixel in the luminance channel of the tunnel inner wall image are used as input parameters of the projection compression model. The projection compression model outputs the grayscale compression value corresponding to each pixel, and the original grayscale value is replaced by the compressed grayscale value to obtain the luminance projection compression map corresponding to the tunnel inner wall image. The power-law transform function is a nonlinear transform function widely used in the field of image processing. Finally, the target tunnel inner wall region under the current illumination conditions is extracted from the luminance projection compression map. The grayscale projection feature is obtained by setting a high brightness judgment value for the brightness projection compression map, comparing the high brightness judgment value with the pixel value of each pixel in the brightness projection compression map, and taking the pixels corresponding to the high brightness judgment value as high frequency pixels. Then, the high frequency pixel statistics are extracted, which represent the total number of all high frequency pixels. The quotient of the high frequency pixel statistics and the total number of pixels in the brightness projection compression map is taken as the grayscale projection feature of the target tunnel inner wall area under the current illumination conditions. In addition, other methods can be used to calculate the grayscale projection feature of the target tunnel inner wall area under the current illumination conditions, which are not limited here.

[0088] It should be noted that, in this embodiment, the grayscale statistical parameters represent a set of numerical indicators describing the grayscale value distribution characteristics of the tunnel inner wall image in the luminance channel. These grayscale statistical parameters include the grayscale mean and grayscale standard deviation, effectively reflecting information such as the overall brightness, contrast, and grayscale variation trend of the image. In this embodiment, the adaptive adjustment factor represents a parameter used to adjust the degree of grayscale compression. In this embodiment, the luminance projection compression map represents a new image obtained after compressing and projecting the grayscale information of the tunnel inner wall image in the luminance channel. This luminance projection compression map retains the main luminance information of the tunnel inner wall structure and compresses the grayscale to enhance structural features, providing effective data for subsequent feature extraction and analysis of the tunnel inner wall. Data support; In this application, grayscale projection features represent the feature information extracted after projecting and compressing the tunnel inner wall image onto the brightness channel. The grayscale projection features characterize the brightness distribution pattern and structural characteristics of the tunnel inner wall. In tunnel safety monitoring, the grayscale distribution of defects such as cracks and spalling is often different from the background. Grayscale projection features can help identify these abnormal structures. In addition, the lighting conditions inside the tunnel are complex. Directly using the original brightness value may be affected by the intensity of the light source and shadows. However, grayscale projection features extract the brightness distribution pattern through statistical methods, which can effectively identify abnormal structures that may exist in the tunnel inner wall image, avoid the identification interference caused by uneven lighting inside the tunnel, and effectively reduce the impact of local lighting changes.

[0089] It should also be noted that, in this application, projection compression refers to the process of compressing the grayscale of an image. Specifically, projection compression of the tunnel inner wall image is performed based on the grayscale compression coefficient and the grayscale statistical parameters of the luminance channel. This involves: acquiring the grayscale data of the tunnel inner wall image in the luminance channel; determining the grayscale statistical parameters of the luminance channel based on the grayscale data; setting an adaptive adjustment factor for projection compression of the tunnel inner wall image based on the grayscale statistical parameters; performing projection compression on the grayscale of each pixel in the luminance channel of the tunnel inner wall image using the adaptive adjustment factor and the grayscale compression coefficient to obtain a luminance projection compression map corresponding to the tunnel inner wall image; and extracting the grayscale projection features of the target tunnel inner wall region under the current illumination conditions from the luminance projection compression map, thus completing the projection compression of the tunnel inner wall image.

[0090] In step 104, heterogeneous feature fusion is performed on the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain the identification feature map of the inner wall region when the target tunnel is identified for security purposes.

[0091] In some embodiments, the heterogeneous feature fusion of the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain the identification feature map of the inner wall region when the target tunnel is identified for security purposes can be achieved by the following steps:

[0092] Based on the structural edge features and grayscale projection features, the edge fusion coefficient and brightness fusion coefficient of the tunnel inner wall are determined respectively when the target tunnel is being monitored for safety.

[0093] Extract the edge detail map and brightness projection compression map corresponding to the tunnel inner wall image;

[0094] The edge detail map and the brightness projection compression map are decomposed into multi-scale sub-maps at different scale levels to obtain edge detail sub-maps and brightness projection compression sub-maps at different scale levels.

[0095] Select a scale level as the selected scale level, and fuse the edge detail sub-image and the brightness projection compression sub-image under the selected scale level based on the edge fusion coefficient and the brightness fusion coefficient to obtain the fused sub-image corresponding to the selected scale level.

[0096] Continue to determine the fusion subgraphs corresponding to the remaining scale levels;

[0097] Generate a feature map of the inner wall region for security identification of the target tunnel based on all fused sub-graphs.

[0098] In specific implementation, firstly, the edge blending coefficient and brightness blending coefficient of the tunnel inner wall during safety monitoring of the target tunnel are determined based on the structural edge features and grayscale projection features, respectively. Specifically, the structural edge features and grayscale projection features are normalized using minimum-maximum normalization, and the normalized structural edge features and grayscale projection features are used as the edge blending coefficient and brightness blending coefficient of the tunnel inner wall during safety monitoring of the target tunnel, respectively. The normalized structural edge features and grayscale projection features have a unified dimension. Furthermore, in other embodiments, other calculation methods can also be used to calculate the target... The edge blending coefficient and brightness blending coefficient of the tunnel inner wall during tunnel safety monitoring are not limited here. Secondly, the edge detail map and brightness projection compression map corresponding to the tunnel inner wall image in the tunnel safety monitoring database are used. Further, the edge detail map and brightness projection compression map are decomposed into multiple scales to obtain edge detail sub-maps and brightness projection compression sub-maps at different scale levels. Specifically, the Laplacian pyramid is used to decompose the edge detail map and brightness projection compression map into multiple scales to obtain edge detail sub-maps and brightness projection compression sub-maps at different scale levels. For example, a Gaussian pyramid can be used to decompose the edge detail map and brightness projection compression map into multiple scales. The detail image and the luminance projection compressed image are downsampled to obtain edge detail sub-images and luminance projection compressed sub-images at different scale levels. Alternatively, other multi-scale decomposition algorithms can be used for decomposition, which are not limited here. The edge detail sub-images and the luminance projection compressed sub-images are images of the same scale. Further, a scale level is selected as the chosen scale level. Based on the edge fusion coefficient and the luminance fusion coefficient, the edge detail sub-images and luminance projection compressed sub-images at the chosen scale level are fused to obtain the fused sub-image corresponding to the chosen scale level. That is, the edge fusion coefficient and the luminance projection compressed sub-image are fused together. The fusion coefficients are used as weights for the edge detail sub-image and the brightness projection compression sub-image at the selected scale level, respectively, and are weighted and fused based on the Laplacian pyramid to obtain the fused sub-image corresponding to the selected scale level. Further details are omitted here. In other embodiments, other fusion methods can also be used, which are not limited here. Then, the remaining scale levels' fused sub-images are determined by fusing the edge detail sub-image and the brightness projection compression sub-image at the selected scale level based on the edge fusion coefficients and the brightness fusion coefficients to obtain the fused sub-image corresponding to the selected scale level.Finally, pyramid reconstruction is used to perform an inverse transformation on all fused sub-images to generate a recognition feature map of the inner wall region for security identification of the target tunnel. Specifically, starting with the highest-scale fused sub-image, the current fused sub-image is upsampled layer by layer (usually by restoring the size through bilinear interpolation or Gaussian interpolation). The upsampled image is then superimposed pixel-by-pixel with the next-level fused sub-image to form a new intermediate image. This process is repeated until all scale-level fused sub-images have been processed, ultimately restoring a recognition feature map with the same size as the original image.

[0099] It should be noted that, in this embodiment, the edge blending coefficient represents a value used to adjust the blending effect of the edge detail map, and the brightness blending coefficient represents a value used to adjust the blending effect of the brightness projection compression map; in this embodiment, the edge detail sub-map represents edge information sub-maps at different scale levels decomposed from the edge detail map, and the brightness projection compression sub-map represents brightness information sub-maps at different scale levels decomposed from the edge brightness projection compression map; in this embodiment, the blended sub-map represents sub-maps generated at different scale levels. Specifically, the blended sub-map refers to the sub-maps generated at different scale levels after weighted fusion of the edge detail sub-map and the brightness projection compression sub-map during the multi-scale decomposition process; the identification in this application... The feature map represents the tunnel inner wall feature map obtained after fusing multiple features. Specifically, the recognition feature map is a feature image generated by heterogeneously fusing edge detail map and brightness projection compression map based on multi-scale decomposition, which can identify the safety status of the target tunnel. The recognition feature map is used for safety monitoring, crack detection or defect identification of the tunnel inner wall. This image fuses structural edge features (provided by edge detail map) and grayscale projection features (provided by brightness projection compression map), which can simultaneously retain the edge information and brightness feature information of the tunnel structure. This allows the recognition feature map to adapt to tunnel environments with different lighting, materials and damage types, improve the generalization ability of the algorithm, and effectively enhance the ability to identify tunnel defects.

[0100] It should also be noted that heterogeneous feature fusion in this application refers to the process of fusing multiple image features. Specifically, heterogeneous feature fusion of the tunnel inner wall image is performed based on the structural edge features and the grayscale projection features. This involves: determining the edge fusion coefficient and brightness fusion coefficient of the tunnel inner wall for safety monitoring based on the structural edge features and grayscale projection features, respectively; extracting the edge detail map and brightness projection compression map corresponding to the tunnel inner wall image; performing multi-scale decomposition on the edge detail map and brightness projection compression map to obtain edge detail sub-maps and brightness projection compression sub-maps at different scale levels; selecting a scale level as the selected scale level, and fusing the edge detail sub-map and brightness projection compression sub-map at the selected scale level based on the edge fusion coefficient and the brightness fusion coefficient to obtain the fused sub-map corresponding to the selected scale level; continuing to determine the fused sub-maps corresponding to the remaining scale levels; and generating a recognition feature map of the inner wall region for safety identification of the target tunnel based on all the fused sub-maps, thus completing the heterogeneous feature fusion of the tunnel inner wall image.

[0101] In step 105, the safety hazards of the inner wall of the target tunnel are analyzed based on the identified feature map, and the edge computing node outputs the safety hazard information of the inner wall of the target tunnel to the cloud platform.

[0102] In some embodiments, the analysis of safety hazards on the inner wall of the target tunnel based on the identified feature map, and the output of the safety hazard information of the inner wall of the target tunnel to the cloud platform by the edge computing node, can be achieved by the following steps:

[0103] Crack identification is performed on the identified feature map to obtain the crack identification features of the target tunnel;

[0104] The edge computing node compares the crack identification features with preset crack comparison features and outputs the safety hazard information of the target tunnel inner wall to the cloud platform.

[0105] In specific implementation, firstly, Canny edge detection is used to identify cracks in the feature map to obtain crack identification features of the target tunnel inner wall. The crack identification feature is the crack length. Then, the edge computing node compares the crack identification feature with a preset crack comparison feature using Euclidean distance. If the crack identification feature is greater than or equal to the preset crack comparison feature, it is determined that there is a safety hazard in the target tunnel inner wall, and the edge computing node outputs the corresponding safety hazard information (i.e., information on the current safety hazard in the tunnel inner wall) to the cloud platform. If the crack identification feature is less than the preset crack comparison feature, it is determined that the target tunnel inner wall is within the safe range and no action is taken.

[0106] It should be noted that the preset crack comparison feature in this embodiment represents a pre-defined standard crack feature used to identify and evaluate tunnel cracks. It is usually established based on historical crack data, engineering specifications, experimental analysis and other information, which will not be elaborated here.

[0107] Furthermore, in another aspect of this application, in some embodiments, this application provides a tunnel safety monitoring system based on edge computing. This system includes sensors, edge computing nodes, a cloud platform, and a monitoring unit. (Refer to...) Figure 4 The figure is a schematic diagram of the structure of a monitoring unit according to some embodiments of this application. The monitoring unit 200 includes: a data acquisition module 201, a processing module 202, and an execution module 203, which are described below:

[0108] The acquisition module 201 in this application is mainly used to acquire images of the inner wall of the target tunnel through sensors deployed at various monitoring points of the target tunnel, and transmit the images of the inner wall of the tunnel to the edge computing node in real time.

[0109] Processing module 202, in this application, is mainly used to determine the edge convolution kernels in different angle directions when the edge computing node extracts the edge of the tunnel inner wall image, and to perform multi-directional inner wall edge recognition on the tunnel inner wall image based on the edge convolution kernels in each angle direction, so as to obtain the structural edge features of the target tunnel inner wall region under the current monitoring state.

[0110] The processing module 202 is further configured to determine the grayscale compression coefficient of the tunnel inner wall image in the brightness channel, and to perform projection compression on the tunnel inner wall image based on the grayscale compression coefficient and the grayscale statistical parameters of the brightness channel, so as to obtain the grayscale projection features of the target tunnel inner wall area under the current illumination conditions.

[0111] In addition, the processing module 202 is also used to perform heterogeneous feature fusion on the tunnel inner wall image based on the structural edge features and the grayscale projection features to obtain the identification feature map of the inner wall area when the target tunnel is identified for security purposes;

[0112] The execution module 203 in this application is mainly used to analyze the safety hazards of the inner wall of the target tunnel based on the identification feature map, and output the safety hazard information of the inner wall of the target tunnel to the cloud platform by the edge computing node.

[0113] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described edge computing-based tunnel security monitoring method.

[0114] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device implementing an edge computing-based tunnel safety monitoring method according to some embodiments of this application. The edge computing-based tunnel safety monitoring method in the above embodiments can... Figure 5 The computer device shown is used to implement this, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0115] The processor 301 can be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more devices used to control the execution of the edge computing-based tunnel security monitoring method in this application.

[0116] The communication bus 302 can be used to transmit information between the aforementioned components.

[0117] The memory 303 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or it may be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via a communication bus 302. The memory 303 may also be integrated with the processor 301.

[0118] The memory 303 stores program code for executing the scheme of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the determination of the tunnel security monitoring method based on edge computing can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0119] Communication interface 304 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0120] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0121] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0122] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described tunnel security monitoring method based on edge computing.

[0123] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0124] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. An edge computing-based tunnel safety monitoring method for a tunnel safety monitoring system to perform tunnel safety monitoring, the tunnel safety monitoring system comprising sensors, edge computing nodes and a cloud platform, the sensors being arranged at respective monitoring points of a target tunnel, characterized in that, The method comprises the following steps: Collecting a tunnel inner wall image of the target tunnel through sensors arranged at each monitoring point of the target tunnel, and transmitting the tunnel inner wall image to an edge computing node in real time; Determining edge convolution kernels in different angle directions when the edge computing node performs edge extraction on the tunnel inner wall image, and performing multi-directional inner wall edge recognition on the tunnel inner wall image according to the edge convolution kernels in each angle direction to obtain structural edge features of the target tunnel inner wall region in the current monitoring state; Determining a gray scale compression coefficient of the tunnel inner wall image on a brightness channel, and performing projection compression on the tunnel inner wall image based on the gray scale compression coefficient and the gray scale statistical parameters of the brightness channel to obtain a gray scale projection feature of the target tunnel inner wall region under the current lighting condition; Fusing the structural edge features and the gray scale projection feature to obtain an identification feature map of the inner wall region when the target tunnel is subjected to safety identification; Analyzing the safety hazards of the target tunnel inner wall according to the identification feature map, and outputting the safety hazard information of the target tunnel inner wall to a cloud platform by the edge computing node.

2. The method of claim 1, wherein, The determination of the edge convolution kernels in different angle directions when the edge computing node performs edge extraction on the tunnel inner wall image specifically comprises: Obtaining historical crack monitoring data of the target tunnel; Predicting the distribution of the crack direction of the target tunnel according to the historical crack monitoring data to obtain a prediction distribution probability of different crack directions of the target tunnel; Calculating a prediction distribution entropy of the cracks of the target tunnel through the prediction distribution probability of each different crack direction; Calibrating the angle direction when the edge computing node performs edge extraction on the tunnel inner wall image based on the prediction distribution entropy to obtain a plurality of different angle directions; Setting edge convolution kernels in different angle directions based on an edge detection operator to obtain the edge convolution kernels in different angle directions when the tunnel inner wall image is subjected to edge extraction.

3. The method of claim 1, wherein, The multi-directional inner wall edge recognition of the tunnel inner wall image according to the edge convolution kernels in each angle direction to obtain the structural edge features of the target tunnel inner wall region in the current monitoring state specifically comprises: Extracting convolution blocks corresponding to each pixel point in the tunnel inner wall image; Selecting a pixel point as a selected pixel point, convolving the convolution block of the selected pixel point with the edge convolution kernels in each angle direction to obtain a convolution pixel of the selected pixel point; Continuing to determine the convolution pixels of the remaining pixel points; Constructing an edge detail map corresponding to the tunnel inner wall image according to all the convolution pixels; Identifying the structural edge features of the target tunnel inner wall region in the current monitoring state from the edge detail map.

4. The method of claim 1, wherein, The determination of the gray scale compression coefficient of the tunnel inner wall image on the brightness channel specifically comprises: Converting the tunnel inner wall image into a gray scale inner wall image; Calculating a brightness change rate sequence corresponding to the gray scale inner wall image; Determining the gray scale compression coefficient of the tunnel inner wall image on the brightness channel through the brightness change rate sequence.

5. The method of claim 1, wherein, The tunnel inner wall image is projected and compressed based on the gray scale compression coefficient and the gray scale statistical parameter of the brightness channel, to obtain a gray scale projection feature of the target tunnel inner wall region under the current lighting condition, and the gray scale projection feature specifically comprises: Obtaining gray scale data of the tunnel inner wall image on the brightness channel; Determining a gray scale statistical parameter of the brightness channel according to the gray scale data; Setting an adaptive adjustment factor for the projection and compression of the tunnel inner wall image based on the gray scale statistical parameter; Projecting and compressing the gray scale of each pixel of the tunnel inner wall image on the brightness channel through the adaptive adjustment factor and the gray scale compression coefficient, to obtain a brightness projection compressed image corresponding to the tunnel inner wall image; Extracting a gray scale projection feature of the target tunnel inner wall region under the current lighting condition from the brightness projection compressed image.

6. The method of claim 1, wherein, The sensor comprises a displacement sensor based on Beidou technology and an image acquisition device based on radar vision technology.

7. The method of claim 6, wherein, The displacement sensor based on Beidou technology is arranged at key positions of the tunnel, and is used to obtain high-precision displacement data, wherein the key positions of the tunnel include a vault, a haunch and a side wall; the image acquisition device based on radar vision technology is installed at each monitoring point on the inner wall of the tunnel, and continuously acquires tunnel inner wall images of the target tunnel in the safety monitoring process through the image acquisition device based on radar vision technology.

8. An edge computing based tunnel safety monitoring system, the system comprising sensors, edge computing nodes, a cloud platform and a monitoring unit, the sensors are arranged at each monitoring point of a target tunnel, characterized in that, The monitoring unit comprises: An acquisition module, which is configured to acquire tunnel inner wall images of the target tunnel through sensors arranged at each monitoring point of the target tunnel, and transmit the tunnel inner wall images to an edge computing node in real time; A processing module, which is configured to determine edge convolution kernels in different angle directions when the edge computing node performs edge extraction on the tunnel inner wall images, and perform multi-directional inner wall edge recognition on the tunnel inner wall images according to the edge convolution kernels in each angle direction, to obtain structural edge features of the target tunnel inner wall region under the current monitoring state; The processing module is further configured to determine a gray scale compression coefficient of the tunnel inner wall image on the brightness channel, and perform projection compression on the tunnel inner wall image based on the gray scale compression coefficient and the gray scale statistical parameter of the brightness channel, to obtain a gray scale projection feature of the target tunnel inner wall region under the current lighting condition; The processing module is further configured to perform heterogeneous feature fusion on the tunnel inner wall image according to the structural edge features and the gray scale projection feature, to obtain an identification feature map of the inner wall region when the target tunnel is subjected to safety identification; An execution module, which is configured to analyze safety hazards of the target tunnel inner wall according to the identification feature map, and output safety hazard information of the target tunnel inner wall from the edge computing node to a cloud platform.

9. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the edge computing-based tunnel safety monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the edge computing-based tunnel safety monitoring method according to any one of claims 1 to 7.

Images