A subway tunnel segment deformation detection method, equipment and medium

Abstract

The present invention relates to the field of image processing technology, and in particular to a method, device and medium for detecting deformation of a subway tunnel segment. The method uses pixel points as objects to comprehensively evaluate the basic weight value of each pixel point based on semantic segmentation results of different dimensions, and then performs aggregate evaluation on connected domains to obtain reference evaluation values ​​corresponding to each second connected domain. The image adjustment coefficient is determined based on the reference distance between first connected domains, and then morphological operations are performed on the basic weight image to ensure that the morphological operation removes noise as much as possible without destroying the structural integrity. Aggregate evaluation is then performed on the subway segment image, thereby realizing hierarchical aggregation of deformation severity and improving the reliability of subway tunnel segment deformation detection.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, equipment and medium for detecting deformation of a subway tunnel segment. Background Art

[0002] As the core infrastructure of urban rail transit, the structural safety of subway tunnels is directly linked to operational safety and the safety of public life and property. Subway tunnel segments, the primary load-bearing structure of the tunnel, are susceptible to deformation (such as cracks, misalignment, and settlement) during long-term operation due to factors such as changing geological conditions, train vibration, and surrounding construction disturbances. If deformation is not promptly detected and repaired, it can lead to structural instability, water leakage, and even serious safety accidents.

[0003] Currently, subway tunnel segment deformation detection primarily relies on manual inspections, laser scanning, and automated image acquisition. Image-based detection methods are widely used due to their low cost, high efficiency, and automation capabilities. However, existing image-based monitoring methods suffer from noise interference, limited segmentation accuracy, and other factors, leading to fluctuations in the accuracy of quantitative deformation evaluation, resulting in poor reliability in subway tunnel segment deformation detection.

[0004] Therefore, how to improve the reliability of subway tunnel segment deformation detection has become an urgent problem to be solved. Summary of the Invention

[0005] In response to the above technical problems, the technical solution adopted by the present invention is a method for detecting deformation of a subway tunnel segment. The method comprises the following steps:

[0006] S101 , collecting subway segment images, and inputting the subway segment images into a trained first semantic segmentation model to obtain a first segmented image.

[0007] S102: Input the subway segment image into a trained second semantic segmentation model to obtain a second segmented image.

[0008] S103 : For any pixel point, determine a basic weight value corresponding to the pixel point according to a first segmentation category corresponding to the pixel point in the first segmented image and a second segmentation category corresponding to the pixel point in the second segmented image.

[0009] S104: forming a basic weight image according to the basic weight values ​​corresponding to all the pixels.

[0010] S105 , performing connected domain analysis on the first segmented image to obtain M first connected domains, and determining an image adjustment coefficient N according to a reference distance between every two first connected domains, where M and N are both positive integers.

[0011] S106 , performing N dilation processing on the basic weight image, and then performing N erosion processing on the basic weight image to obtain a target weight image, and performing binarization processing on the target weight image to obtain a target binary image.

[0012] S107 , performing connected component analysis on the target binary image to obtain K second connected components, where K is a positive integer.

[0013] S108 : For any second connected domain, determine a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains, and the target weight image.

[0014] S109 , determining a deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the respective second connected domains.

[0015] The present invention also provides a subway tunnel segment deformation detection device, which includes:

[0016] The image acquisition module is used to acquire subway pipe segment images and input the subway pipe segment images into the trained first semantic segmentation model to obtain a first segmented image.

[0017] The image segmentation module is used to input the subway segment image into the trained second semantic segmentation model to obtain a second segmented image.

[0018] The basic weight determination module is used to determine, for any pixel point, a basic weight value corresponding to the pixel point based on the first segmentation category corresponding to the pixel point in the first segmented image and the second segmentation category corresponding to the pixel point in the second segmented image.

[0019] The image forming module is used to form a basic weight image according to the basic weight values ​​corresponding to all pixel points.

[0020] The coefficient determination module is used to perform connected domain analysis on the first segmented image to obtain M first connected domains, and determine the image adjustment coefficient N according to the reference distance between each two first connected domains, where M and N are both positive integers.

[0021] The morphological processing module is used to perform N dilation processes on the basic weight image and then perform N erosion processes to obtain a target weight image, and perform binarization processing on the target weight image to obtain a target binary image.

[0022] The connected domain analysis module is used to perform connected domain analysis on the target binary image to obtain K second connected domains, where K is a positive integer.

[0023] The connected domain evaluation module is configured to determine, for any second connected domain, a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains and the target weight image.

[0024] The deformation evaluation module is used to determine the deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to each second connected domain.

[0025] The present invention also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned subway tunnel segment deformation detection method is implemented.

[0026] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned subway tunnel segment deformation detection method is implemented.

[0027] The present invention has at least the following beneficial effects: according to the semantic segmentation results of different dimensions, the basic weight value of each pixel point is comprehensively evaluated with the pixel points as the object, and then the connected domain is used as the object for aggregation evaluation to obtain the reference evaluation value corresponding to each second connected domain, and the image adjustment coefficient is determined according to the reference distance between the first connected domains, and then the basic weight image is subjected to morphological operations to ensure that the morphological operation removes noise as much as possible without destroying the structural integrity, and then the subway pipe segment image is used as the object for aggregation evaluation, thereby realizing the hierarchical aggregation of deformation severity and improving the reliability of subway tunnel pipe segment deformation detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0029] Figure 1 A schematic flow chart of a method for detecting deformation of a subway tunnel segment provided in the first embodiment of the present invention;

[0030] Figure 2 This is a structural schematic diagram of a subway tunnel segment deformation detection device provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It is understood that, where appropriate, the above-mentioned terms used to distinguish similar objects can be interchanged so that the present invention can also implement other embodiments other than the above-mentioned illustrated embodiments or described embodiments. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] Example 1

[0034] This embodiment provides a method for detecting deformation of a subway tunnel segment. Figure 1 FIG. 1 is a flow chart of a method for detecting deformation of a subway tunnel segment provided in a first embodiment of the present invention. The method for detecting deformation of a subway tunnel segment includes the following steps:

[0035] S101, collecting a subway segment image, and inputting the subway segment image into a trained first semantic segmentation model to obtain a first segmented image;

[0036] S102, inputting the subway segment image into a trained second semantic segmentation model to obtain a second segmented image;

[0037] S103: For any pixel point, determine a basic weight value corresponding to the pixel point according to a first segmentation category corresponding to the pixel point in the first segmented image and a second segmentation category corresponding to the pixel point in the second segmented image;

[0038] S104, forming a basic weight image according to the basic weight values ​​corresponding to all the pixels;

[0039] S105, performing a connected domain analysis on the first segmented image to obtain M first connected domains, and determining an image adjustment coefficient N based on a reference distance between every two first connected domains, where M and N are both positive integers;

[0040] S106, performing N dilation processing on the basic weight image, and then performing N erosion processing on the basic weight image to obtain a target weight image, and performing binarization processing on the target weight image to obtain a target binary image;

[0041] S107, performing connected component analysis on the target binary image to obtain K second connected components, where K is a positive integer;

[0042] S108, for any second connected domain, determining a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains, and the target weight image;

[0043] S109 , determining a deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the respective second connected domains.

[0044] Among them, the subway pipe segment image can be collected by an industrial camera. The first semantic segmentation model and the second semantic segmentation model have the same architecture and can both adopt the U-Net model architecture, but the training processes of the first semantic segmentation model and the second semantic segmentation model are different. For the same sample image, the first annotated image used to train the first semantic segmentation model includes the first category label corresponding to each pixel point, and the second annotated image used to train the second semantic segmentation model includes the second category label corresponding to each pixel point.

[0045] The basic weight value may indicate the contribution of the corresponding pixel point to the deformation degree evaluation. The higher the basic weight value, the greater the contribution of the corresponding pixel point to the deformation degree evaluation.

[0046] The image adjustment coefficient is used to determine the scale of subsequent morphological processing, while preserving the key spatial structure information in the basic weight image while eliminating misjudgment points and noise points as much as possible.

[0047] In a specific embodiment, the first segmentation categories include at least a segment body category, a joint category, a bolt category, a grouting hole category, and a background category, and the second segmentation categories include at least a crack category, a misalignment category, a defect category, a protrusion category, a deformation category, and a non-defect category.

[0048] The determining, according to the first segmentation category corresponding to the pixel in the first segmented image and the second segmentation category corresponding to the pixel in the second segmented image, a basic weight value corresponding to the pixel includes:

[0049] According to the first segmentation category corresponding to the pixel point in the first segmented image and the second segmentation category corresponding to the pixel point in the second segmented image, a preset mapping table is used to map and obtain a basic weight value corresponding to the pixel point, wherein the mapping table includes a mapping relationship between category combinations and basic weight values, and the category combination includes the first segmentation category and the second segmentation category.

[0050] Among them, the mapping relationships contained in the mapping table may include: the basic weight value corresponding to the joint category and the misalignment category is 0.9, the basic weight value corresponding to the joint category and the crack category is 0.8, the basic weight value corresponding to the bolt category and the protrusion category is 0.85, the basic weight value corresponding to the segment body category and the crack category is 0.6, the basic weight value corresponding to the segment body category and the defect category is 0.4, the basic weight value corresponding to the segment body category and the deformation category is 0.9, the basic weight value corresponding to the grouting hole category and the defect category is 0.75, etc. The implementer can set the mapping relationship in the mapping table and adjust the basic weight value corresponding to the mapping relationship based on actual experience, expert opinions and other factors.

[0051] Specifically, there are three situations for other category combinations: non-defective category, background category, and the first segmentation category and the second segmentation category do not correspond. Pixels in the non-defective category do not belong to defects, pixels in the background category do not need to be considered, and the situation where the first segmentation category and the second segmentation category do not correspond is a misjudgment.

[0052] In a specific embodiment, performing connected domain analysis on the first segmented image to obtain M first connected domains, and determining the image adjustment coefficient N based on a reference distance between each two first connected domains, includes:

[0053] According to the first segmentation categories corresponding to the respective pixel points, the pixel values ​​of the pixel points belonging to the preset segmentation category are set to the first preset value, and the pixel values ​​of the pixel points not belonging to the preset segmentation category are set to the second preset value, to form a temporary binary image;

[0054] Performing a connected domain analysis on the temporary binary image to obtain the M first connected domains;

[0055] For any two first connected domains, determining the centroid points corresponding to the two first connected domains, and forming a reference connecting line between the centroid points corresponding to the two first connected domains;

[0056] Counting the number of pixel points on the reference connection line that are the second preset value and obtaining the number of pixel points as the reference distance between the two first connected domains;

[0057] The minimum value of all reference distances is taken as the target distance R;

[0058] Let f(R / 2) be the image adjustment coefficient N, where f( ) is a floor function.

[0059] The first preset value may be 1, the second preset value may be 0, and the preset segmentation categories may include a seam category and a bolt category.

[0060] In a specific embodiment, performing N dilation processes on the basic weight image and then performing N erosion processes to obtain a target weight image includes:

[0061] For any dilation process, if it is the first dilation process, the basic weight image is used as the first temporary image; if it is not the first dilation process, the result of the last dilation process is used as the first temporary image;

[0062] The dilation processing only performs local maximum mapping on each pixel point with a pixel value of zero in the first temporary image to obtain a processing result of the dilation processing;

[0063] For any erosion process, if it is the first erosion process, the result of the Nth dilation process is used as the second temporary image; if it is not the first erosion process, the result of the last erosion process is used as the second temporary image;

[0064] The corrosion process performs local minimum mapping on all non-zero pixel points in the second temporary image to obtain a processing result of the corrosion process;

[0065] The processing result of the Nth corrosion process is used as the target weight image.

[0066] In any dilation process, local maximum mapping is performed only on each pixel point with a zero pixel value in the first temporary image. For a single pixel point with a zero pixel value, this embodiment uses a 3×3 template to extract the neighborhood of the pixel point and maps the maximum pixel value of each pixel point included in the neighborhood to the pixel point, thereby ensuring that the pixel value of the non-zero pixel point is not changed by the dilation process.

[0067] Similarly, for any corrosion process, the corrosion process only performs local minimum value mapping on each pixel point with non-zero pixel value in the second temporary image. For a single pixel point with non-zero pixel value, this embodiment uses a 3×3 size template to extract the neighborhood of the pixel point, and maps the minimum pixel value of each pixel point contained in the neighborhood to the pixel point.

[0068] In a specific embodiment, the binarization of the target weight image to obtain a target binary image includes:

[0069] The pixel values ​​of the pixels whose pixel values ​​are not zero in the target weighted image are set to a first preset value to obtain the target binary image.

[0070] The first preset value may be 1, and the target binary image may be used for subsequent connected component analysis.

[0071] In a specific embodiment, for any second connected domain, determining a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains, and the target weight image includes:

[0072] For any second connected domain, determine a first connected domain closest to the second connected domain from the M first connected domains as an associated connected domain of the second connected domain;

[0073] Calculating the intersection-and-union ratio of the second connected domain and its associated connected domain to obtain a scaling factor;

[0074] Calculate the mean of the pixel values ​​corresponding to the pixel points in the target weight image, to obtain a basic evaluation value;

[0075] The scaling coefficient corresponding to the second connected domain is multiplied by the basic evaluation value, and the multiplication result is used as the reference evaluation value of the second connected domain.

[0076] For any second connected domain, the centroid distance between the centroid point of the second connected domain and the centroid points of the M first connected domains is calculated, and the first connected domain corresponding to the minimum centroid distance is used as the associated connected domain of the second connected domain.

[0077] Specifically, the intersection-and-union ratio of the second connected domain and its associated connected domain is calculated. If the intersection-and-union ratio is not zero, it is considered that the defective area belongs to the corresponding joint structure area or the bolt structure area, and the intersection-and-union ratio is used as the scaling factor. If the intersection-and-union ratio is not zero, it is considered that the defective area belongs to the pipe segment body area, and the first preset value is used as the scaling factor, and the first preset value is 1.

[0078] In a specific embodiment, determining the deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the second connected domains includes:

[0079] An average value is calculated based on the reference evaluation values ​​corresponding to each second connected domain, and the average calculation result is used as the deformation detection evaluation value of the subway segment image.

[0080] Among them, the deformation detection evaluation value can be used to characterize the deformation degree of the current subway segment image. The implementer can combine the preset evaluation threshold to generate deformation warning information when the deformation detection evaluation value is greater than the evaluation threshold. The implementer can also use the deformation detection evaluation values ​​corresponding to multiple subway segment images obtained in time series, combined with the time series prediction model to predict the deformation detection evaluation value at a specific time point, thereby achieving early warning.

[0081] In the first embodiment of the present invention, based on the semantic segmentation results of different dimensions, the basic weight value of each pixel point is comprehensively evaluated with the pixel points as the object, and then the connected domain is used as the object for aggregation evaluation to obtain the reference evaluation value corresponding to each second connected domain, and the image adjustment coefficient is determined according to the reference distance between the first connected domains, and then the basic weight image is subjected to morphological operations to ensure that the morphological operation removes noise as much as possible without destroying the structural integrity, and then the subway segment image is used as the object for aggregation evaluation, thereby realizing the hierarchical aggregation of deformation severity and improving the reliability of subway tunnel segment deformation detection.

[0082] Example 2

[0083] This embodiment 2 provides a subway tunnel segment deformation detection device, such as Figure 2 FIG. 1 is a schematic diagram of a structure of a subway tunnel segment deformation detection device provided by a second embodiment of the present invention. The subway tunnel segment deformation detection device includes:

[0084] An image acquisition module 201 is configured to acquire a subway segment image and input the subway segment image into a trained first semantic segmentation model to obtain a first segmented image;

[0085] An image segmentation module 202 is configured to input the subway segment image into a trained second semantic segmentation model to obtain a second segmented image;

[0086] A basic weight determination module 203 is configured to determine, for any pixel point, a basic weight value corresponding to the pixel point based on a first segmentation category corresponding to the pixel point in the first segmented image and a second segmentation category corresponding to the pixel point in the second segmented image;

[0087] An image forming module 204 is configured to form a basic weight image based on the basic weight values ​​corresponding to all pixels;

[0088] A coefficient determination module 205 is configured to perform a connected domain analysis on the first segmented image to obtain M first connected domains, and determine an image adjustment coefficient N based on a reference distance between each two first connected domains, where M and N are both positive integers.

[0089] The morphological processing module 206 is configured to perform N dilation processes and then N erosion processes on the basic weight image to obtain a target weight image, and perform binarization processing on the target weight image to obtain a target binary image;

[0090] A connected component analysis module 207 is configured to perform a connected component analysis on the target binary image to obtain K second connected components, where K is a positive integer;

[0091] A connected domain evaluation module 208 is configured to determine, for any second connected domain, a reference evaluation value corresponding to the second connected domain based on the second connected domain, the M first connected domains, and the target weight image;

[0092] The deformation evaluation module 209 is configured to determine a deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the respective second connected domains.

[0093] It should be noted that the specific limitations of the subway tunnel segment deformation detection device can be found in the limitations of the subway tunnel segment deformation detection method described above and will not be repeated here. The information interaction and execution process between the aforementioned modules, as well as other details, are based on the same concept as the method embodiments of the present invention. Their specific functions and technical effects can be found in the method embodiments and will not be repeated here.

[0094] Example 3

[0095] This third embodiment provides a computer device, which may be a server. The computer device may include a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for detecting deformation of a subway tunnel segment is implemented.

[0096] Example 4

[0097] This fourth embodiment provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program implements the subway tunnel segment deformation detection method described in the above embodiment. To avoid repetition, this computer program is not described here. Alternatively, when executed by a processor, the computer program implements the functions of the various modules / units described in the above embodiment of the subway tunnel segment deformation detection device. To avoid repetition, this is not described here.

[0098] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0099] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0100] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any form. Although the present invention has been disclosed as above in terms of preferred embodiments, they are not intended to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for detecting deformation of a subway tunnel segment, characterized in that: The subway tunnel segment deformation detection method comprises the following steps: S101, collecting a subway segment image, and inputting the subway segment image into a trained first semantic segmentation model to obtain a first segmented image; S102, inputting the subway segment image into a trained second semantic segmentation model to obtain a second segmented image; S103: For any pixel point, determine a basic weight value corresponding to the pixel point according to a first segmentation category corresponding to the pixel point in the first segmented image and a second segmentation category corresponding to the pixel point in the second segmented image; S104, forming a basic weight image according to the basic weight values ​​corresponding to all the pixels; S105, performing a connected domain analysis on the first segmented image to obtain M first connected domains, and determining an image adjustment coefficient N based on a reference distance between every two first connected domains, where M and N are both positive integers; S106, performing N dilation processing on the basic weight image, and then performing N erosion processing on the basic weight image to obtain a target weight image, and performing binarization processing on the target weight image to obtain a target binary image; S107, performing connected component analysis on the target binary image to obtain K second connected components, where K is a positive integer; S108, for any second connected domain, determining a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains, and the target weight image; S109 , determining a deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the respective second connected domains.

2. The subway tunnel segment deformation detection method according to claim 1, characterized in that: The first segmentation categories include at least a segment body category, a joint category, a bolt category, a grouting hole category, and a background category; the second segmentation categories include at least a crack category, a misalignment category, a defect category, a protrusion category, a deformation category, and a non-defect category; The determining, according to the first segmentation category corresponding to the pixel in the first segmented image and the second segmentation category corresponding to the pixel in the second segmented image, a basic weight value corresponding to the pixel includes: According to the first segmentation category corresponding to the pixel point in the first segmented image and the second segmentation category corresponding to the pixel point in the second segmented image, a preset mapping table is used to map and obtain a basic weight value corresponding to the pixel point, wherein the mapping table includes a mapping relationship between category combinations and basic weight values, and the category combination includes the first segmentation category and the second segmentation category.

3. The subway tunnel segment deformation detection method according to claim 2, characterized in that: The performing of connected domain analysis on the first segmented image to obtain M first connected domains, and determining an image adjustment coefficient N based on a reference distance between every two first connected domains, includes: According to the first segmentation categories corresponding to the respective pixel points, the pixel values ​​of the pixel points belonging to the preset segmentation category are set to the first preset value, and the pixel values ​​of the pixel points not belonging to the preset segmentation category are set to the second preset value, to form a temporary binary image; Performing a connected domain analysis on the temporary binary image to obtain the M first connected domains; For any two first connected domains, determining the centroid points corresponding to the two first connected domains, and forming a reference connecting line between the centroid points corresponding to the two first connected domains; Counting the number of pixel points on the reference connection line that are the second preset value and obtaining the number of pixel points as the reference distance between the two first connected domains; The minimum value of all reference distances is taken as the target distance R; Let f(R / 2) be the image adjustment coefficient N, where f( ) is a floor function.

4. The subway tunnel segment deformation detection method according to claim 1, characterized in that: The step of performing N dilation processes and N erosion processes on the basic weight image to obtain a target weight image includes: For any dilation process, if it is the first dilation process, the basic weight image is used as the first temporary image; if it is not the first dilation process, the result of the last dilation process is used as the first temporary image; The dilation processing only performs local maximum mapping on each pixel point with a pixel value of zero in the first temporary image to obtain a processing result of the dilation processing; For any erosion process, if it is the first erosion process, the result of the Nth dilation process is used as the second temporary image; if it is not the first erosion process, the result of the last erosion process is used as the second temporary image; The corrosion process performs local minimum mapping on all non-zero pixel points in the second temporary image to obtain a processing result of the corrosion process; The processing result of the Nth corrosion process is used as the target weight image.

5. The subway tunnel segment deformation detection method according to claim 1, characterized in that: The binarization process is performed on the target weight image to obtain a target binary image, including: The pixel values ​​of the pixels whose pixel values ​​are not zero in the target weighted image are set to a first preset value to obtain the target binary image.

6. The subway tunnel segment deformation detection method according to claim 1, characterized in that: The step of determining, for any second connected domain, a reference evaluation value corresponding to the second connected domain according to the second connected domain, the M first connected domains, and the target weight image includes: For any second connected domain, determine a first connected domain closest to the second connected domain from the M first connected domains as an associated connected domain of the second connected domain; Calculating the intersection-and-union ratio of the second connected domain and its associated connected domain to obtain a scaling factor; Calculate the mean of the pixel values ​​corresponding to the pixel points in the target weight image, to obtain a basic evaluation value; The scaling coefficient corresponding to the second connected domain is multiplied by the basic evaluation value, and the multiplication result is used as the reference evaluation value of the second connected domain.

7. The subway tunnel segment deformation detection method according to claim 1, characterized in that: Determining the deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to the second connected domains includes: An average value is calculated based on the reference evaluation values ​​corresponding to each second connected domain, and the average calculation result is used as the deformation detection evaluation value of the subway segment image.

8. A subway tunnel segment deformation detection device, characterized in that: The subway tunnel segment deformation detection device comprises: An image acquisition module is used to acquire subway segment images and input the subway segment images into a trained first semantic segmentation model to obtain a first segmented image; an image segmentation module, configured to input the subway segment image into a trained second semantic segmentation model to obtain a second segmented image; a basic weight determination module, configured to determine, for any pixel point, a basic weight value corresponding to the pixel point based on a first segmentation category corresponding to the pixel point in the first segmented image and a second segmentation category corresponding to the pixel point in the second segmented image; An image forming module, configured to form a basic weighted image based on the basic weight values ​​corresponding to all pixel points; a coefficient determination module, configured to perform a connected domain analysis on the first segmented image to obtain M first connected domains, and determine an image adjustment coefficient N based on a reference distance between every two first connected domains, where M and N are both positive integers; A morphological processing module is used to perform N dilation processes on the basic weight image and then perform N erosion processes to obtain a target weight image, and to perform binarization processing on the target weight image to obtain a target binary image; A connected domain analysis module is used to perform a connected domain analysis on the target binary image to obtain K second connected domains, where K is a positive integer; a connected domain evaluation module, configured to determine, for any second connected domain, a reference evaluation value corresponding to the second connected domain based on the second connected domain, the M first connected domains, and the target weight image; The deformation evaluation module is used to determine the deformation detection evaluation value of the subway segment image according to the reference evaluation values ​​corresponding to each second connected domain.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the subway tunnel segment deformation detection method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the subway tunnel segment deformation detection method according to any one of claims 1 to 7 is implemented.

Images