Tunnel image-based disease spatial coupling integration method and system, and storage medium

Abstract

This invention discloses a method, system, and storage medium for spatial coupling and integration of tunnel defects based on tunnel imagery. The method includes: acquiring tunnel images; segmenting image layers corresponding to physical single rings based on ring joint features; assigning unique identifiers with spatial locations to form a structured image set; establishing an engineering semantic coordinate system based on ring numbers and clock angles; calculating the spatial scale by combining tunnel physics and pixel size; constructing a bidirectional coordinate mapping engine based on a polar coordinate model to achieve mutual conversion between the two types of coordinates; importing defect data; completing coordinate matching with the mapping engine; building a spatial coupling database for defects; and conducting multi-dimensional spatial coupling analysis of defects; calling the database and mapping engine to complete the visual annotation of defects on the image layers; and outputting the final integrated results. This provides reliable technical support for the safe operation of underground transportation facilities.

Description

Technical Field

[0001] This invention belongs to the field of tunnel defect detection technology, and more specifically, relates to a spatial coupling and integration method, system and storage medium for defects based on tunnel images. Background Technology

[0002] Currently, the scale of underground infrastructure construction in my country continues to expand, with a large number of tunnel projects being put into formal operation. During long-term service, tunnels are easily affected by multiple factors such as geological conditions, traffic loads, and environmental changes, making them prone to various defects such as structural deformation, foundation settlement, water leakage in the inner wall, and lining cracking. These defects not only affect the normal use of the tunnels but also gradually threaten the structural operational safety. Therefore, routine and meticulous tunnel defect detection and operation and maintenance management are of paramount importance.

[0003] With the continuous iteration and upgrading of intelligent operation and maintenance technologies, automated defect identification based on high-definition tunnel images has become the mainstream application in the tunnel inspection field. At the image acquisition level, existing equipment can complete full-area high-definition scanning of the tunnel interior, stably generating complete digital images of the tunnel that clearly present various defect features on the lining surface, providing ample data support for defect identification. In the defect identification stage, the industry generally adopts target detection algorithms to directly locate and label defect targets in the images, initially achieving automation of defect identification and effectively improving inspection efficiency.

[0004] However, the existing technological system still has significant shortcomings. The core issue lies in the failure to establish an accurate and unified correspondence between image pixel locations and the actual physical spatial locations of tunnels. In traditional methods, location descriptions commonly used on-site, such as ring numbers, angles, and structural zoning, cannot be efficiently converted into image pixel coordinates. Location information familiar to engineers is difficult to directly match with image footage, resulting in a significant gap in location representation. Furthermore, due to the lack of a unified spatial reference benchmark, defect data collected from different tunnel segment rings and different inspection periods are independent of each other, making data integration, horizontal comparison, and vertical tracing difficult, thus hindering the full exploitation of data value.

[0005] In addition, existing image recognition systems can only complete visual labeling of defects and cannot identify tunnel engineering areas such as the arch waist, invert arch, and arch crown. They also cannot conduct in-depth analysis such as regional statistics and defect correlation assessment based on tunnel structural characteristics. Currently, the industry urgently needs a new technical solution that can map the semantic location of engineering structures to the pixel space of two-dimensional images without building complex 3D models. This would break down data barriers between multiple sources of defects, support comprehensive defect analysis, improve the intelligence and precision of tunnel defect maintenance, and ensure the long-term safe and stable operation of underground transportation facilities. Summary of the Invention

[0006] This invention aims to solve the problems in existing tunnel image defect detection, such as the inability to map pixel positions to engineering physical locations and the difficulty in integrating and analyzing data. It constructs a coordinate mapping system based on tunnel images, achieving bidirectional conversion between engineering semantics and pixel coordinates, completing coupled analysis and visual annotation of defect data, and improving the precision and intelligence of tunnel defect detection and maintenance.

[0007] To address the aforementioned deficiencies or improvement needs of existing technologies, as a first aspect of this invention, the present invention provides a spatial coupling and integration method for defects based on tunnel images, comprising: S1. Acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; S2. Construct an engineering semantic coordinate system based on ring number and clock angle, and obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, construct a bidirectional coordinate mapping engine between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate; S3. Obtain tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes; S4. Based on the disease spatial coupling database and the bidirectional coordinate mapping engine, the disease attribute information is mapped to the corresponding image pixel coordinates, and a visual annotation containing the spatial location and engineering semantic features of the disease is generated on the ring image layer, and the disease spatial coupling integration result is output.

[0008] Further, in step S1, the tunnel image data is segmented based on the ring joint features to obtain multiple ring image layers corresponding to the physical single ring, specifically including: The tunnel image data is converted to grayscale, and the projection curve of the pixel grayscale value is calculated along the longitudinal direction of the tunnel. The projection curve is smoothed and filtered, and the local minimum points in the smoothed projection curve are extracted as candidate locations for the annular seam. Based on the actual physical width constraints of adjacent segments, the validity of the candidate positions of the annular seam is verified and corrected to determine the final pixel coordinates of the annular seam boundary. The tunnel image data is cropped based on the final ring seam boundary pixel coordinates to obtain the multiple ring image layers.

[0009] Furthermore, in S2, constructing an engineering semantic coordinate system based on the ring number and clock angle specifically includes: With the geometric center of the tunnel cross-section as the pole and the top of the tunnel arch as the zero point of the polar axis, the 360-degree circle is divided into continuous clock angles in a clockwise direction. Based on the stress characteristics and structural zoning of the tunnel project, the clock angle is divided into multiple angle regions with specific engineering semantic attributes. The angle regions include at least: the crown region, the left arch waist region, the right arch waist region, the left arch foot region, the right arch foot region, and the inverted arch region. Establish a mapping dictionary between each angle region and its corresponding clock angle interval, so that any clock angle can be automatically resolved into the name of the corresponding engineering structure part.

[0010] Furthermore, in step S2, based on the spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system, wherein the engine includes ring numbers. and clock angle Converting engineering semantic location to image pixel coordinates The calculation formula is: In the formula, The pixel height of the unfolded image of a single-ring tunnel; The pixel width of the unfolded image of a single-ring tunnel; The physical radius of the tunnel cross-section; For the clock angle The converted radian value, and ; This is the modulo operator.

[0011] Furthermore, in step S2, obtaining the spatial scale of the tunnel physical dimensions and image pixel dimensions specifically includes an adaptive scale calculation mechanism: Get the image resolution of the current ring image layer And the actual outer diameter of the corresponding physical pipe segment. and actual ring width ; Calculate the horizontal spatial scale and vertical spatial scale ; When a difference in resolution or physical size is detected between different ring image layers, the bidirectional coordinate mapping engine dynamically updates the... and ; Among them, the longitudinal spatial scale The configuration is as follows: when the tunnel defect data includes the longitudinal physical offset distance within the ring, the longitudinal spatial scale is used. The vertical physical offset distance within the ring is converted into a vertical pixel offset within the ring and then superimposed onto the vertical pixel coordinates. In this process, the vertical pixel-level positioning of the disease within the single-ring image layer is completed.

[0012] Furthermore, when a difference in resolution or physical size is detected between different ring image layers, the bidirectional coordinate mapping engine dynamically updates the... and Specifically, it includes: Traverse the structured image layer set and read the attribute metadata associated with each ring image layer. The attribute metadata includes at least the actual image resolution of the current ring. and the actual outer diameter of the corresponding physical pipe segment and actual ring width ,in For the ring number; The attribute metadata of the current ring is compared with the preset benchmark reference value or the attribute metadata of the adjacent previous ring. When it is determined that the deviation of resolution or physical size exceeds the preset tolerance threshold, the scale update mechanism is triggered. Based on the attribute metadata of the current ring, independently calculate the horizontal spatial scale of the current ring. and vertical spatial scale and the ring number With the corresponding , Stored as key-value pairs in the scale mapping dictionary; During the execution of the bidirectional coordinate mapping, the bidirectional coordinate mapping engine dynamically retrieves the current ring number from the scale mapping dictionary based on the target ring number where the defect data is located. and Perform pixel coordinate calculation.

[0013] Furthermore, in step S3, a multi-dimensional spatial coupling analysis of the disease is performed based on the engineering semantic attributes, specifically including at least one of the following analysis dimensions: Intra-ring coupling analysis: Within the same ring number, based on the clock angle and pixel boundary of the disease, calculate the spatial proximity between different disease types and identify clusters of associated diseases; Inter-ring coupling analysis: In multiple consecutive ring numbers, extract similar defects within the same engineering semantic attribute area, perform longitudinal connectivity calculation, and identify longitudinal through cracks or continuous seepage zones; Temporal evolution analysis: For the same spatial index location, compare the area, length or grade data of the disease at different collection periods to calculate the spatiotemporal expansion rate and deterioration trend of the disease.

[0014] Furthermore, in step S4, a visual annotation containing the spatial location and engineering semantic features of the lesion is generated on the ring image layer, specifically including an anti-overlap rendering strategy: Based on the type and severity level of the disease, the underlying drawing is performed by matching the preset primitive shape and color coding rules; Generate semantic tags containing ring number, clock angle and engineering part name next to the graphic element; When multiple semantic tags are detected to have overlapping bounding boxes in the same local pixel area, an anti-overlap method is triggered: the tags are sorted according to the priority of the disease level, the high-priority tags are kept in their original positions, the low-priority tags are offset and pulled radially or vertically, and connected to the corresponding disease primitives through the lead wire.

[0015] As a second aspect of the present invention, a spatially coupled integrated system for defects based on tunnel images is also provided, comprising: The image segmentation and layer construction unit is used to acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; The coordinate construction and bidirectional mapping unit is used to construct an engineering semantic coordinate system based on ring number and clock angle, and to obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate. The defect database construction and coupling analysis unit is used to acquire tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes. The visualization annotation and output unit is used to map disease attribute information to corresponding image pixel coordinates based on the disease spatial coupling database and the bidirectional coordinate mapping engine, generate visualization annotations containing disease spatial location and engineering semantic features on the ring image layer, and output the disease spatial coupling integration results.

[0016] As a third aspect of the invention, a computer-readable storage medium is also provided, on which a computer program is stored, which is executed by a processor as described in any one of the claims, a method for spatial coupling and integration of defects based on tunnel images.

[0017] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: 1. The tunnel image-based spatial coupling integration method of this invention segments tunnel images based on annular joint features, determines the annular joint boundaries by combining grayscale projection, smoothing filtering, and physical size verification, and cuts the overall image into image layers corresponding to the physical single ring of the tunnel. Each layer is then assigned a unique identifier with spatial location information, building a structured image layer set. This method establishes a one-to-one correspondence between continuous tunnel images and the actual tunnel segment structure, standardizes and organizes the original image data, eliminates the problem of disorganized image data, and provides standardized and locatable basic image data for subsequent coordinate transformation, defect matching, and data analysis, ensuring that all subsequent operations can be carried out in an orderly manner based on the actual physical structure of the tunnel.

[0018] 2. The tunnel image-based spatial coupling and integration method of this invention constructs an engineering semantic coordinate system centered on ring numbers and clock angles. It calculates an adaptive spatial scale by combining tunnel physical dimensions and image pixel dimensions, and builds a bidirectional coordinate mapping engine between the two coordinate systems based on a polar coordinate transformation model, enabling the mutual conversion between engineering semantic locations and image pixel coordinates. This method establishes a unified spatial reference system, allowing commonly used structural location descriptions in the engineering field to correspond to image pixel locations. It is also adaptable to tunnel ring images of different resolutions and physical specifications, solving the problems of unintuitive location descriptions and inconsistent spatial benchmarks in traditional technologies. Furthermore, it provides stable and reliable conversion capabilities for cross-coordinate matching of defect data.

[0019] 3. The tunnel image-based spatial coupling integration method of this invention synchronously maps defect data to two coordinate systems using a bidirectional coordinate mapping engine, constructing a defect spatial coupling database indexed by ring number and clock angle. Multi-dimensional spatial coupling analysis of defects is conducted based on engineering semantic attributes. The mapping engine then annotates defect information on the ring image layer and completes the visualization output. This method achieves integrated storage and correlation analysis of defect data across multiple time periods and ring segments, objectively identifying the distribution patterns and development trends of defects. Simultaneously, it presents defect information using appropriate rendering methods, overcoming the shortcomings of traditional engineering analysis capabilities and fully realizing the entire process of tunnel defect data integration, analysis, and intuitive display. Attached Figure Description

[0020] Figure 1 This is a flowchart of a spatial coupling and integration method for defects based on tunnel images, according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the effect of pixel coordinates being projected into a scanned image according to an embodiment of the present invention; Figure 3This is a schematic diagram of the system units in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0022] Example 1 Please refer to Figure 1 This embodiment 1 provides a spatial coupling and integration method for defects based on tunnel images, including: S1. Acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; S2. Construct an engineering semantic coordinate system based on ring number and clock angle, and obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, construct a bidirectional coordinate mapping engine between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate; S3. Obtain tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes; S4. Based on the disease spatial coupling database and the bidirectional coordinate mapping engine, the disease attribute information is mapped to the corresponding image pixel coordinates, and a visual annotation containing the spatial location and engineering semantic features of the disease is generated on the ring image layer, and the disease spatial coupling integration result is output.

[0023] This embodiment 1 further elaborates on the above steps.

[0024] (1) Image segmentation and layer construction In this embodiment, obtaining the original tunnel image data is a fundamental prerequisite for constructing a structured image layer set. Let the obtained original tunnel image data be... ,in Represents the horizontal pixel coordinates of the image. This represents the longitudinal pixel coordinates of the image along the tunnel's longitudinal direction. To eliminate color interference and highlight the physical characteristics of light and shadow at the annular joint, it is necessary to modify this original tunnel image data. Perform grayscale conversion to obtain a grayscale image. After obtaining the grayscale image, the projection curve of the pixel grayscale value is calculated along the longitudinal direction of the tunnel. The projection curve The calculation process involves converting grayscale images... In each row of vertical coordinates All horizontal pixel coordinates on The corresponding grayscale values ​​are summed, and the horizontal range of the summation is limited by the total horizontal pixel width of the image. .

[0025] Since directly calculated projection curves often contain noise interference caused by dirt on the tunnel surface or uneven lighting, smoothing filtering must be performed on the projection curves to extract candidate locations for the annular joint. A kernel size is set. The smoothing filter for the projection curve Perform convolution operation to obtain the projection curve after smoothing and filtering. The smoothed projection curve The algorithm iterates upwards to find local minima where the function value is less than the function value of its neighboring ordinate points. The ordinate values ​​corresponding to these local minima are then extracted to form a set of candidate locations for the annular seam. Each element in the set Representing the The vertical pixel coordinates of each candidate slit position in the image.

[0026] Candidate locations extracted solely based on the minimum grayscale values ​​of an image are susceptible to artifacts. Therefore, it is necessary to introduce prior physical knowledge to filter and determine the true annular joint boundary. This involves obtaining the actual physical width of adjacent tunnel segments. and the vertical spatial resolution during image acquisition The actual physical width With vertical spatial resolution Divide to calculate the theoretical pixel spacing of the annular seam. Set a proportional coefficient to control the spacing tolerance. Constructing a system based on the theoretical pixel spacing of the circumferential seam Centered on, with proportional coefficient Compared with the theoretical annular slit pixel spacing The product represents the validity verification interval within the tolerance range. Traverse the set of candidate circumferential seam locations. elements in The actual pixel spacing between adjacent candidate positions is calculated. If the actual pixel spacing falls within the validity verification interval, the corresponding candidate position is determined to be valid; otherwise, position correction or elimination is performed. After verification and correction, the final set of pixel coordinates of the circumferential seam boundary is determined. The elements in this set Representing the The final determined vertical pixel coordinates of the slit boundary in the image.

[0027] Once the boundaries of each physical ring in the image are clearly defined, the elongated continuous image can be physically segmented to generate independent layer files. This is based on the final set of pixel coordinates of the ring boundaries. Middle adjacent elements and The defined vertical pixel range is relevant to the original tunnel image data. or grayscale image Perform a full-width horizontal crop to obtain multiple loop image layers corresponding to the physical single loop. subscript This represents the physical arrangement number of the ring image layer along the longitudinal direction of the tunnel.

[0028] To imbue the cropped discrete images with engineering spatial properties, each ring image layer needs to be assigned a spatially recognizable label. Assign a unique identifier containing spatial location information This unique identifier It encapsulates the tunnel design mileage information, segment assembly ring number, and relative index of the layer within the overall imagery corresponding to the current panoramic image layer. All elements are accompanied by unique identifiers. Ring image layer By aggregating and associating images, a structured collection of image layers is constructed.

[0029] (2) Coordinate construction and bidirectional mapping Constructing an engineering semantic coordinate system based on ring numbers and clock angles is the core foundation for realizing the spatial location of tunnel defects. First, with the geometric center of the tunnel cross-section as the pole and the top of the tunnel arch as the polar zero, the 360-degree circle is divided into continuous clock angles in a clockwise direction. For example, 12 o'clock = 0°, 3 o'clock = 90°, 6 o'clock = 180°, and 9 o'clock = 270°. Based on the stress characteristics and structural zoning of the tunnel engineering, these clock angles are... The system is divided into multiple angular regions with specific engineering semantic attributes. These angular regions include at least: the crown region, the left arch waist region, the right arch waist region, the left arch foot region, the right arch foot region, and the invert arch region. A mapping dictionary is established between each angular region and its corresponding clock angle range. For example, the crown region corresponds to the 11 o'clock to 1 o'clock direction, i.e., the angle range [330º, 30º]; the left arch waist region corresponds to the 8 o'clock to 10 o'clock direction, i.e., the angle range [240º, 300º]; and the right arch waist region corresponds to the 2 o'clock to 4 o'clock direction, i.e., the angle range [60º, 120º]. ​​This mapping dictionary allows for any clock angle... All of them can be automatically resolved into the corresponding names of engineering structural parts, thus giving clear engineering physical meaning to the pure geometric perspective.

[0030] After establishing the engineering semantic coordinate system, it is necessary to obtain the physical dimensions of the tunnel and the pixel dimensions of the images to calculate the spatial scale, and introduce an adaptive scale calculation mechanism to cope with the heterogeneity of the data source. The structured image layer set is traversed, and the attribute metadata associated with each ring image layer is read. This attribute metadata at least includes the actual image resolution of the current ring. and the actual outer diameter of the corresponding physical pipe segment and actual ring width ,in This refers to the ring number. The attribute metadata of the current ring is compared with a preset benchmark value or the attribute metadata of the adjacent previous ring. When the deviation in resolution or physical size exceeds a preset tolerance threshold, a scale update mechanism is triggered. Based on the attribute metadata of the current ring itself, the lateral spatial scale of the current ring is independently calculated. and vertical spatial scale The ring number Corresponding horizontal spatial scale Vertical spatial scale The data is stored as key-value pairs in the scale mapping dictionary, enabling the bidirectional coordinate mapping engine to dynamically retrieve the current ring number from the scale mapping dictionary based on the target ring number where the disease data is located. and Perform pixel coordinate calculation.

[0031] Based on the calculated spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic positions and image pixel coordinates. This includes the ring number... and clock angle Converting engineering semantic location to image pixel coordinates When calculating, the following formula is used: Vertical pixel coordinates Horizontal pixel coordinates In the formula, The pixel height of the unfolded image of a single-ring tunnel; The pixel width of the unfolded image of a single-ring tunnel; The physical radius of the tunnel cross-section; For the clock angle The converted radian value, and satisfying ; The modulo operator is used to ensure the periodic closure of the horizontal pixel coordinates within the image width range.

[0032] For cases where the disease is not absolutely located at the edge of the annular seam but exhibits longitudinal distribution differences within a single ring, the longitudinal spatial scale is used. This is configured as a key parameter for processing intra-loop positioning. When the tunnel defect data includes the intra-loop longitudinal physical offset distance, the longitudinal spatial scale is used. The intra-ring vertical physical offset distance is converted into an intra-ring vertical pixel offset, and this intra-ring vertical pixel offset is superimposed on the vertical pixel coordinates. In this process, the vertical pixel-level positioning of the disease within the single-ring image layer is completed.

[0033] (3) Disease database construction and coupling analysis Obtaining tunnel defect data is a prerequisite for initiating spatial coupling analysis and database construction. Let the obtained tunnel defect data set be... Each disease record in this set is denoted as This includes disease type identifiers. Ring number Initial clock angle and the set of diseased pixel boundaries The built bidirectional coordinate mapping engine is invoked to record the defects. The ring number in relative to the initial clock angle Mapped to horizontal pixel coordinates in the image pixel coordinate system with vertical pixel coordinates At the same time, the initial clock angle is determined using a mapping dictionary. Analysis into engineering structural component names The horizontal pixel coordinates generated by the mapping Vertical pixel coordinates Names of structural components in the project Compared with the original disease record To merge and construct based on the ring number and initial clock angle A spatially coupled database of diseases with a combined spatial index.

[0034] Intra-ring coupling analysis was conducted using a spatial coupling database of defects to uncover associated relationships among defects within the same pipe segment ring. Within a specified target ring number, the initial clock angle of all defect records was extracted. Set of pixels with disease boundary The spatial proximity between any two records of different disease types is calculated. The spatial proximity is calculated based on the ratio of the minimum Euclidean distance between the boundary sets of two disease pixels to a preset proximity threshold. When the spatial proximity meets the preset aggregation condition, the corresponding multiple disease records are aggregated and marked as a cluster of co-existing diseases, thereby revealing the symbiotic mechanism of diseases such as seepage and cracks within the same ring.

[0035] Inter-ring coupling analysis was performed based on the longitudinal structural characteristics of the tunnel to identify continuous defects that develop across rings. Within multiple consecutive ring number sequences, components sharing the same structural name were selected. And disease type identification For the same subset of disease records, the longitudinal connectivity of the diseases within that subset is calculated. The quantification of longitudinal connectivity depends on the pixel overlap and longitudinal spacing of adjacent disease ring boundaries. When the pixel overlap is greater than the overlap threshold or the longitudinal spacing is less than the spacing threshold, it is determined that the diseases of adjacent rings have a spatial connectivity relationship. The connected disease sequence is then identified as a longitudinally penetrating crack or a continuous seepage zone, providing a quantitative basis for assessing the overall structural stability of the tunnel in the longitudinal dimension.

[0036] A time dimension is introduced to perform temporal evolution analysis to understand the dynamic development patterns of diseases. This is applied to the spatially coupled disease database, which is based on the ring number... and initial clock angle For the same spatial index location, quantitative disease characteristic data corresponding to different collection periods are extracted. This quantitative disease characteristic data includes disease area, disease length, or disease severity. The difference in the quantitative disease characteristic data within the time interval is calculated, and the spatiotemporal spread rate of the disease is calculated in combination with the time interval. At the same time, the deterioration trend of the disease is determined based on the changing direction of the quantitative disease characteristic data, ultimately forming a complete multidimensional coupled analysis result of the disease covering spatial distribution and temporal evolution.

[0037] (4) Visual annotation and output of results Please refer to Figure 2 Based on the constructed disease spatial coupling database and bidirectional coordinate mapping engine, the visualization rendering of the disease spatial coupling integration results is performed, mapping the abstract disease attribute information set to the panoramic image layer. In a two-dimensional physical space. First, a set of disease attribute information is extracted one by one from the disease spatial coupling database. This set includes disease types. Severity level and image pixel coordinates Call the bidirectional coordinate mapping engine to verify the image pixel coordinates. In the panoramic image layer After confirming the effectiveness, based on the type of disease. Severity level Match the corresponding primitive shape and color code in the preset primitive configuration library, and use image pixel coordinates. As anchor points, in the surround image layer The underlying layer completes the instantiation and drawing of disease primitives.

[0038] After the underlying defect features are drawn, text annotations containing the spatial location and engineering semantic features of the defects need to be generated next to the features to assist in manual interpretation. For each drawn defect feature, its associated ring number is extracted. Clock angle and the name of the engineering part These three core semantic information items are concatenated and combined to generate corresponding semantic tags. The semantic tags are then calculated on the surrounding image layer. The 2D pixel bounding box used during rendering is defined as the bounding box. All bounding boxes in the current viewport or a specified local pixel region are traversed, and a pairwise collision detection algorithm is executed. When any two or more bounding boxes are detected to have geometric intersection within the local pixel region, it is determined that there is semantic label overlap, and an anti-overlap rendering strategy is triggered to ensure the clarity of visual information.

[0039] The core of the anti-overlap rendering strategy lies in differentiated layout adjustments and visual guidance based on the severity of the damage. This involves extracting the severity level corresponding to all overlapping semantic tags. The semantic tags are then converted into quantifiable disease priority levels. Based on these priority levels, overlapping semantic tags are sorted in descending order, with tags of high priority displayed in their initial calculated positions. For the remaining low-priority tags, an anti-overlap offset traction vector is calculated. This vector is strictly limited to either the radial direction along the tunnel cross-section or the longitudinal direction along the tunnel axis. The low-priority tags are translated along this vector until their bounding boxes are no longer in contact with any other bounding boxes. After the low-priority tags have been offset, the shortest unobstructed path between the center point of the offset semantic tag and the corresponding disease element anchor point is automatically calculated. A leader line is then drawn along this path, and finally, the complete image, including the anti-overlap layout and the leader line connection, is encapsulated and output as a standard format disease spatial coupling integration result.

[0040] Example 2 Please refer to Figure 3 This embodiment 2 provides a spatially coupled integrated system for defects based on tunnel images, including: The image segmentation and layer construction unit is used to acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; The coordinate construction and bidirectional mapping unit is used to construct an engineering semantic coordinate system based on ring number and clock angle, and to obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate. The defect database construction and coupling analysis unit is used to acquire tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes. The visualization annotation and output unit is used to map disease attribute information to corresponding image pixel coordinates based on the disease spatial coupling database and the bidirectional coordinate mapping engine, generate visualization annotations containing disease spatial location and engineering semantic features on the ring image layer, and output the disease spatial coupling integration results.

[0041] Example 3 This embodiment 3 also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement any step of a spatial coupling integration method for defects based on tunnel images.

[0042] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0043] For a description of the computer-readable storage medium provided in this application, please refer to the above method embodiments; further details will not be repeated here.

[0044] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A spatially coupled integrated method for defects based on tunnel images, characterized in that, include: S1. Acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; S2. Construct an engineering semantic coordinate system based on ring number and clock angle, and obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, construct a bidirectional coordinate mapping engine between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate; S3. Obtain tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes; S4. Based on the disease spatial coupling database and the bidirectional coordinate mapping engine, the disease attribute information is mapped to the corresponding image pixel coordinates, and a visual annotation containing the spatial location and engineering semantic features of the disease is generated on the ring image layer, and the disease spatial coupling integration result is output.

2. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S1, the tunnel image data is segmented based on the ring joint features to obtain multiple ring image layers corresponding to the physical single ring, specifically including: The tunnel image data is converted to grayscale, and the projection curve of the pixel grayscale value is calculated along the longitudinal direction of the tunnel. The projection curve is smoothed and filtered, and the local minimum points in the smoothed projection curve are extracted as candidate locations for the annular seam. Based on the actual physical width constraints of adjacent segments, the validity of the candidate positions of the annular seam is verified and corrected to determine the final pixel coordinates of the annular seam boundary. The tunnel image data is cropped based on the final ring seam boundary pixel coordinates to obtain the multiple ring image layers.

3. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S2, constructing an engineering semantic coordinate system based on the ring number and clock angle specifically includes: With the geometric center of the tunnel cross-section as the pole and the top of the tunnel arch as the zero point of the polar axis, the 360-degree circle is divided into continuous clock angles in a clockwise direction. Based on the stress characteristics and structural zoning of the tunnel project, the clock angle is divided into multiple angle regions with specific engineering semantic attributes. The angle regions include at least: the crown region, the left arch waist region, the right arch waist region, the left arch foot region, the right arch foot region, and the inverted arch region. Establish a mapping dictionary between each angle region and its corresponding clock angle interval, so that any clock angle can be automatically resolved into the name of the corresponding engineering structure part.

4. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S2, based on the spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system, which includes ring numbers. and clock angle Converting engineering semantic location to image pixel coordinates The calculation formula is: In the formula, The pixel height of the unfolded image of a single-ring tunnel; The pixel width of the unfolded image of a single-ring tunnel; The physical radius of the tunnel cross-section; For the clock angle The converted radian value, and ; This is the modulo operator.

5. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S2, the spatial scale for calculating the tunnel physical dimensions and image pixel dimensions is obtained, specifically including an adaptive scale calculation mechanism: Get the image resolution of the current ring image layer And the actual outer diameter of the corresponding physical pipe segment. and actual ring width ; Calculate the horizontal spatial scale and vertical spatial scale ; When a difference in resolution or physical size is detected between different ring image layers, the bidirectional coordinate mapping engine dynamically updates the... and ; Among them, the longitudinal spatial scale The configuration is as follows: when the tunnel defect data includes the longitudinal physical offset distance within the ring, the longitudinal spatial scale is used. The vertical physical offset distance within the ring is converted into a vertical pixel offset within the ring and then superimposed onto the vertical pixel coordinates. In this process, the vertical pixel-level positioning of the disease within the single-ring image layer is completed.

6. The method for spatial coupling and integration of defects based on tunnel images according to claim 5, characterized in that, When a difference in resolution or physical size is detected between different ring image layers, the bidirectional coordinate mapping engine dynamically updates the... and Specifically, it includes: Traverse the structured image layer set and read the attribute metadata associated with each ring image layer. The attribute metadata includes at least the actual image resolution of the current ring. and the actual outer diameter of the corresponding physical pipe segment and actual ring width ,in For the ring number; The attribute metadata of the current ring is compared with the preset benchmark reference value or the attribute metadata of the adjacent previous ring. When it is determined that the deviation of resolution or physical size exceeds the preset tolerance threshold, the scale update mechanism is triggered. Based on the attribute metadata of the current ring, independently calculate the horizontal spatial scale of the current ring. and vertical spatial scale and the ring number With the corresponding , Stored as key-value pairs in the scale mapping dictionary; During the execution of the bidirectional coordinate mapping, the bidirectional coordinate mapping engine dynamically retrieves the current ring number from the scale mapping dictionary based on the target ring number where the defect data is located. and Perform pixel coordinate calculation.

7. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S3, a multi-dimensional spatial coupling analysis of the disease is performed based on the engineering semantic attributes, specifically including at least one of the following analysis dimensions: Intra-ring coupling analysis: Within the same ring number, based on the clock angle and pixel boundary of the disease, calculate the spatial proximity between different disease types and identify clusters of associated diseases; Inter-ring coupling analysis: In multiple consecutive ring numbers, extract similar defects within the same engineering semantic attribute area, perform longitudinal connectivity calculation, and identify longitudinal through cracks or continuous seepage zones; Temporal evolution analysis: For the same spatial index location, compare the area, length or grade data of the disease at different collection periods to calculate the spatiotemporal expansion rate and deterioration trend of the disease.

8. The method for spatial coupling and integration of defects based on tunnel images according to claim 1, characterized in that, In step S4, a visual annotation containing the spatial location of the lesion and engineering semantic features is generated on the ring image layer, specifically including an anti-overlap rendering strategy: Based on the type and severity level of the disease, the underlying drawing is performed by matching the preset primitive shape and color coding rules; Generate semantic tags containing ring number, clock angle and engineering part name next to the graphic element; When multiple semantic tags are detected to have overlapping bounding boxes in the same local pixel area, an anti-overlap method is triggered: the tags are sorted according to the priority of the disease level, the high-priority tags are kept in their original positions, the low-priority tags are offset and pulled radially or vertically, and connected to the corresponding disease primitives through the lead wire.

9. A spatially coupled integrated system for defects based on tunnel images, characterized in that, include: The image segmentation and layer construction unit is used to acquire tunnel image data, segment the tunnel image data based on ring joint features to obtain multiple ring image layers corresponding to a physical single ring, and assign a unique identifier containing spatial location information to each ring image layer to construct a structured image layer set; The coordinate construction and bidirectional mapping unit is used to construct an engineering semantic coordinate system based on ring number and clock angle, and to obtain the spatial scale for calculating the tunnel physical size and image pixel size; based on the spatial scale and polar coordinate transformation model, a bidirectional coordinate mapping engine is constructed between the engineering semantic coordinate system and the image pixel coordinate system to complete the mutual conversion between engineering semantic position and image pixel coordinate. The defect database construction and coupling analysis unit is used to acquire tunnel defect data, map the defect data to the engineering semantic coordinate system and the image pixel coordinate system through the bidirectional coordinate mapping engine, construct a defect spatial coupling database with ring number and clock angle as spatial index, and perform multi-dimensional defect spatial coupling analysis based on the engineering semantic attributes. The visualization annotation and output unit is used to map disease attribute information to corresponding image pixel coordinates based on the disease spatial coupling database and the bidirectional coordinate mapping engine, generate visualization annotations containing disease spatial location and engineering semantic features on the ring image layer, and output the disease spatial coupling integration results.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor as described in any one of claims 1-8: a spatial coupling integration method for defects based on tunnel images.

Images