A method and system for detecting the quality of a spiral-wound inner liner pipe

Abstract

The application discloses a kind of spiral winding inner lining pipe quality detection method and system, belong to the technical field of trenchless repair detection;Including the following steps: after shield segment assembly, before winding, collect segment inner surface point cloud and image, establish baseline model containing geometric parameters and defects;Get forming area image in winding construction, through cylindrical development and image analysis, detect pitch anomaly, trajectory drift and locking edge lap defect;After grouting, collect completion point cloud and image, identify apparent defect and roundness, center offset, ring joint misalignment and other deformation anomaly area, and difference with baseline, identify wall after grouting or grouting anomaly in combination with grouting parameter;Construction defects are retested in completion data, distinguish between closed and not closed;Not closed construction defects, apparent defects, deformation anomaly area and grouting anomaly are mapped to three-dimensional point cloud, to generate defect map containing defect type, mileage, azimuth and size;The application realizes whole-process closed-loop quality control and digital delivery.

Description

Technical Field

[0001] This invention relates to the field of trenchless repair and testing technology, and in particular to a method and system for quality testing of spiral wound lining pipes. Background Technology

[0002] Spiral winding lining is a trenchless repair technique used inside shield tunnels. It involves spirally winding strip material onto the inner surface of existing shield tunnel segments and securing the edges with locking mechanisms. Grout is then injected into the annulus between the lining and the segments to form a functional layer that shares the load with the existing structure. In sensitive conditions such as deep drainage tunnels, this technique requires precise fit between the lining and the segments, and the construction process must be continuous over the entire length in confined spaces. Therefore, quality inspection becomes a crucial element in ensuring structural safety and functional compliance.

[0003] In existing technologies, the quality inspection of spiral wound linings mainly refers to standards such as ASTM F1741 and F1698, relying on material compliance inspections and random checks after installation. In terms of tunnel structure inspection, machine vision technology has been used to identify surface defects such as cracks and leaks in concrete linings, and 3D point cloud technology has been used for convergence measurement of tunnel cross sections. However, when spiral wound linings are applied inside shield tunnel segments, the above-mentioned detection methods have the following shortcomings: First, traditional spot checks and manual inspections cannot cover the entire process and full-section quality acceptance of long-distance tunnels, and it is difficult to unify the process parameters of the winding construction (such as pitch and edge continuity), the geometric shape of the lining after forming (such as roundness and misalignment), and the appearance defects into the same data chain, making it impossible to form a traceable closed-loop quality control; Second, existing visual inspection and point cloud inspection are mostly used in isolation, lacking an integrated monitoring and fusion positioning method for the combined process of "shield tunnel segment assembly - spiral winding construction - grouting", making it difficult to distinguish between the inherent errors of segment assembly and the new deviations added by the lining construction, and also unable to accurately locate the surface defects identified by vision in the three-dimensional point cloud space, resulting in low defect verification efficiency and incomplete acceptance evidence chain. Summary of the Invention

[0004] The purpose of this invention is to provide a quality inspection method and system for spiral wound inner lining tubes, which realizes closed-loop quality control and digital delivery throughout the entire process.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide a method for quality inspection of spirally wound liner-lined tubes, comprising the following steps: After the shield tunnel segments are assembled and before the spiral winding construction, the initial three-dimensional point cloud data and initial image data of the inner surface of the segments are obtained, and a baseline model containing cross-sectional geometric parameters and existing defect information is established based on the initial three-dimensional point cloud data and initial image data. During the spiral winding process, construction image data of the winding area is acquired; the construction image data is then subjected to cylindrical unfolding and image analysis to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping. After grouting is completed, the as-built 3D point cloud data and as-built image data of the inner wall of the lining are acquired; the appearance defects are identified based on the as-built image data; the geometric indices of roundness, center offset and circumferential joint misalignment of the lining are analyzed based on the as-built 3D point cloud data to identify areas of abnormal geometric deformation; and the as-built 3D point cloud data is spatially registered and geometrically differs from the baseline model. Combined with the process parameters during the grouting process, areas of voids or abnormal grouting behind the wall are identified. Construction defects detected during construction are re-measured using as-built image data and as-built 3D point cloud data. Based on the re-measurement results, closed construction defects are distinguished from unclosed construction defects. Unclosed construction defects, apparent defects identified after grouting, areas of abnormal geometric deformation, and areas of hollowness or abnormal grouting behind the wall are uniformly mapped to a three-dimensional point cloud space constructed based on as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth, and size information.

[0006] In some optional embodiments, the specific steps for performing cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping are as follows: A cylindrical unfolding transformation is performed on each frame of the construction image data to map the spiral winding lines on the inner surface of the tunnel into oblique straight lines on a two-dimensional plane. Monitor the slope change of the straight line relative to the preset nominal slope. When the slope change exceeds the preset threshold, it is determined to be an abnormal winding pitch. Obtain the position offset of the oblique straight line relative to the nominal path. When the position offset exceeds the preset allowable deviation, it is determined to be a winding trajectory drift. The construction image data is input into a deep learning instance segmentation network to segment and extract the seam lines or lap lines between adjacent strips, and to identify the break point, break length and opening width of the seam lines or lap lines. Based on the length of the break and the width of the opening, the severity of the discontinuity in the edge locking or overlapping is determined, and a corresponding level of continuity alarm signal is output.

[0007] In some optional embodiments, the geometric indices of the lining's roundness, center offset, and circumferential joint misalignment are analyzed based on the as-built 3D point cloud data to identify areas of abnormal geometric deformation. The specific steps are as follows: Continuous cross-sections are extracted from the completed 3D point cloud data along the tunnel's direction of travel. Perform circular or elliptical fitting on the point cloud of each cross section to obtain the roundness or ellipticity of the cross section, as well as the offset of the cross section center relative to the design axis. The misalignment of the point cloud at the circumferential joint of adjacent segments is obtained; Generate index curves along the mileage direction for roundness or ellipticity, center offset, and misalignment. The roundness or ellipticity, center offset and misalignment of each mileage section are expanded according to the circumferential orientation, and the index values ​​of each circumferential position are mapped by color to generate a circumferential heat map. Based on the index curve and the circumferential heat map, areas with roundness or ellipticity, center offset and misalignment exceeding the preset threshold and exhibiting spatial clustering are identified as areas of geometric deformation anomalies.

[0008] In some optional embodiments, the steps of spatially registering and geometrically analyzing the as-built 3D point cloud data with the baseline model, and combining the process parameters during grouting to identify areas of voids or abnormal grouting behind the wall, are as follows: Spatial registration is performed between the completed 3D point cloud data and the initial 3D point cloud data in the baseline model, so that the two are in the same coordinate system. Obtain the normal distance field from the registered completed point cloud surface to the initial point cloud surface. Based on the difference between the normal distance field and the sum of the inner lining wall thickness and the grouting layer design thickness, determine the geometric difference behind the wall. Areas where the geometric difference behind the wall exceeds the preset allowable range are designated as candidate areas for anomalies behind the wall. Retrieve the grouting process curve for the corresponding mileage segment. The grouting process curve includes the changes in grouting pressure, flow rate, and cumulative grouting volume over time. When the geometric difference behind the wall in the candidate area of ​​abnormality behind the wall is less than the design thickness, and the pressure is consistently low and the cumulative grouting volume exceeds the theoretical grouting volume during the grouting process, it is determined to be a hollow area behind the wall. When the geometric difference behind the wall of the candidate area of ​​abnormality is greater than the design thickness, and abnormal pressure fluctuations or peak values ​​exceed the preset value during grouting, it is determined to be an abnormal grouting area.

[0009] In some optional embodiments, the steps for re-measuring construction defects detected in construction image data in as-built 3D point cloud data and as-built image data are as follows: The location information of the two-dimensional image of the construction defect is mapped to the as-built three-dimensional point cloud space through the extrinsic parameter calibration parameters between the camera and the laser scanner to determine the spatial location of the defect. In the as-built image data, check whether the area corresponding to the spatial location has the same defect features; In the as-built 3D point cloud data, check whether the geometry of the spatial location meets the design tolerance requirements.

[0010] In some optional embodiments, the specific steps for distinguishing between closed and unclosed construction defects based on the retest results are as follows: If the construction defect no longer exists in the as-built image data and the as-built 3D point cloud data, or the defect size has been reduced to the preset allowable range, it is determined to be a closed construction defect. Otherwise, it is judged as an unclosed construction defect.

[0011] In some optional embodiments, the unclosed construction defects, apparent defects identified after grouting, areas of abnormal geometric deformation, and areas of hollowness or abnormal grouting behind the wall are uniformly mapped to a three-dimensional point cloud space constructed based on the as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth, and size information. The specific steps are as follows: Assign a unique defect instance identifier to each defect; Record the defect type, mileage coordinates, circumferential azimuth angle, geometric dimensions, and severity level for each defect; Record the closure status of the re-measurement of construction defects; The 3D bounding boxes or masks of each defect are overlaid and rendered on the 3D model constructed from the as-built 3D point cloud data to generate an interactive 3D defect map.

[0012] In some optional embodiments, the cross-sectional geometric parameters include cross-sectional ellipticity, circumferential joint misalignment, and cross-sectional center offset; the existing defect information includes the distribution and size of cracks, damage, and leakage marks on the segment surface.

[0013] In some optional embodiments, the defect 3D map, along with the full-line point cloud model, geometric index curves along the mileage, defect list reports, and quality traceability records from process early warning, rework handling to re-inspection closure, are also included as digital delivery results.

[0014] Embodiments of the present invention also provide a quality inspection system for spirally wound lined tubes, characterized in that the system comprises: The baseline establishment module is used to acquire the initial three-dimensional point cloud data and initial image data of the inner surface of the tunnel segment after the shield segment is assembled and before the spiral winding construction. Based on the initial three-dimensional point cloud data and initial image data, a baseline model containing cross-sectional geometric parameters and existing defect information is established. The process monitoring module is used to acquire construction image data of the winding and forming area during the spiral winding construction process; it performs cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping. The as-built analysis module is used to acquire as-built 3D point cloud data and as-built image data of the inner wall of the lining after grouting; identify apparent defects based on the as-built image data; analyze the geometric indices of the lining's roundness, center offset, and circumferential joint misalignment based on the as-built 3D point cloud data, and identify areas of abnormal geometric deformation; and perform spatial registration and geometric difference analysis between the as-built 3D point cloud data and the baseline model, combined with the process parameters during the grouting process, to identify areas of voids or abnormal grouting behind the wall. The retesting and verification module is used to retest the construction defects detected in the construction image data in the as-built 3D point cloud data and as-built image data, and to distinguish between closed and unclosed construction defects based on the retesting results. The defect map generation module is used to uniformly map unclosed construction defects, apparent defects, abnormal geometric deformation areas, and areas with hollow walls or abnormal grouting to a three-dimensional point cloud space constructed based on as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth angle, and size information.

[0015] The quality inspection method for spirally wound lined tubes provided by this invention has at least the following beneficial effects: This invention integrates the initial acquisition of shield tunnel segments, monitoring of the winding process, and post-grouting final inspection into a unified quality control closed loop, achieving end-to-end data connectivity and traceability management. Through spatiotemporal fusion of multi-source sensor data, winding process parameters, lining geometry, and apparent defects are unified onto the same spatial benchmark, establishing an evidence chain of "process discovery—final inspection—defect location": winding defects detected during construction are automatically re-inspected at the final stage, distinguishing between closed and unclosed defects; apparent defects, abnormal geometric deformation areas, and areas of hollowness or grouting abnormality identified during final inspection are assigned precise mileage, azimuth, and dimensional information and mapped to a three-dimensional point cloud space, forming a structured defect map. This method overcomes the coverage blind spots and subjective reliance of traditional spot checks and manual inspections in long-distance confined spaces, providing objective, quantifiable, and traceable digital deliverables for final acceptance. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0017] Figure 1 This is a flowchart of a quality inspection method for spiral wound inner lining tubes according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a shield tunnel spiral winding lining repair structure according to an embodiment of the present invention; Figure 3This is a schematic diagram of the shield tunnel spiral winding lining construction process according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating key indicators and implementation methods for quality inspection of spiral wound inner lining of shield tunnels according to an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the principle of spiral winding construction cylinder unfolding and transformation according to an embodiment of the present invention; Figure 6 This is a schematic diagram of a three-dimensional point cloud acquisition device for detecting the quality of spiral wound inner lining of a shield tunnel according to an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0019] One embodiment of the present invention relates to a quality inspection method for spirally wound lined tubes. The implementation details of the quality inspection method for spirally wound lined tubes of this embodiment are described in detail below. The following implementation details are provided for ease of understanding and are not necessary for implementing this solution.

[0020] The specific process of the quality inspection method for spiral wound lining tubes in this embodiment can be as follows: Figure 1 As shown, it includes: Step 101: Obtain the initial three-dimensional point cloud data and initial image data of the inner surface of the shield tunnel segment after the segment is assembled and before the spiral winding construction. Based on the initial three-dimensional point cloud data and initial image data, establish a baseline model containing cross-sectional geometric parameters and existing defect information. In this embodiment, a mobile inspection platform equipped with a laser scanner and camera is used to collect continuous cross-sectional point cloud and panoramic image data of the inner surface of the shield tunnel segments along the tunnel's travel direction. For long-distance tunnels, the mobile inspection platform can travel along a pre-set track or road surface within the tunnel to achieve full coverage.

[0021] This baseline model serves not only as a boundary constraint for subsequent construction but also as the sole basis for distinguishing between "inherent segment errors" and "lining construction errors." Specifically, the cross-sectional geometric parameters include cross-sectional ellipticity, circumferential joint misalignment, and cross-sectional center offset; the existing defect information includes the distribution and size of surface cracks, damage, and leakage marks on the segments. After subsequent grouting, by comparing with this baseline model, geometric deviations and defects already present during the shield tunnel segment assembly stage can be identified, avoiding misjudgment as new defects caused by spiral winding lining construction or grouting processes.

[0022] Step 102: Obtain construction image data of the winding and forming area during the spiral winding construction process; perform cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping. During the spiral winding process, a machine vision system is deployed on a mobile carrier, using an industrial camera on the mobile inspection platform to continuously monitor the winding area via video. Unlike traditional pipeline CCTV inspection, this invention shifts from "defect identification" to "process closed-loop": it transforms invisible process parameters (pitch, seam locking) into visible and controllable quality data, not only identifying problems but also verifying the stability of the process.

[0023] In some optional embodiments, the specific steps for performing cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping are as follows: Each frame of the construction image data undergoes a cylindrical unfolding transformation, mapping the helical winding lines on the tunnel's inner surface to oblique straight lines on a two-dimensional plane. Since the tunnel's inner wall is cylindrical, the helical winding lines appear as curves in the images, making direct analysis difficult. The cylindrical unfolding transformation maps the cylindrical coordinate system to a Cartesian coordinate system, converting the spatial helix into a planar oblique line, providing a geometric basis for subsequent pitch and trajectory analysis.

[0024] The system monitors the slope change of the straight line relative to a preset nominal slope. When the slope change exceeds a preset threshold, it is determined to be an abnormal winding pitch. There is a definite conversion relationship between the slope of the helix and the winding pitch: on the unfolded diagram, the lateral distance is the tunnel circumference (a fixed value, representing one revolution of the machine), the longitudinal distance is the distance the machine travels in one revolution (i.e., the pitch), and the slope equals the pitch divided by the circumference. Therefore, as long as the slope remains constant, it indicates that the pitch is uniform and stable; once the slope changes significantly, it is determined to be an abnormal pitch.

[0025] The machine acquires the positional offset of the straight line relative to the nominal path. When the positional offset exceeds the preset allowable deviation, it is determined to be a winding trajectory drift. The nominal path is the ideal winding route specified in the design drawings. By comparing the deviation between the actual route taken by the machine and the route planned in the drawings in real time, an alarm is immediately triggered once it is found that the machine is going astray or is unable to keep up with the previous turn, supporting on-site shutdown and correction.

[0026] Construction image data is input into a deep learning instance segmentation network to segment and extract the seam lines or overlap lines between adjacent strips, identifying the breakpoint location, break length, and opening width of the seam lines or overlap lines. The spiral-wound liner is connected by connecting adjacent rings of strip through seam or overlap methods. Seam refers to a locked connection where the male thread is inserted into the female thread, while overlap refers to a layered connection where one strip overlaps the other. Both connection methods appear as continuous, thin lines in the image. If a section of the line suddenly disappears, it indicates a break or disengagement; if the line suddenly thickens or forks, it indicates a loose connection or opening.

[0027] Based on the length of the break and the width of the opening, the severity of the discontinuity in the overlock or lap joint is determined, and a corresponding level of continuous alarm signal is output. This tiered alarm system distinguishes between minor defects and serious flaws, providing a basis for on-site handling.

[0028] Step 103: Obtain the as-built 3D point cloud data and as-built image data of the inner wall of the lining after grouting; identify apparent defects based on the as-built image data; analyze the geometric indices of the roundness, center offset, and circumferential joint misalignment of the lining based on the as-built 3D point cloud data, and identify areas with abnormal geometric deformation; and perform spatial registration and geometric difference analysis between the as-built 3D point cloud data and the baseline model, and identify areas with hollow areas or abnormal grouting by combining the process parameters during the grouting process. After grouting is completed, a mobile testing platform is used again to perform point cloud and image scanning of the entire line.

[0029] Identify apparent defects based on as-built image data, including visible defects such as cracks, damage, and leakage marks that appear on the lining surface after grouting.

[0030] In some optional embodiments, the geometric indices of the lining's roundness, center offset, and circumferential joint misalignment are analyzed based on the as-built 3D point cloud data to identify areas of abnormal geometric deformation. The specific steps are as follows: The as-built 3D point cloud data is continuously segmented along the tunnel's direction of travel. Circles or ellipses are fitted to the point clouds of each segment to obtain the roundness or ellipticity of the segment, as well as the offset of the segment center relative to the design axis. The misalignment is obtained from the point clouds at the circumferential joints of adjacent tunnel segments. Ellipticity reflects whether the tunnel cross-section has been compressed into an "egg shape," center offset reflects whether this section of the tunnel has "gone astray," and misalignment reflects the height difference at the joint between two adjacent slabs.

[0031] The roundness or ellipticity, center offset, and misalignment are generated as index curves along the mileage direction. The roundness or ellipticity, center offset, and misalignment of each mileage section are then expanded circumferentially, and color-mapped to represent the index values ​​at each circumferential location, generating a circumferential heatmap. This allows for the automatic spatial clustering and screening of weak deformation sections in the tunnel structure through continuous visualization along the mileage direction and circumferential color mapping of the discrete section detection data.

[0032] Based on the index curve and the circumferential heat map, areas with roundness or ellipticity, center offset and misalignment exceeding the preset threshold and exhibiting spatial clustering are identified as areas of geometric deformation anomalies.

[0033] In some optional embodiments, the steps of spatially registering and geometrically analyzing the as-built 3D point cloud data with the baseline model, and combining the process parameters during grouting to identify areas of voids or abnormal grouting behind the wall, are as follows: Spatially register the as-built 3D point cloud data with the initial 3D point cloud data in the baseline model to place them in the same coordinate system; obtain the normal distance field from the surface of the as-built point cloud to the surface of the initial point cloud after registration; and determine the geometric difference behind the wall based on the difference between the normal distance field and the sum of the inner lining wall thickness and the grouting layer design thickness.

[0034] This invention employs a "three-dimensional differential evaluation" method, namely Δd distance field analysis, which can accurately quantify the annular thickness and shape deviation, providing objective "forensic" evidence for grouting quality and structural safety. Through this differential analysis, the problem of "segment-lining" coupled deformation can be effectively solved: in the sensitive condition of subway tunneling, it can prove whether the deformation of the lining is caused by grouting compression or by inherent initial geometric deviations in the segments themselves.

[0035] Areas where the geometric difference behind the tunnel wall exceeds a preset allowable range are designated as candidate areas for anomalies. Grouting process curves for the corresponding mileage section are retrieved, including changes in grouting pressure, flow rate, and cumulative grouting volume over time. While geometric difference alone can identify the location of the anomaly, it cannot determine its cause. Cross-validation using the grouting process curves can differentiate between grouting anomalies and defects in the tunnel lining segments themselves, preventing misjudgments that could lead to incorrect construction.

[0036] When the geometric difference behind the wall in the candidate area of ​​an anomaly is less than the design thickness, and the pressure remains consistently low and the cumulative grouting volume exceeds the theoretical grouting volume during grouting, the grout filling in this area is determined to be incomplete, with voids present, and it is identified as a void area behind the wall. When the geometric difference behind the wall in the candidate area of ​​an anomaly is greater than the design thickness, and abnormal pressure fluctuations or peak values ​​exceed the preset value during grouting, the grouting pressure is determined to be too high, causing deformation of the lining, and it is identified as a grouting anomaly area.

[0037] Through the above steps, non-destructive and precise screening of hidden voids and abnormal grouting areas is achieved, and the evaluation of grouting quality is upgraded from subjective experience judgment to data-driven objective judgment.

[0038] Step 104: The construction defects detected in the construction image data are re-tested in the as-built 3D point cloud data and as-built image data. Based on the re-test results, the closed construction defects are distinguished from the unclosed construction defects. In some optional embodiments, the construction defects detected during the construction process are re-measured using as-built image data and as-built 3D point cloud data. The specific steps are as follows: The location information of the two-dimensional image of the construction defect is mapped to the as-built three-dimensional point cloud space through the extrinsic parameter calibration parameters between the camera and the laser scanner, thus determining the spatial location of the defect. The extrinsic parameter calibration is used to determine the relative spatial relationship between the camera and the laser scanner, including translation and rotation parameters, which is the basis for the accurate mapping of image pixels to point cloud spatial coordinates.

[0039] In the as-built image data, check whether the area corresponding to the spatial location has the same defect features; in the as-built 3D point cloud data, check whether the geometry of the spatial location meets the design tolerance requirements.

[0040] In some optional embodiments, the specific steps for distinguishing between closed and unclosed construction defects based on the retest results are as follows: if the construction defect no longer exists in the as-built image data and the as-built 3D point cloud data, or the defect size has been reduced to a preset allowable range, it is determined to be a closed construction defect; otherwise, it is determined to be an unclosed construction defect.

[0041] Step 105: Unclosed construction defects, apparent defects, abnormal geometric deformation areas, and areas with hollow walls or abnormal grouting are uniformly mapped to a three-dimensional point cloud space constructed based on the as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth angle, and size information.

[0042] This step achieves the fusion and delivery of multi-source data. By using the extrinsic calibration parameters between the camera and the laser scanner, the visually recognized defect mask (including edge openings, derailment, damage, leakage, etc.) is projected onto the point cloud to obtain the three-dimensional coordinates (mileage, azimuth), size, and geometric background information of the defects. This process overcomes the challenge of image and point cloud registration in narrow spaces, achieving the unification of "macro-geometry (point cloud)" and "micro-texture (visual)".

[0043] In some optional embodiments, the unclosed construction defects, apparent defects identified after grouting, areas of abnormal geometric deformation, and areas of hollowness or abnormal grouting behind the wall are uniformly mapped to a three-dimensional point cloud space constructed based on the as-built three-dimensional point cloud data to generate a defect three-dimensional map containing defect type, mileage, azimuth, and size information. The specific steps are as follows: assign a unique defect instance identifier to each defect; record the defect type, mileage coordinates, circumferential azimuth, geometric size, and severity level of each defect; record the re-measurement closure status of the construction defects; and overlay and render the three-dimensional bounding box or mask of each defect onto the three-dimensional model constructed from the as-built three-dimensional point cloud data to generate an interactive defect three-dimensional map.

[0044] In some optional embodiments, the defect 3D map, along with the full-line point cloud model, geometric index curves along the mileage, defect list reports, and quality traceability records from process early warning, rework handling to re-inspection closure, are also included as digital delivery results.

[0045] Through the aforementioned digital delivery, the final product delivered to the owner is not merely a paper report, but a digital twin model containing geometric information and health status of every meter of the tunnel. The resulting point cloud model, defect list, and 3D map can be directly used in the operation and maintenance phase, realizing the transformation of inspection data into digital twin assets and significantly enhancing the technological content of project delivery. In the sensitive situation of passing close to an existing subway line, data acquisition and closed-loop control throughout the entire process ensure that construction disturbances are within a controllable range, providing data protection for safe passage. Simultaneously, objective data demonstrates the fit between the tunnel segments and the lining, avoiding quality disputes caused by subjective judgment.

[0046] In this application, construction defects focus on compliance with process parameters during construction; apparent defects focus on the integrity of the surface after curing; geometric deformation anomalies focus on the macroscopic dimensional accuracy after pipe formation; and wall voids or grouting anomalies focus on the density of the filling layer between the lining and the original pipe. Although some construction defects may manifest as apparent or geometric anomalies after completion, this method, through phased and multimodal detection, can trace the root cause of the defects (whether it is due to improper construction operations or later stress deformation), thereby achieving more precise quality control.

[0047] A schematic diagram of the spiral winding lining repair structure for shield tunnels is shown below. Figure 2 As shown in the figure, the core structure of the spiral winding lining repair of the shield tunnel is presented. The outer shield segments are the original tunnel structure and are assembled and connected through circumferential and longitudinal bolt holes. The inner spiral winding pipe is the new lining pipe formed by the repair construction. The two constitute a double-layer tunnel structure of "original pipe-new lining", which provides a clear structural reference for the whole process quality inspection plan and clarifies the spatial relationship between the inspection object and each key component.

[0048] A schematic diagram of the spiral winding lining construction process for shield tunnels is shown below. Figure 3 As shown in the figure, the construction scene of the spiral winding lining is presented. The figure clearly marks the circumferential and longitudinal joints of the shield tunnel segments, the winding machine used for construction, and the interlayer interface between the formed lining and the original tunnel segments. It fully restores the construction scene during the process monitoring stage, providing an intuitive structural reference for real-time monitoring of the pitch, trajectory, and edge defects of the spiral winding, and also clarifies the key focus objects for the interlayer interface quality inspection in the subsequent completion stage.

[0049] A schematic diagram illustrating key indicators and implementation methods for quality inspection of spiral wound inner lining in shield tunnels is shown below. Figure 4 As shown in the figure, the core content of the whole process quality inspection is presented. The 3D point cloud inspection vehicle in the figure is the core tool for data acquisition, corresponding to the point cloud acquisition process in the baseline establishment and as-built analysis stages. The marked indicators such as segment roundness, segment misalignment, and joint width correspond to the baseline detection of the original error of the shield tunnel segment. Center offset, winding pitch, and back wall void correspond to the key quality control points in the spiral winding lining construction and completion stages. It fully demonstrates the core idea of ​​distinguishing between original errors and construction defects and realizing the quantitative detection of multi-dimensional quality indicators, providing clear detection indicators and implementation path references for the whole process closed-loop quality control.

[0050] A schematic diagram illustrating the principle of spiral winding construction and cylindrical unfolding transformation is shown below. Figure 5 As shown in the figure, the core principle of construction process monitoring is presented. Through cylindrical unfolding transformation, the spiral winding trajectory on the cylindrical surface of the tunnel is transformed into a straight line on a two-dimensional plane. The pitch and trajectory deviation, which were originally difficult to quantify directly, are transformed into the slope k and pitch d that can be calculated in the planar image. This provides a simplified and standardized analytical basis for real-time monitoring of construction defects (such as pitch abnormalities and trajectory drift) using machine vision. It solves the problem of direct detection of spiral trajectories in confined spaces and is a key technical support for realizing online quality control of the construction process.

[0051] A schematic diagram of a 3D point cloud acquisition device for quality detection of spiral wound inner lining in shield tunnels is shown below. Figure 6 As shown in the figure, the core data acquisition hardware used in the baseline establishment and as-built analysis phases is presented. The device consists of a point cloud acquisition sensor, a scanner bracket, a seismic isolation base, an inspection vehicle frame, wheels, and a data processing unit. It can move and operate inside the tunnel, and acquire high-precision three-dimensional point cloud data of the shield tunnel segments and the spiral wound lining. This provides key data support for building the segment baseline model, analyzing the geometric deformation of the lining, and screening for void defects behind the wall. It is the hardware foundation for realizing closed-loop quality inspection of the entire process from baseline to process to completion.

[0052] Another embodiment of the present invention relates to a quality inspection system for spiral wound lined tubes. The implementation details of the quality inspection system for spiral wound lined tubes of this embodiment are described in detail below. The following implementation details are provided for ease of understanding and are not necessary for implementing this solution.

[0053] Specifically, the baseline establishment module is used to acquire initial three-dimensional point cloud data and initial image data of the inner surface of the tunnel segment after the shield tunnel segment is assembled and before the spiral winding construction, and to establish a baseline model containing cross-sectional geometric parameters and existing defect information based on the initial three-dimensional point cloud data and initial image data. The process monitoring module is used to acquire construction image data of the winding area during the spiral winding construction process; to perform cylindrical unfolding and image analysis on the construction image data, and to detect construction defects such as abnormal winding pitch, trajectory drift and discontinuous locking or overlapping. The as-built analysis module is used to acquire as-built 3D point cloud data and as-built image data of the inner wall of the lining after grouting is completed; identify apparent defects based on the as-built image data; analyze the geometric indices of the lining's roundness, center offset, and circumferential joint misalignment based on the as-built 3D point cloud data, and identify areas of abnormal geometric deformation; and perform spatial registration and geometric difference analysis between the as-built 3D point cloud data and the baseline model, combined with the process parameters during the grouting process, to identify areas of voids or abnormal grouting behind the wall. The retesting and verification module is used to retest construction defects detected during the construction process on as-built image data and as-built 3D point cloud data, and to distinguish between closed and unclosed construction defects based on the retesting results. The defect map generation module is used to uniformly map unclosed construction defects, apparent defects identified after grouting, abnormal geometric deformation areas, and areas with voids or abnormal grouting behind the wall to a three-dimensional point cloud space constructed based on as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth angle, and size information.

[0054] The detection method in this embodiment has the following beneficial effects: First, closed-loop control throughout the entire process: For the first time, the "shield tunneling - segment tunneling - winding - grouting" process is incorporated into a unified quality evidence chain, achieving data connectivity from process early warning to final delivery. Second, a quantitative indicator system: Traditional empirical inspections are transformed into calculable process parameters (pitch, edge continuity) and geometric indicators (roundness, misalignment, bulge), supporting statistical process control and hierarchical alarms. Third, three-dimensional traceability of defects: Overcoming the limitation of two-dimensional images not being able to accurately locate defects, it achieves millimeter-level three-dimensional coordinate mapping and spatial clustering analysis of defects, significantly improving review efficiency and evidence reliability. Fourth, adaptability to complex working conditions: Through adaptive supplementary lighting, polarization suppression of reflection, mileage constraint registration, and other technologies, it adapts to harsh environments such as long distances, confined spaces, and wet surface reflections.

[0055] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the protection scope of this invention.

[0056] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0057] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes 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.

[0058] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A method for quality inspection of spirally wound inner lining tubes, characterized in that, The method includes: After the shield tunnel segments are assembled and before the spiral winding construction, the initial three-dimensional point cloud data and initial image data of the inner surface of the segments are obtained. Based on the initial three-dimensional point cloud data and initial image data, a baseline model containing cross-sectional geometric parameters and existing defect information is established. Acquire construction image data of the winding area during the spiral winding process; perform cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping. Acquire as-built 3D point cloud data and as-built image data of the inner wall of the lining after grouting; identify apparent defects based on the as-built image data; analyze the geometric indices of roundness, center offset and circumferential joint misalignment of the lining based on the as-built 3D point cloud data, and identify areas of abnormal geometric deformation; and perform spatial registration and geometric difference analysis between the as-built 3D point cloud data and the baseline model, combined with the process parameters during the grouting process, to identify areas of voids or abnormal grouting behind the wall. Construction defects detected in construction image data are re-measured in as-built 3D point cloud data and as-built image data. Based on the re-measurement results, closed construction defects are distinguished from unclosed construction defects. Unclosed construction defects, apparent defects, areas with abnormal geometric deformation, and areas with hollow areas or abnormal grouting behind the wall are uniformly mapped to a three-dimensional point cloud space constructed based on the as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth angle, and size information.

2. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The specific steps for performing cylindrical unfolding and image analysis on construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping are as follows: A cylindrical unfolding transformation is performed on each frame of the construction image data to map the spiral winding lines on the inner surface of the tunnel into oblique straight lines on a two-dimensional plane. Monitor the slope change of the straight line relative to the preset nominal slope. When the slope change exceeds the preset threshold, it is determined to be an abnormal winding pitch. Obtain the position offset of the oblique straight line relative to the nominal path. When the position offset exceeds the preset allowable deviation, it is determined to be a winding trajectory drift. The construction image data is input into a deep learning instance segmentation network to segment and extract the seam lines or lap lines between adjacent strips, and to identify the break point, break length and opening width of the seam lines or lap lines. Based on the length of the break and the width of the opening, the severity of the discontinuity in the edge locking or overlapping is determined, and a corresponding level of continuity alarm signal is output.

3. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The geometric indices of the lining, including roundness, center offset, and circumferential joint misalignment, are analyzed based on the as-built 3D point cloud data to identify areas of abnormal geometric deformation. The specific steps are as follows: Continuous cross-sections are extracted from the completed 3D point cloud data along the tunnel's direction of travel. Perform circular or elliptical fitting on the point cloud of each cross section to obtain the roundness or ellipticity of the cross section, as well as the offset of the cross section center relative to the design axis. The misalignment of the point cloud at the circumferential joint of adjacent segments is obtained; Generate index curves along the mileage direction for roundness or ellipticity, center offset, and misalignment. The roundness or ellipticity, center offset and misalignment of each mileage section are expanded according to the circumferential orientation, and the index values ​​of each circumferential position are mapped by color to generate a circumferential heat map. Based on the index curve and the circumferential heat map, areas with roundness or ellipticity, center offset and misalignment exceeding the preset threshold and exhibiting spatial clustering are identified as areas of geometric deformation anomalies.

4. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The process involves spatial registration and geometric difference analysis of the as-built 3D point cloud data and the baseline model, combined with process parameters during grouting, to identify areas of voids or abnormal grouting behind the wall. The specific steps are as follows: Spatial registration is performed between the completed 3D point cloud data and the initial 3D point cloud data in the baseline model, so that the two are in the same coordinate system. Obtain the normal distance field from the registered completed point cloud surface to the initial point cloud surface. Based on the difference between the normal distance field and the sum of the inner lining wall thickness and the grouting layer design thickness, determine the geometric difference behind the wall. Areas where the geometric difference behind the wall exceeds the preset allowable range are designated as candidate areas for anomalies behind the wall. Retrieve the grouting process curve for the corresponding mileage segment. The grouting process curve includes the changes in grouting pressure, flow rate, and cumulative grouting volume over time. When the geometric difference behind the wall in the candidate area of ​​abnormality behind the wall is less than the design thickness, and the pressure is consistently low and the cumulative grouting volume exceeds the theoretical grouting volume during the grouting process, it is determined to be a hollow area behind the wall. When the geometric difference behind the wall of the candidate area of ​​abnormality is greater than the design thickness, and abnormal pressure fluctuations or peak values ​​exceed the preset value during grouting, it is determined to be an abnormal grouting area.

5. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The specific steps for re-measuring construction defects detected in construction image data in as-built 3D point cloud data and as-built image data are as follows: The location information of the two-dimensional image of the construction defect is mapped to the as-built three-dimensional point cloud space through the extrinsic parameter calibration parameters between the camera and the laser scanner to determine the spatial location of the defect. In the as-built image data, check whether the area corresponding to the spatial location has the same defect features; In the as-built 3D point cloud data, check whether the geometry of the spatial location meets the design tolerance requirements.

6. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The specific steps for distinguishing between closed and unclosed construction defects based on the retest results are as follows: If the construction defect no longer exists in the as-built image data and the as-built 3D point cloud data, or the defect size has been reduced to the preset allowable range, it is determined to be a closed construction defect. Otherwise, it is judged as an unclosed construction defect.

7. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The process involves mapping unclosed construction defects, apparent defects identified after grouting, areas of abnormal geometric deformation, and areas of hollowness or abnormal grouting behind the wall to a unified 3D point cloud space constructed based on as-built 3D point cloud data. This generates a 3D defect map containing defect type, mileage, azimuth, and size information. The specific steps are as follows: Assign a unique defect instance identifier to each defect; Record the defect type, mileage coordinates, circumferential azimuth angle, geometric dimensions, and severity level for each defect; Record the closure status of the re-measurement of construction defects; The 3D bounding boxes or masks of each defect are overlaid and rendered on the 3D model constructed from the as-built 3D point cloud data to generate an interactive 3D defect map.

8. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, The cross-sectional geometric parameters include cross-sectional ellipticity, circumferential joint misalignment, and cross-sectional center offset; the existing defect information includes the distribution and size of cracks, damage, and leakage marks on the segment surface.

9. The quality inspection method for spirally wound inner lining tubes as described in claim 1, characterized in that, It also includes the output of 3D defect maps, point cloud models of the entire line, geometric index curves along the mileage, defect list reports, and quality traceability records from process early warning, rework handling to re-inspection closure, all as digital delivery results.

10. A quality inspection system for spirally wound inner lining tubes, characterized in that, The system includes: The baseline establishment module is used to acquire the initial three-dimensional point cloud data and initial image data of the inner surface of the tunnel segment after the shield segment is assembled and before the spiral winding construction. Based on the initial three-dimensional point cloud data and initial image data, a baseline model containing cross-sectional geometric parameters and existing defect information is established. The process monitoring module is used to acquire construction image data of the winding and forming area during the spiral winding construction process; it performs cylindrical unfolding and image analysis on the construction image data to detect construction defects such as abnormal winding pitch, trajectory drift, and discontinuous locking or overlapping. The as-built analysis module is used to acquire as-built 3D point cloud data and as-built image data of the inner wall of the lining after grouting; identify apparent defects based on the as-built image data; analyze the geometric indices of the lining's roundness, center offset, and circumferential joint misalignment based on the as-built 3D point cloud data, and identify areas of abnormal geometric deformation; and perform spatial registration and geometric difference analysis between the as-built 3D point cloud data and the baseline model, combined with the process parameters during the grouting process, to identify areas of voids or abnormal grouting behind the wall. The retesting and verification module is used to retest the construction defects detected in the construction image data in the as-built 3D point cloud data and as-built image data, and to distinguish between closed and unclosed construction defects based on the retesting results. The defect map generation module is used to uniformly map unclosed construction defects, apparent defects, abnormal geometric deformation areas, and areas with hollow walls or abnormal grouting to a three-dimensional point cloud space constructed based on as-built three-dimensional point cloud data, generating a three-dimensional defect map containing defect type, mileage, azimuth angle, and size information.

Images