Measuring device and method for performance evaluation of tunnel lining crack width measuring system

Abstract

The application discloses a kind of tunnel lining crack width measurement system performance evaluation measurement device and method, belong to measurement technical field.Measurement device includes test plate, multiple simulation cracks are formed on test plate, multiple simulation cracks form with first direction M angle, the angle of M angle is different, with first angle set, second angle set and third angle set, third angle set contains element greater than or equal to 2;First angle set value range is 0 and / or 180, second angle set value range is 90, third angle set value range is: first angle range and / or second angle range, first angle range is 0~90, second angle range is 90~180.The scheme can truly reflect the complex form of crack multi-angle, multidirectional expansion in actual tunnel.For tunnel lining crack detection equipment, especially the scientific, objective, standardized evaluation of crack width measurement precision provides solid theoretical basis and practical experimental tool.

Description

Technical Field

[0001] This invention belongs to the field of measurement technology, and specifically relates to a measuring device and method for evaluating the performance of a tunnel lining crack width measuring system. Background Technology

[0002] Tunnel lining cracks are a critical defect affecting structural safety and durability, making accurate and efficient detection crucial. The morphology of the cracks, especially their width, directly relates to the structural safety assessment and maintenance decisions. Therefore, developing accurate and efficient tunnel lining crack detection technology is urgently needed for timely and effective preventative maintenance.

[0003] To address this need, various tunnel lining crack detection technologies have emerged. Currently, tunnel inspection has shifted from traditional methods such as visual inspection and manual measurement to automated methods, such as ultrasonic flaw detectors and ground-penetrating radar. It's important to note that while these instruments can effectively assess the internal condition of the lining, their contact-based or low-speed-movement measurement methods are inefficient and can easily interfere with normal tunnel traffic. In recent years, vision-based inspection methods have gained widespread application due to their high efficiency and non-contact advantages. For example, using machine vision combined with a CCD camera to acquire images of the lining surface has achieved crack identification with an error margin of less than 10%. To further improve detection accuracy and obtain richer structural information, some researchers are also attempting to combine laser point cloud scanning to obtain high-precision three-dimensional coordinates and texture information of the tunnel lining, constructing digital models to identify structural defects. Meanwhile, intelligent neural network technologies such as deep learning have shown great potential in automatic crack identification and are gradually becoming a research hotspot.

[0004] However, no matter how advanced the measuring instruments or technology themselves are, a scientific and objective evaluation of their measurement performance remains the cornerstone of ensuring the reliability of the test results. To accurately evaluate the measurement performance of a tunnel lining crack measurement system, simply identifying cracks is insufficient. Many domestic and international standards and specifications have stipulated and required the measurement of crack width in tunnel linings. For example, the "Specifications for Design of Highway Tunnels" JTG 3370.1-2018 requires that the calculated crack width of reinforced concrete components be ≤0.20 mm (such as arches and sidewalls); the "Specifications for Durability Design of Concrete Structures in Highway Engineering" JTG / T 3310 requires that the crack width be ≤0.15 mm in corrosive environments (such as marine areas and chemical plant areas); and the "Specifications for Durability Design of Concrete Structures" GB / T 51355 requires that the bending crack width of self-waterproofing concrete components be ≤0.20 mm, etc. Different crack widths correspond to different repair methods. For example, the International Tunneling Association (ITA) recommends Grade A (≥ 0.50 mm): immediate structural repair; Grade B (0.30–0.50 mm): grouting reinforcement; and Grade C (0.10–0.30 mm): seal monitoring.

[0005] However, the actual shape of tunnel lining cracks is irregular, and the direction of the skeleton curve is untraceable, making their measurement and calibration quite difficult. Figure 1 Currently, the definition of crack width is unclear, including definitions and calculation methods for global maximum crack width, local maximum crack width, and average width. The completely irregular crack orientation and different width measurement and calculation methods result in a chaotic and inconsistent state of crack width measurement. This also makes it impossible to accurately quantify and evaluate the true measurement performance of tunnel lining crack width measurement systems.

[0006] Therefore, there is an urgent need to establish a performance evaluation method for a tunnel lining crack width measurement system. Summary of the Invention

[0007] To address the problems existing in the prior art, the present invention provides a measuring device for evaluating the performance of a tunnel lining crack width measuring system, comprising:

[0008] A test board has multiple simulated cracks formed on it. These simulated cracks form M angles with a first direction, and the M angles are all different. The M angles have three sets: a first set of angles, a second set of angles, and a third set of angles. The third set of angles contains at least two elements. The first set of angles ranges from 0° to 180°, the second set of angles ranges from 90°, and the third set of angles ranges from the first set of angles to the second set of angles. The first set of angles ranges from 0° to 90°, and the second set of angles ranges from 90° to 180°.

[0009] Optionally, the simulated cracks have a straight direction, the number of simulated cracks is M, and the width of the simulated cracks is equal.

[0010] Optionally, the M simulated cracks are distributed in a fan shape, a 360° full circle, an irregular random distribution, or an asymmetric array distribution generated based on a probability distribution.

[0011] Optionally, the simulated crack width set has a first width subset, a second width subset, and a third width subset. The width range in the first width subset is: a first set width threshold H1 to a second identification width threshold H2. The width range in the second width subset is: a second identification width threshold H2 to a third identification width threshold H3. There are multiple widths in the second width subset. The width range in the third width subset is: a third identification width threshold H3 to a fourth set width threshold H4. H2 is less than H3 and all of them are widths of simulated cracks that can be identified by the tunnel lining crack width measurement system.

[0012] The width of the M simulated cracks covers all widths in the set of simulated crack widths.

[0013] Optionally, M is greater than the number of widths in the set of simulated crack widths, and the set of simulated crack widths corresponds to the widths of the simulated cracks in a random distribution.

[0014] Optionally, H2 is 0.1 mm, H3 is 1 mm, M=19, and the interval angle between adjacent simulated cracks is the same.

[0015] On the other hand, a method for evaluating the performance of a tunnel lining crack width measurement system using the aforementioned measuring device is provided, comprising:

[0016] Set up a test environment and fix multiple test boards in the test environment. The width distribution of multiple simulated cracks in each test board is different.

[0017] A target tunnel lining surface image is dynamically acquired by the tunnel lining crack width measurement system in the test environment, and the target tunnel lining surface image contains the test plate;

[0018] The surface image of the target tunnel lining is processed to obtain measurement results, which include the width of each simulated crack in each test plate;

[0019] Based on the measurement results, the performance of the tunnel lining crack width measurement system is evaluated from two dimensions: simulated crack recognition rate and relative error of crack width.

[0020] Optionally, the test environment is a simulated tunnel;

[0021] The acquisition of the target tunnel lining surface image dynamically collected by the tunnel lining crack width measurement system in the test environment includes:

[0022] The tunnel lining crack width measurement system to be detected is driven into the simulated tunnel at the detection speed;

[0023] Obtain the surface image of the target tunnel lining acquired by the tunnel lining crack width measurement system.

[0024] Optionally, the detection speed is 60–100 km / h.

[0025] The beneficial effects of the above technical solutions in the embodiments of the present invention are as follows:

[0026] This scheme accurately reflects the complex morphology of cracks propagating from multiple angles and directions in actual tunnels. It provides a solid theoretical basis and practical experimental tools for the scientific, objective, and standardized evaluation of tunnel lining crack detection equipment, especially the accuracy of crack width measurement. This not only facilitates horizontal comparison and optimization guidance of equipment performance but will also promote the development of tunnel structural health monitoring technology towards greater precision, efficiency, and standardization. Attached Figure Description

[0027] Figure 1 A schematic diagram of an actual tunnel lining crack provided for an embodiment of the present invention;

[0028] Figure 2 This is a schematic diagram of the structure of a measuring device for evaluating the performance of a tunnel lining crack width measuring system, provided in an embodiment of the present invention.

[0029] Figure 3 This is a schematic diagram of the physical structure of a measuring device for evaluating the performance of a tunnel lining crack width measurement system, provided in an embodiment of the present invention.

[0030] Figure 4 This is a schematic diagram of the tunnel lining test environment used in a performance evaluation method for a tunnel lining crack width measurement system provided in an embodiment of the present invention.

[0031] Figure 5 A grayscale image schematic diagram of a test board provided in an embodiment of the present invention;

[0032] Figure 6 A flowchart illustrating a performance evaluation method for a tunnel lining crack width measurement system provided in an embodiment of the present invention;

[0033] Figure 7 A schematic diagram illustrating the correlation between the crack width measurement value wt and the standard value wD provided in an embodiment of the present invention;

[0034] Figure 8 This is a schematic diagram illustrating the repeatability comparison of measurement values ​​provided in an embodiment of the present invention;

[0035] Figure 9 This is a schematic diagram illustrating the absolute error of crack width under different crack angles θ, provided as an embodiment of the present invention.

[0036] Figure 10 A schematic diagram illustrating the relative error of crack width provided in an embodiment of the present invention;

[0037] Figure 11 This is a schematic diagram showing the comparison of measurement system scores under different crack widths, provided by an embodiment of the present invention.

[0038] Figure 12This is a schematic diagram showing the comparison of measurement system scores under different crack angles θ, provided as an embodiment of the present invention.

[0039] Explanation of the numbering in the diagram:

[0040] 1 test plate, 10 mounting holes, 11 simulated cracks. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0042] See Figure 2-3 This invention provides a measuring device for evaluating the performance of a tunnel lining crack width measurement system. The measuring device includes a test plate 1. Multiple simulated cracks 11 are formed on the test plate 1 to simulate tunnel lining cracks. Each simulated crack 11 forms M angles with a first direction, and these M angles are all different. The first direction can be the direction in which the tunnel extends, such as a horizontal direction. Figure 1 The center represents the left-right direction. The test plate can be manufactured by machining, and its structure is highly reproducible. In other embodiments, it can also be made of polymer composite materials, ceramic or resin-based composite materials through additive manufacturing technologies such as injection molding and 3D printing to reduce processing costs and improve deformation stability. This embodiment does not limit this.

[0043] The number of simulated cracks 11 and the number of included angles can be the same. In this case, each simulated crack 11 forms an included angle with the first direction. This can be because the included angles between the simulated cracks 11 and the first direction are all the same, or one marker position is set for each simulated crack 11, and this marker position forms an included angle with the first direction. This marker position can be used to identify the crack width. Alternatively, the number of simulated cracks 11 and the number of included angles can be different. In this case, any one of the simulated cracks 11 forms at least two included angles with the first direction. For example, multiple marker positions can be set, and each marker position forms an included angle with the first direction. This embodiment does not limit this. The M included angles have: a first set of angles, a second set of angles, and a third set of angles, where the third set contains at least two elements.

[0044] The first angle set can take values ​​in the range of 0° and / or 180°. In other words, the first angle set can be 0°, meaning the first angle set contains one element of 0°, indicating that among the multiple simulated cracks 11, there is a simulated crack (or the marked position of the simulated crack) with an angle of 0° with the first direction. In this case, the simulated crack (or the marked position of the simulated crack) can simulate a circumferential crack. The first angle set can also be 180°, meaning the first angle set contains one element of 180°, indicating that among the multiple simulated cracks 11, there is a simulated crack (or the marked position of the simulated crack) with an angle of 180° with the first direction. In this case, the simulated crack (or the marked position of the simulated crack) can simulate a circumferential crack. The first angle set can also be 0° and 180°, meaning the first angle set contains two elements of 0° and 180° respectively, indicating that among the multiple simulated cracks 11, there is a simulated crack 11 with an angle of 0° with the first direction, and at the same time, among the multiple simulated cracks 11, there is also a simulated crack 11 with an angle of 180° with the first direction. The second angle set has a value range of 90°, meaning that the second angle set contains one element that is 90°. This indicates that among the multiple simulated cracks 11, there is a simulated crack 11 with an angle of 90° with the first direction. At this time, the simulated crack (or the marked position of the simulated crack) can simulate a longitudinal crack.

[0045] The range of the third angle set is: the range of the first angle and / or the range of the second angle, where the first angle range is 0°~90° and the second angle range is 90°~180°. In other words, the range of the third angle set can be the first angle range, meaning that each element in the third angle set belongs to the first angle range (0°~90°), indicating that among the multiple simulated cracks 11, there is a simulated crack (or the marked position of the simulated crack) whose angle with the first direction is between 0° and 90°; or the range of the third angle set can be the second angle range, meaning that each element in the third angle set belongs to the second angle range (90°~180°), indicating that among the multiple simulated cracks 11, there is a simulated crack (or the marked position of the simulated crack) whose angle with the first direction is between 0° and 90°. The included angle ranges from 90° to 180°; the range of the third angle set can also be the range of the first angle and the range of the second angle, that is, some elements in the third angle set belong to the first angle range and some elements belong to the second angle range, indicating that the included angle between the simulated cracks and the first direction in the multiple simulated cracks 11 ranges from 0° to 90°. At the same time, there are also simulated cracks in the multiple simulated cracks that have included angles between the simulated cracks and the first direction ranging from 90° to 180°. In this case, 0° to 90° and 90° to 180° correspond to oblique cracks with different tilt angles simulated.

[0046] This design allows all simulated cracks 11 on test plate 1 to simulate lining cracks with more expansion angles, achieving multi-angle simulation. This facilitates a reasonable evaluation of the measurement performance of the tunnel lining crack width measurement system and has a better scientific basis.

[0047] Since the width of actual tunnel lining cracks is not constant, and even the width of the same crack can vary at different locations, the ideal method is to accurately measure a large number of actual tunnel lining cracks and use them as standard specimens. However, this method is extremely costly. Considering that the skeleton lines of actual lining cracks are arbitrarily oriented, this embodiment simplifies irregular lining cracks into straight line segments with different slopes, and further simplifies them to a uniform width. This width simplification method also ensures that the final calculation results based on the calculation method for most crack widths are consistent.

[0048] The simulated crack 11 is machined from test plate 1. The simulated crack has a certain depth and high accuracy in size and shape. The simulated crack 11 is simplified to a straight line for ease of machining and measurement. For irregularly shaped simulated cracks 11, corners may exist during machining. A crucial step in identifying and calculating the crack width during measurement is extracting the crack's boundary and normal. For irregularly shaped simulated cracks, errors in boundary and normal extraction can lead to deviations in crack width calculation. For straight-line simulated cracks, the boundaries on both sides of the finely machined crack are extremely clear, and the boundary and skeleton lines are easily extracted with high accuracy, allowing the actual crack width to be calculated using the normal. When the simulated crack 11 follows a straight line, for M included angles, M simulated cracks 11 are formed, each forming an angle with the first direction. The number of simulated cracks 11 corresponding to the third angle set is N, where 1 < N < M, meaning the third angle set contains at least two elements.

[0049] In other embodiments, the simulated crack 11 may also be a Bezier curve, a sinusoidal waveform, or an exponentially decaying continuously widening crack to more realistically simulate the asymmetry, bifurcation, or uneven width distribution along the length of actual tunnel lining cracks.

[0050] To simulate the multidirectional propagation characteristics of cracks in the tunnel lining and to simulate cracks with more propagation angles, the simulated cracks on test plate 1 were designed in a fan-shaped arrangement. In other words, multiple simulated cracks form the framework of the fan-shaped arrangement, distributed in a fan shape within a 180° range, as shown below. Figure 2-3As shown. In other embodiments, the arrangement of multiple simulated cracks can be a 360° full-circle distribution, an irregular random distribution, or an asymmetric array distribution generated based on probability distribution, to adapt to the performance evaluation requirements of different tunnel lining crack width measurement systems (such as omnidirectional scanning systems). These alternative layouts not only maintain the spatial coverage integrity of the simulated cracks, but also achieve precise positioning of crack location and width through image recognition and spatial registration technology, thereby ensuring the accuracy and comparability of the evaluation data.

[0051] The included angle between adjacent simulated cracks in multiple simulated cracks 11 can be the same, i.e., evenly distributed at equal intervals, such as 5°, 10°, 15°, or other equal intervals. When the included angle is 10°, the number of simulated cracks distributed in a fan shape within a 180° range is 19. When the angle is other, the number of simulated cracks set on the test plate 1 will be more or less than 19. Generally, the more simulated cracks, the better, indicating that the angle subdivision of the crack propagation direction is finer, which can more comprehensively evaluate the performance of the tunnel lining crack width measurement system. On the other hand, a larger number of cracks indicates that there will be more simulated cracks of the same width, which is more conducive to subsequent analysis of the tunnel lining crack width measurement system's measurement performance for a certain width using a repeatable approach.

[0052] A set of simulated crack widths is established, comprising a first subset, a second subset, and a third subset. The width range in the first subset is from a first set width threshold H1 to a second identifier width threshold H2. The width range in the second subset is from a second identifier width threshold H2 to a third identifier width threshold H3, and there are multiple widths in the second subset. The width range in the third subset is from a third identifier width threshold H3 to a fourth set width threshold H4. H2 is less than H3, and both are widths of identifiable simulated cracks identified by the tunnel lining crack width measurement system. H2 can be the minimum width of an identifiable simulated crack identified by the tunnel lining crack width measurement system. Each simulated crack has a certain width, and the widths of the M simulated cracks cover all widths in the set of simulated crack widths; that is, each width in the set of simulated crack widths can be identified by at least one simulated crack among the M simulated cracks.

[0053] In tunnel lining crack detection, the actual crack widths exhibit significant differences. Some cracks are extremely small, at the sub-millimeter level, while others are more pronounced, reaching the millimeter level. The goal of tunnel lining detection is to effectively detect all cracks, especially those with extremely small widths. Currently, the highest requirement for identifying and measuring tunnel lining crack width in the industry is the second identification width threshold. The specific measurement accuracy varies depending on the system; for example, some system manufacturers claim their systems can achieve 0.1mm, and some tunnel lining crack width measurement systems also use 0.1mm as a selling point. Therefore, when designing the test plate, it is necessary to ensure that there are simulated cracks with widths smaller than the second identification width threshold H2. Thus, the simulated cracks have widths belonging to the first width subset, and the element values ​​range from the first set width threshold H1 to the second identification width threshold H2. When H2 is 0.1mm, H1 can be 0.07mm, and correspondingly, the first width subset can contain one element, which can be 0.08mm. For simulated cracks with a width exceeding the third identifier width threshold H3, such as 1mm, there is no technical difficulty for current image recognition equipment. Moreover, in engineering, crack widths greater than 1mm are already quite significant, and the impact of cracks, such as crack expansion rate, is analyzed through methods like fixed-point monitoring, rather than rapid detection and identification methods. The detection and identification method used in the tunnel lining crack width measurement system described in this embodiment is a rapid detection and identification method, i.e., detection and identification while in motion. Therefore, this embodiment sets a third width subset, whose width range is from the third identifier width threshold H3 to the fourth set width threshold H4. When H3 is 1mm, H4 can be 1.2mm. Correspondingly, the third width subset can contain one element, which can be 1.15mm. It should be noted that setting a crack width larger is not very meaningful. Within the range greater than the third identifier width threshold H3, only the parameter 1.15mm is set.

[0054] A second width subset is set, with its width range being the second identifier width threshold H2 to the third identifier width threshold H3. The second width subset contains multiple widths. Theoretically, the crack width within this range can be arbitrarily set. In this scheme, it roughly covers the second identifier width threshold H2 to the third identifier width threshold H3 in increments of 0.1 mm. For example, when H2 is 0.1 mm and H3 is 1 mm, values ​​could be 0.25 mm, 0.35 mm, 0.45 mm, 0.55 mm, 0.65 mm, 0.75 mm, 0.85 mm, 0.95 mm, etc. To evaluate the performance of the tunnel lining crack measurement system, specifically, 12 width combinations of 0.08 mm, 0.12 mm, 0.18 mm, 0.25 mm, 0.35 mm, 0.45 mm, 0.55 mm, 0.65 mm, 0.75 mm, 0.85 mm, 0.95 mm, and 1.15 mm were selected to form a set of simulated crack widths, representing the differences in crack width. The first width subset was 0.08 mm, the second width subset was 0.12 mm, 0.18 mm, 0.25 mm, 0.35 mm, 0.45 mm, 0.55 mm, 0.65 mm, 0.75 mm, 0.85 mm, and 0.95 mm, and the third width subset was 1.15 mm.

[0055] To ensure the accuracy of the standard values ​​represented by the simulated cracks, high-precision machining and calibration techniques (such as crack width gauges) are used to measure the crack width. In other embodiments, this calibration method can be replaced by other techniques, such as using standard gauge blocks, laser scanning confocal microscopes, digital holographic interferometers, or micro-displacement sensor arrays based on microelectromechanical systems (MEMS) for three-dimensional topography scanning and width inversion. Alternatively, micro-gratings or conductive microelectrodes can be embedded inside the crack to achieve online measurement via optical or electrical signals. While these alternative calibration techniques differ in equipment cost and operational difficulty, they all provide stable and traceable width reference values.

[0056] To increase differentiation and unpredictability, the number M of simulated cracks is greater than the number of elements in the simulated crack width set. The simulated crack width set corresponds to the simulated crack widths in a random distribution. When the simulated crack width set contains the aforementioned 12 widths, these 12 widths are randomly distributed on the test plate. Typically, performance testing involves deploying multiple test plates, each with a different width distribution. This ensures that cracks with the same width value but distributed at different angles on different test plates can be equivalent to repeated reproduction and measurement of the same width value. To avoid the influence of human factors, the crack widths are randomly distributed at different angles or on different test plates. If a random distribution is not used, and multiple simulated cracks on the same test plate have the same width, such as 0.5mm, the tunnel lining crack width measurement system acquires the apparent images of tunnel lining cracks during high-speed driving. Due to vehicle vibration, road bumps, equipment accuracy issues, etc., image distortion, blurring, or other imaging problems are unavoidable. Assuming a simulated crack at a 0° angle can be identified and measured, other angles may be misidentified as 0.2mm or 0.6mm due to distortion. To ensure a good performance score for the measurement system, human factors might cause cracks at other angles to be output as 0.5mm, thus rendering the measurement system's evaluation score unscientific. The test plate should not be too thick, ideally less than 5mm.

[0057] Currently, the performance evaluation methods for tunnel lining crack measurement systems (especially crack width measurement accuracy) still have the following obvious shortcomings / deficiencies.

[0058] ①Most studies rely on actual, diverse cracks to verify equipment performance. This results in evaluations that are limited to specific actual crack samples, making them difficult to quantify, standardize, and reproduce, and are also inefficient.

[0059] ② Many measurement systems focus on “identifying” cracks rather than “precisely measuring” them, especially paying insufficient attention to the quantitative accuracy of crack width.

[0060] ③ Static or low-speed testing is used during calibration or excitation, which cannot reflect the true error at operating vehicle speeds (60–100 km / h).

[0061] ④ There are many different tunnel crack measurement systems, and there is no unified method for quantitative evaluation and assessment of their measurement performance. As a result, there is a lack of clear and recognized measurement standards for key indicators such as measurement accuracy, reliability, and stability of various instrument and equipment systems.

[0062] To objectively evaluate the overall performance of the measurement device for measuring the width of cracks in tunnel lining, see [reference needed]. Figure 4-12This embodiment provides a performance evaluation method for a tunnel lining crack width measurement system. When executed, it utilizes the measurement device for evaluating the performance of the tunnel lining crack width measurement system mentioned in the above embodiment. The method includes:

[0063] Step 101: Set up the test environment. In the test environment, fix multiple test boards and the width distribution of multiple simulated cracks in each test board is different.

[0064] Multiple test plates are fixedly arranged in the testing environment, with different width distributions of simulated cracks on each plate. During the experiment, multiple test plates were configured, and it should be noted that the crack widths on these plates are randomly distributed. In practical applications, there is no specific requirement for the exact number of test plates, but more is better, as this allows for a more accurate evaluation of the tunnel lining crack width measurement system's performance and avoids errors in evaluation scores introduced by a single random error.

[0065] Multiple test boards are placed in the testing environment to simulate the occurrence of cracks in an actual tunnel. The boards are designed for detachable connection for quick installation and removal. This can be achieved through magnetic connections (with mounting holes 10 on the back or edge of the test board to accommodate magnetic materials such as neodymium iron boron magnets), adhesive connections, screw connections (with pre-drilled threads on the edge of the test board), or snap-fit ​​connections (with slotted structures on the edge of the test board). All of these structures support efficient deployment at multiple locations and heights within the testing environment, meeting the requirements for rapid replacement under dynamic testing conditions.

[0066] The installation location of the test panels is designed with a specific purpose in mind. Vehicle-mounted tunnel lining crack detection systems typically employ multiple cameras to collaboratively photograph the test environment where the test panels are installed. Due to limitations in camera field of view, a complete test environment cross-section usually requires more than six cameras. Test panels placed at different heights will be photographed by different cameras. Therefore, test panels are positioned at different heights or locations. Figure 4 Four test boards were set up. One test board was set up at the first position, and three test boards were set up at the second position. One of the three test boards was the same height as the test board at the first position, and the other two test boards were set on both sides of the first test board. The first and second positions were set at intervals along the driving direction / tunnel extension direction.

[0067] Step 102: Obtain a target tunnel lining surface image dynamically acquired by the tunnel lining crack width measurement system in the test environment. The target tunnel lining surface image contains a test plate.

[0068] The testing environment can be a simulated tunnel. Multiple test plates are placed on the simulated tunnel lining surface along the driving direction. The vehicle-mounted tunnel lining crack width measurement system is driven into the simulated tunnel at the detection speed. During the journey, the system captures images of the entire simulated tunnel lining surface, thus obtaining an image of the target tunnel lining surface. The detection speed can be 60–100 km / h, allowing dynamic testing at actual vehicle speeds within the simulated tunnel to verify the stability of the measurement system under real operating conditions.

[0069] In other embodiments, the test environment can also be a uniform-speed moving platform, such as a tracked mobile chassis, set up in a closed test field. The uniform-speed moving platform carries a tunnel lining crack width measurement system to perform uniform or variable-speed tests, thereby realizing dynamic testing. Although the above alternative test methods differ slightly in terms of dynamic realism, they can still effectively reflect the performance of the system in evaluating its robustness under multiple working conditions and have equivalent verification value.

[0070] Step 103: Process the surface image of the target tunnel lining to obtain measurement results, which include the number of simulated cracks identified in each test plate and the width of the simulated crack.

[0071] Step 104: Evaluate the performance of the tunnel lining crack width measurement system based on the measurement results from two dimensions: simulated crack recognition rate and relative error of crack width.

[0072] In critical scenarios such as tunnel lining quality and safety inspection, the performance evaluation of a tunnel lining crack width measurement system must ensure both the completeness of detection and identification—that is, accurately identifying the existence of cracks—and the confidence level of the measurement values—that is, accurately measuring the crack width. Traditional single-dimensional evaluation methods have inherent limitations: over-focusing on high detection rates may lead to coarse measurements, while unilaterally emphasizing measurement accuracy may weaken risk coverage. Therefore, this embodiment constructs a progressive two-dimensional evaluation system, comprehensively and systematically characterizing the performance of the tunnel lining crack width measurement system from the underlying detection capability to the upper-level measurement accuracy.

[0073] The completeness of crack detection and identification in a tunnel lining crack width measurement system serves as the underlying support for system performance and directly determines the completeness of structural safety risk identification. This method uses the identification ratio (IR) as a key indicator to measure the completeness of the system's detection and identification. This ratio represents the ratio of correctly identified cracks to the total number of actual cracks, characterizing the system's ability to perceive crack morphology. A low IR indicates an increased rate of missed / false detections, potentially leading to the omission of critical structural risks. The calculation method for the identification ratio IR is shown in the formula. .

[0074]

[0075] In the formula, N0 represents the total number of actual cracks, N T This indicates the number of cracks that were correctly identified.

[0076] Based on the aforementioned detection integrity, relative error (RE) is used as the measurement accuracy index for the tunnel lining crack width measurement system. The relative error RE for the i-th crack is... i The calculation method is shown in the formula. This indicator is the measured value w. ti Compared with the standard value w Di The error between the two and the standard value w Di The absolute value of the ratio quantifies the reliability of the geometric parameters of the identified cracks. High measurement bias still carries the risk of misjudgment in the assessment of the bearing capacity of the tunnel lining structure.

[0077]

[0078]

[0079] To objectively evaluate the overall performance of the tunnel lining crack width measurement system, this method constructs a dual-index weighted evaluation model based on the aforementioned recognition rate (IR) and relative error (RE), as shown in the formula. This model establishes a quantitative evaluation system for crack measurement systems by integrating two key parameters: recognition rate (IR) and relative error (RE). The final output is a comprehensive score ranging from [0, 100], enabling a direct comparison of the measurement performance of tunnel lining crack width measurement systems.

[0080]

[0081] In the formula, parameters a and b represent the weights of the recognition rate IR and the relative measurement error RE, respectively, and a + b = 1. Without loss of generality, in some scenarios, both weight values ​​can be set to 0.5, indicating that crack recognition is as important as crack width calculation.

[0082] In other embodiments, a weighted nonlinear scoring function can be used to better align with human subjective judgment of the system's overall performance; or a dynamic weighting mechanism can be introduced, for example, automatically adjusting the relative weights of IR and RE based on the crack width range (e.g., <0.2 mm for safety warning, >0.5 mm for emergency repair), to achieve differentiated evaluation based on the specific problem. Such intelligent alternatives can also achieve the "standardized, quantifiable, and comparable" evaluation goals of this invention.

[0083] The following section uses an experimental design as an example to explain this method in detail.

[0084] To investigate and verify the correctness of the proposed quantitative method for measuring the crack width of tunnel lining, this method was tested in a simulated tunnel at a test site. The simulated tunnel's radius was designed to be approximately 6m, similar to the dimensions of an actual highway tunnel. Based on the testing requirements of the tunnel lining crack width measurement system, a total of 12 test plates were customized for the experiment (a photograph of one of the test plates can be found here). Figure 3 ), and are magnetically attached to different positions on the tunnel lining surface along the direction of travel (see Figure 4 ).

[0085] The experiment employed a domestically produced vehicle-mounted tunnel lining crack detection system, which consists of multiple industrial area array cameras, encoders, and a data processing unit. The main experimental procedures are as follows:

[0086] ① Twelve test plates were pasted at different heights and positions on the tunnel lining surface to simulate real tunnel cracks;

[0087] ② The tunnel lining crack width measurement system was driven into the simulated tunnel scene at a typical detection speed (close to the actual traffic flow speed) and continuously acquired images of the tunnel lining surface through multiple industrial area array cameras.

[0088] ③ The encoder records the spatial position corresponding to the acquisition time of each image in real time, realizing the precise correspondence between the image and the position of the tunnel structure, and ensuring the spatial traceability of the data;

[0089] ④ Based on the collected visual image data, the instrument under test manually / automatically completes image processing and analysis, identifies the location and direction of the crack, quantifies the width of each straight segment of the crack, and finally outputs the crack width measurement results.

[0090] Image acquisition results for some test boards can be found in [link to image acquisition results]. Figure 5 .

[0091] In this test experiment, these 12 test plates (or crack plates) provided a total of 12×19 standard crack widths w. DThe instrument simultaneously measured cracks of different widths on 12 test plates, obtaining a total of 12 × 19 measurement values. t The crack width measurement performance of the tested instrument is based on these data.

[0092] Results of conventional qualitative analysis methods:

[0093] Theoretically, if the performance of the tunnel lining crack width measurement system is ideal, then the width test value w t Should be consistent with the standard value w D They are completely identical, and their correlation also approaches 1. The correlation analysis measures the value w. t Compared with the standard crack width value w D ,See Figure 7 The results show that the measured value w t Fluctuations within a certain range indicate the relative stability of the measurement results. The mean w of the measured values... T With the standard width value w D The fitted Pearson correlation coefficient exceeded 0.99, indicating a high correlation between the two. Further linear regression analysis showed that the slope of the regression line was 0.906 (less than 1). This result suggests that when the crack width is small, the measured value w... t There is a problem of systemic over-largeness.

[0094] Based on the design and fabrication of the cracks, it is known that there are multiple cracks of the same width, randomly distributed across different cracked plates. Therefore, a repetitive analysis approach can be used. Figure 8 Repeatability measurement results of a tunnel lining crack width measurement system under different crack widths are presented. The results show that the coefficient of variation (Cv) gradually decreases with increasing crack width. When the crack width exceeds 0.2 mm, the Cv drops below 20%, and further to 5.13% when the crack width reaches 1.15 mm. However, this result does not reflect an improvement in measurement performance with increasing crack width. In fact, the Cv is not only related to the standard deviation of the measured values ​​but also affected by the crack width as the denominator. As the crack width increases, the denominator increases, and even if the standard deviation remains constant or changes only slightly, Cv will naturally decrease. Therefore, the decrease in Cv reflects more the change in crack width itself than an absolute improvement in measurement performance. Thus, using the Cv to measure the performance of a crack measurement system does not seem to be a good choice.

[0095] To thoroughly evaluate the measurement performance of the tested tunnel lining crack width measurement system, the measurement errors for each crack width were statistically analyzed. (See attached data.) Figure 9The horizontal axis of the figure represents the angle θ between the crack skeleton line and the horizontal line, in degrees; the vertical axis represents the absolute error values ​​corresponding to the 12 cracked plates, in mm. The results show that the measurement error distribution of crack width is similar under different angles θ, with both positive and negative values, and no significant trend changing with θ. This result indicates that the tested instrument has similar measurement performance for cracks at any angle.

[0096] The results from the perspectives of correlation, repeatability, and measurement error indicate that the performance of the tested instrument is relatively stable, but there is still a certain gap between its performance and expectations. However, the above results do not provide more quantitative information.

[0097] Quantitative analysis results:

[0098] First, looking at the recognition rate (IR), all 12×19 cracks were identified, i.e., IR=100%. This value indicates that the tested instrument performed perfectly, but this is clearly not what we expected, and it is even inconsistent with the results of the aforementioned conventional analysis.

[0099] Secondly, the measured value w for each crack t Compared with the standard value w D Relative error RE i See Figure 10 Because many crack widths share the same standard value, the horizontal axis in the graph exhibits significant repetition. For each standard crack width, RE i The distribution of values ​​is relatively dispersed, and the maximum value tends to decrease as the crack width increases, even exceeding 100% when the crack width is small. This performance indicates that the tested instrument has poor performance or is unstable.

[0100] Finally, according to the formula A comprehensive quantitative scoring analysis was conducted on the measurement performance of the tunnel lining crack width measurement system (also known as the tested instrument or the measured device), resulting in a final quantitative score of 90.91. This result indicates that the performance of the tested instrument is relatively good, but not yet perfectly ideal. A perfectly ideal score would be 100. This evaluation method clearly confirms the performance of the tested instrument. For tunnel management and maintenance departments, this allows for the targeted selection of better-performing tunnel lining crack width measurement equipment, leading to more effective detection and monitoring of tunnel linings.

[0101] Furthermore, this quantitative evaluation scheme can also conduct in-depth evaluation of the measurement performance of the test instrument from two other aspects: the evaluation is carried out from the perspectives of width and crack angle, and the calculations involved in the evaluation are all based on formula (4).

[0102] ① Measurement performance for cracks of different widths; ② Measurement performance for cracks with different opening angles. These two performance scores reflect different measurement problems of the instrument under test and can provide different optimization suggestions. Let's assume the scores for these two aspects are Score_1 and Score_2, respectively. A low Score_1 usually indicates that the device is insufficient in identifying small cracks, which may lead to missed detections. In this case, it is necessary to increase the acquisition resolution or improve the image noise reduction accuracy. Conversely, a low Score_2 indicates that the device's image may be distorted, making it difficult to identify and calculate cracks in different directions, which may lead to incorrect detection and calculation. In this case, targeted image correction processing is required.

[0103] It should be noted that: Score_1 calculates the quantitative performance score of the tested device for different widths. For example, if a tested device scores 83.05 and 89.38 for cracks with widths of 0.1mm and 0.2mm respectively, it indicates that the tested device's measurement performance for small-sized cracks is slightly weaker. Score_2 calculates the quantitative performance score of the tested device for different tilt angles. For example, if a tested device scores 83.05, 80.61, and 89.38 for cracks with angles of 0°, 50°, and 90° respectively, it indicates that the tested device's measurement performance for transverse and longitudinal cracks is relatively good, while its measurement performance for oblique cracks is weak.

[0104] Both Score1 and Score2 require pre-selective grouping of all measurement results, such as grouping those with the same width or the same angle, before calculating according to Formula 4. Alternatively, one can directly calculate the overall score of the device using Formula 4 without classifying or processing by angle or width. However, this overall score is less instructive than Score_1 and Score_2. Instead, it most directly reflects the overall performance of the device.

[0105] The quantitative scores of the measured performance of the tested instrument under different crack widths are as follows: Figure 11 As shown in the figure, the trend of the quantization score shows that the quantization score of the tested instrument gradually increases with the increase of the crack width, reaching a maximum of 95.29. However, when identifying cracks with a width of less than 0.2 mm, the quantization score of the tested instrument is less than 70.00, exhibiting poor measurement performance. This result indicates that the measurement accuracy of the tested instrument is not ideal, and its measurement performance can be improved by increasing the image acquisition resolution or enhancing the image processing algorithm.

[0106] The quantitative scores of the tested instrument for cracks with different crack angles θ are as follows: Figure 12As shown, under different included angles θ, the scores of the tested instrument were relatively stable, with upper and lower limits ranging from 86.50 to 95.43, and an average score of 90.91. Analysis revealed that the quantitative score did not exhibit a significant variation with the included angle θ, further confirming the stability of the tested instrument in measuring cracks at different angles.

[0107] In summary, this study proposes a performance evaluation method based on simulated cracks arranged in a fan shape, which quantifies the performance of a tunnel lining crack width measurement system through scoring. It provides a solid theoretical basis and practical experimental tool for the scientific, objective, and standardized evaluation of tunnel lining crack detection equipment, especially the accuracy of crack width measurement, by combining dual-dimensional quantitative indicators (IR and RE). This method not only facilitates horizontal comparison and optimization guidance of equipment performance but will also promote the development of tunnel structural health monitoring technology towards greater accuracy, efficiency, and standardization.

[0108] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative in all respects and are not the only ones. All modifications within the scope of this invention or its equivalents are included in this invention.

Claims

1. A measuring device for evaluating the performance of a tunnel lining crack width measurement system, characterized in that, The measuring device includes: The test board has multiple simulated cracks formed on it. The multiple simulated cracks form M angles with a first direction. The angles of the M angles are different from each other. The angles of the M angles have: a first angle set, a second angle set, and a third angle set. The third angle set contains more than or equal to 2 elements. Wherein, the first set of angles has a value range of 0° and / or 180°, the second set of angles has a value range of 90°, and the third set of angles has a value range of the first angle range and / or the second angle range, wherein the first angle range is 0° to 90° and the second angle range is 90° to 180°.

2. The measuring device according to claim 1, characterized in that, The simulated cracks have a straight direction, there are M simulated cracks, and the width of the simulated cracks is equal.

3. The measuring device according to claim 2, characterized in that, The simulated cracks described in M ​​are distributed in a fan shape, a 360° full circle, an irregular random distribution, or an asymmetric array distribution generated based on probability distribution.

4. The measuring device according to claim 2 or 3, characterized in that, The simulated crack width set has a first width subset, a second width subset, and a third width subset. The width range in the first width subset is: a first set width threshold H1 to a second identified width threshold H2. The width range in the second width subset is: a second identified width threshold H2 to a third identified width threshold H3. There are multiple widths in the second width subset. The width range in the third width subset is: a third identified width threshold H3 to a fourth set width threshold H4. H2 is less than H3 and all of them are the widths of simulated cracks that can be identified by the tunnel lining crack width measurement system. The width of the M simulated cracks covers all widths in the set of simulated crack widths.

5. The measuring device according to claim 4, characterized in that, M is greater than the number of widths in the set of simulated crack widths, and the set of simulated crack widths corresponds to the widths of the simulated cracks in a random distribution.

6. The measuring device according to claim 4, characterized in that, H2 is 0.1 mm, H3 is 1 mm, M=19, and the interval angle between adjacent simulated cracks is the same.

7. A method for evaluating the performance of a tunnel lining crack width measurement system using the measuring device according to any one of claims 1-6, characterized in that, The method includes: Set up a test environment and fix multiple test boards in the test environment. The width distribution of multiple simulated cracks in each test board is different. A target tunnel lining surface image is dynamically acquired by the tunnel lining crack width measurement system in the test environment, and the target tunnel lining surface image contains the test plate; The surface image of the target tunnel lining is processed to obtain measurement results, which include the width of each simulated crack in each test plate; Based on the measurement results, the performance of the tunnel lining crack width measurement system is evaluated from two dimensions: simulated crack recognition rate and relative error of crack width.

8. The method according to claim 7, characterized in that, The test environment was a simulated tunnel; The acquisition of the target tunnel lining surface image dynamically collected by the tunnel lining crack width measurement system in the test environment includes: The tunnel lining crack width measurement system to be detected is driven into the simulated tunnel at the detection speed; Obtain the surface image of the target tunnel lining acquired by the tunnel lining crack width measurement system.

9. The method according to claim 8, characterized in that, The detection speed is 60–100 km / h.

Images