Non-contact concrete apparent crack detection method and device and electronic equipment

Abstract

The invention provides a non-contact concrete apparent crack detection method and device and electronic equipment, and relates to the field of data processing. The method comprises the following steps: acquiring full-field reflection intensity data after the concrete surface is irradiated by a digital micromirror array and a spatial light modulator; determining a coding mode matrix of the concrete surface irradiated by the digital micromirror array and the spatial light modulator; according to the full-field reflection intensity data and the coding mode matrix, adopting a compressed sensing reconstruction model to obtain a two-dimensional reflectivity image of the concrete surface; and carrying out adaptive threshold segmentation and morphological refinement on the two-dimensional reflectivity image to obtain apparent crack distribution data. By implementing the technical scheme provided by the invention, the concrete apparent crack detection precision can be conveniently improved.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and in particular to a non-contact concrete surface crack detection method, device and electronic equipment. Background Art

[0002] With the continuous expansion of urbanization and infrastructure in my country, concrete structures such as roads, bridges, tunnels, and underground pipeline corridors are shouldering an increasingly heavy burden of transportation and environmental protection. Concrete cracks, the most direct indicator of structural health, not only accelerate steel corrosion and aggregate loss, but can also cause water seepage, freeze-thaw damage, and in severe cases, even structural failure and accidents. Therefore, it is necessary to detect apparent cracks in concrete.

[0003] Existing techniques for detecting cracks in concrete typically rely solely on visible light images and manual experience. However, the concrete surface itself exhibits significant roughness and irregular speckle noise. Coupled with the large variations in ambient lighting and the alternation of shadows and highlights, conventional camera images struggle to reliably capture the details of micron-level or shallow cracks. Furthermore, the varying degrees of manual experience can lead to distorted crack assessments. Consequently, these methods significantly reduce the accuracy of concrete surface crack detection.

[0004] Therefore, there is an urgent need for a non-contact concrete surface crack detection method, device and electronic equipment. Summary of the Invention

[0005] The present application provides a non-contact concrete surface crack detection method, device and electronic equipment, which are convenient for improving the accuracy of concrete surface crack detection.

[0006] In a first aspect of the present application, a non-contact method for detecting apparent cracks in concrete is provided, the method comprising: obtaining full-field reflection intensity data of a concrete surface after being illuminated by a digital micromirror array and a spatial light modulator; determining a coding pattern matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface; obtaining a two-dimensional reflectivity image of the concrete surface using a compressed sensing reconstruction model based on the full-field reflection intensity data and the coding pattern matrix; and performing adaptive threshold segmentation and morphological refinement on the two-dimensional reflectivity image to obtain apparent crack distribution data.

[0007] By adopting the above-mentioned technical solution, a dynamically controllable illumination coding pattern is formed by combining a digital micromirror array with a spatial light modulator. Utilizing the principle of compressed sensing, a two-dimensional reflectivity image of the concrete surface can be efficiently reconstructed by collecting only a small amount of full-field reflection intensity data. This significantly reduces the amount of data collected and the hardware complexity compared to traditional imaging methods, making it particularly suitable for complex engineering environments such as low illumination and high noise. After imaging, an adaptive threshold segmentation algorithm is introduced to dynamically adjust the crack extraction threshold based on the reflective characteristics of different regions, enhancing the ability to identify small cracks and low-contrast areas. Combined with morphological refinement processing, the continuity and accuracy of crack contours are further improved, significantly reducing false cracks and false detection interference. The overall process offers the advantages of non-contact, high sensitivity, low power consumption, and strong anti-interference ability, enabling rapid, stable, and quantifiable detection of concrete surface cracks. This facilitates improved accuracy in detecting apparent cracks in concrete.

[0008] Optionally, obtaining full-field reflection intensity data of the concrete surface after being irradiated by the digital micromirror array and the spatial light modulator specifically includes: loading a coding pattern into the digital micromirror array and the spatial light modulator according to a preset sequence; controlling a synchronization controller to transmit a trigger signal to the digital micromirror array and the spatial light modulator to control the digital micromirror array and the spatial light modulator to irradiate the concrete surface according to the coding pattern; obtaining a reflection intensity value and an irradiation timestamp of the concrete surface after being irradiated by each of the coding patterns; and performing simple smoothing and de-peaking on the reflection intensity value, the irradiation timestamp, and the coding pattern to obtain the full-field reflection intensity data.

[0009] By employing this technical solution, preset coded illumination patterns are sequentially loaded into a digital micromirror array and a spatial light modulator. A synchronous controller precisely controls both to emit light beams in a synchronized sequence to illuminate the concrete surface, achieving high-precision, programmable dynamic light field modulation of the target area. This illumination method overcomes the limitations of traditional static illumination. Multi-mode coded illumination significantly increases the dimensionality and density of information acquired, providing a rich observational foundation for subsequent compressed sensing imaging. During data acquisition, not only are the reflection intensity values ​​corresponding to each coded pattern acquired, but the illumination timestamp is also recorded simultaneously, ensuring temporal traceability and sequential consistency of the reflectance data, facilitating multi-frame comparison and time-domain processing. Simple smoothing and spike removal of the acquired raw data effectively suppresses interference such as optical noise and sudden reflection changes during field acquisition, enhancing data quality and robustness. Overall, this solution acquires high-dimensional information with a small sampling rate, ensuring precise illumination control while improving the stability and integrity of data acquisition, providing reliable data support for subsequent image reconstruction and crack extraction.

[0010] Optionally, determining the coding pattern matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface specifically includes: obtaining the coding pattern corresponding to each sequence in the preset sequence; generating a mapping table according to the coding pattern and the reflection intensity value, the mapping table including the light field distribution grayscale and trigger timing corresponding to each coding pattern; and outputting the coding pattern matrix in chronological order according to the mapping table.

[0011] By employing the above-mentioned technical solution, each coding pattern in a preset sequence is parsed one by one to obtain its specific light field projection form in the digital micromirror array and spatial light modulator. Combined with the reflection intensity values, a one-to-one light field response mapping table is constructed, thereby accurately linking the coding pattern with the actual illumination response. This mapping table not only contains the spatial grayscale distribution information generated by each coding pattern, but also records the specific timing of its triggering, ensuring that the illumination state corresponding to each pattern can be accurately restored according to the time sequence during the subsequent image reconstruction process. This mapping mechanism effectively compensates for timing errors and light field offsets caused by system response delays, device jitter, or light source fluctuations in actual operation, thereby improving the accuracy and robustness of the reconstructed image. Furthermore, by precisely binding the coding and illumination responses, this method essentially forms a temporally and spatially consistent coded light field database, providing a high-quality, structured input foundation for subsequent compressed sensing reconstruction, avoiding the problem of mismatching image content and illumination patterns in traditional image acquisition.

[0012] Optionally, the two-dimensional reflectivity image of the concrete surface is obtained using a compressed sensing reconstruction model based on the full-field reflection intensity data and the coding pattern matrix, specifically including: determining a measurement vector based on the full-field reflection intensity data and the coding pattern matrix; normalizing and background-subtracting the measurement vector to obtain a vector to be input; standardizing the coding pattern matrix to obtain a matrix to be input; and inputting the vector to be input and the matrix to be input into the compressed sensing reconstruction model to generate the two-dimensional reflectivity image.

[0013] By adopting the technical scheme, the collected full-field reflection intensity data and the corresponding coding mode matrix are jointly processed to construct a measurement vector accurately expressing the projection and response relationship of the light field, thereby effectively improving the initial information quality of the compressed sensing modeling. After the normalization processing of the measurement vector, the data distribution under different intensity scales is unified, and the adaptability of the model to light intensity changes is enhanced. The background subtraction operation eliminates the interference of non-structural reflection, environmental noise and system background light on the imaging result, further improving the contrast and crack detail performance of the reconstructed image. At the same time, the standardization processing of the coding mode matrix makes different light field modes have equal weight characteristics in the model input, avoiding the reconstruction deviation caused by uneven light intensity or mode distribution difference, thereby ensuring the stability and convergence of the model solving process. Finally, the normalized vector and the standardized matrix are jointly input into the compressed sensing reconstruction model, which can realize the efficient recovery of the two-dimensional reflectivity image of the concrete surface, greatly reducing the required sampling number and data redundancy, and accurately restoring the micro crack texture details under the condition of extremely low data amount.

[0014] Optionally, the adaptive threshold segmentation and morphological refinement of the two-dimensional reflectivity image to obtain apparent crack distribution data specifically includes: statistical gray histogram features and local contrast distribution features of the two-dimensional reflectivity image; determining an initial segmentation threshold according to the gray histogram features and the local contrast distribution features; generating a preliminary binary image according to the two-dimensional reflectivity image and the initial segmentation threshold; performing open operation, close operation and skeleton thinning on the preliminary binary image to determine skeleton line coordinates and connected topological structure; and generating the apparent crack distribution data based on the skeleton line coordinates and the connected topological structure.

[0015] By employing this technical solution, the initial segmentation threshold is dynamically determined during the post-processing of the 2D reflectance image by fusing the image's grayscale histogram features with the local contrast distribution. This approach accurately adapts to the complex textures and uneven brightness of concrete surfaces, avoiding the incomplete crack extraction and misidentification issues often encountered by traditional fixed-threshold segmentation methods when processing low-contrast or high-noise images. The local contrast analysis mechanism employed effectively enhances sensitivity to minute crack boundaries, ensuring accurate identification of crack regions even in the presence of uneven illumination or surface contamination. After generating the initial binary image, opening and closing operations are introduced to eliminate isolated noise points and connect broken crack segments, improving the continuity and readability of the crack morphology. The skeletonization and refinement step maximizes image information while preserving the crack topology, enabling accurate restoration of crack paths and facilitating subsequent quantitative analysis and parameter extraction. By extracting skeleton line coordinates and constructing a connected topology, the system not only captures the spatial distribution of cracks but also identifies crack orientations, intersections, and branching structures, providing advanced data support for structural health assessment and damage evolution analysis.

[0016] Optionally, the apparent crack distribution data is generated based on the skeleton line coordinates and the connected topological structure, specifically including: determining the skeleton pixel points and the adjacency relationship of the skeleton pixel points according to the skeleton line coordinates, abstracting them into a graph structure; determining the endpoints, bifurcation nodes and intermediate nodes of the graph structure; splitting the graph structure into crack networks through connected components, and each crack network traverses from the endpoint along the intermediate node to the bifurcation node or the next endpoint; extracting a branch coordinate sequence based on the crack network; calculating the geometric length, local strike angle and bounding box of the branch coordinate sequence to obtain attribute data; and organizing the apparent crack distribution data according to a preset data format based on the branch coordinate sequence and the attribute data.

[0017] By adopting the above technical solution and systematically modeling and decomposing the skeleton line coordinates and their connected topological structure, a crack characterization method for structural understanding and data-based representation was constructed, significantly improving the accuracy, interpretability, and engineering practicality of crack detection results. Specifically, the skeleton pixels and their adjacency relationships were first abstracted into a graph structure, effectively restoring the continuity and distribution of cracks in space and achieving a structured understanding of crack morphology. By identifying endpoints, bifurcations, and intermediate nodes in the graph structure, the system automatically identifies key features such as the start and end points of crack paths and branching changes, laying the foundation for subsequent path extraction and pattern recognition. Furthermore, connected component analysis was used to decompose the graph structure into multiple independent crack networks. Branch coordinate sequences were extracted along the node relationships, allowing each crack path to be accurately isolated and reconstructed. Geometric analysis of the branch sequences was then performed to extract spatial attributes such as length, strike angle, and bounding box. This not only provides rich metrics for quantitative crack analysis but also enables cross-scenario transfer and comparison of crack features.

[0018] Optionally, the method further includes: determining the coordinates of the abnormal position according to the apparent crack distribution data; and sending the coordinates of the abnormal position to a user device corresponding to the inspection personnel.

[0019] By adopting the above technical solution, the practicality of crack detection results, response efficiency and convenience of on-site operation are significantly improved by automatically identifying abnormal position coordinates based on the apparent crack distribution data and accurately pushing them to the user device of the inspector. First, by intelligently distinguishing abnormal features in the crack distribution data, areas with potential safety risks or requiring key review can be quickly screened out, greatly reducing the workload of inspectors in checking massive amounts of data one by one. Secondly, the extracted abnormal position coordinates are highly accurate and spatially locatable. Combined with the three-dimensional model of the structure or the site plan, they can achieve point-to-point rapid positioning of cracks in practical applications, effectively shortening the survey path and assessment cycle, and improving the timeliness of engineering response and the quality of decision-making. More importantly, by sending this type of abnormal coordinate information to the user device carried by the inspector in a timely and accurate manner, a closed-loop process from background intelligent analysis to front-end visual guidance is realized, promoting the deep integration of human-machine collaboration and intelligent assistance, and enhancing the information transparency and collaborative efficiency of on-site work.

[0020] In a second aspect of the present application, a non-contact concrete apparent crack detection device is provided, the detection device comprising an acquisition module and a processing module, wherein the acquisition module is configured to acquire full-field reflection intensity data of a concrete surface irradiated by a digital micromirror array and a spatial light modulator; the processing module is configured to determine a coding mode matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface; the processing module is further configured to obtain a two-dimensional reflectivity image of the concrete surface by using a compressed sensing reconstruction model according to the full-field reflection intensity data and the coding mode matrix; and the processing module is further configured to perform adaptive threshold segmentation and morphological thinning on the two-dimensional reflectivity image to obtain apparent crack distribution data.

[0021] In a third aspect of the present application, an electronic device is provided, comprising a processor, a memory, a user interface and a network interface, the memory is configured to store instructions, the user interface and the network interface are configured to communicate with other devices, and the processor is configured to execute the instructions stored in the memory to enable the electronic device to perform the method described above.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided, which stores instructions that, when executed, perform the method described above.

[0023] In summary, the one or more technical solutions provided in the present application have at least the following technical effects or advantages: By combining a digital micromirror array with a spatial light modulator to form a dynamically controllable light irradiation coding mode, and using the principle of compressed sensing, a two-dimensional reflectivity image of the concrete surface can be efficiently reconstructed by only collecting a small amount of full-field reflection intensity data, which greatly reduces the data acquisition amount and hardware complexity compared to traditional imaging methods, and is particularly suitable for complex engineering environments such as low illumination and high noise. After imaging, by introducing an adaptive threshold segmentation algorithm, the crack extraction threshold can be dynamically adjusted according to the reflection characteristics of different regions, enhancing the ability to identify small cracks and low-contrast regions. Combined with morphological thinning processing, the continuity and accuracy of the crack profile are further improved, significantly reducing false cracks and false detection interference. The overall process has the advantages of non-contact, high sensitivity, low power consumption, strong anti-interference, etc., and can realize rapid, stable and quantifiable detection of concrete surface cracks. Therefore, it is convenient to improve the accuracy of concrete apparent crack detection. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 A flowchart of a non-contact concrete apparent crack detection method provided by an embodiment of the present application is shown; Figure 2Another schematic flow chart of a non-contact concrete surface crack detection method provided in an embodiment of the present application; Figure 3 A schematic diagram of a module of a non-contact concrete surface crack detection device provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0025] Explanation of the reference numerals: 31, acquisition module; 32, processing module; 41, processor; 42, communication bus; 43, user interface; 44, network interface; 45, memory. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0027] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0028] In the description of the embodiments of the present application, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0029] With the continuous advancement of urbanization process and the continuous expansion of infrastructure construction scale in China, concrete structures such as highways, bridges, tunnels and underground pipe galleries play a key role in bearing traffic operation and environmental protection. However, once cracks occur in concrete, not only are they easy to cause steel corrosion, aggregate spalling, but also may cause chain reactions such as water seepage, freeze-thaw damage, and even threaten the stability and safety of the entire structure system. Therefore, timely and accurate detection of concrete surface cracks has become an indispensable important link to ensure the health of engineering structures.

[0030] However, under the existing technical conditions, crack detection usually relies on traditional visible light cameras to obtain images, supplemented by manual experience for interpretation. This way is easy to be affected by environmental factors such as changes in lighting conditions, shadow obstruction and mirror reflection when facing complex texture, high roughness and random speckle interference on the surface of concrete, resulting in large fluctuations in image quality, and crack details, especially micro cracks and early developing cracks, are difficult to be stably identified. At the same time, manual judgment has strong subjectivity and poor consistency, which is easy to cause misjudgment and missed detection of crack identification, thereby significantly limiting the overall detection accuracy and reliability of engineering application.

[0031] In order to solve the above technical problems, the present application provides a non-contact concrete surface crack detection method, referring to Figure 1 , Figure 1 A flowchart of a non-contact concrete surface crack detection method provided by an embodiment of the present application. The method is applied to a server and includes steps S110 to S140, which are as follows: S110, obtaining full-field reflection intensity data of the concrete surface after being irradiated by a digital micromirror array and a spatial light modulator.

[0032] Specifically, the central server in the system is responsible for sending instructions to the digital micromirror array and the spatial light modulator located on the scene, so that they irradiate the concrete surface in turn according to the pre-set encoding mode; then, the system measures and collects the reflection light intensity of the entire illumination area at one time through the condensing device or the single-pixel detector, and uploads these numerical light intensity information together with the corresponding illumination time stamp to the server. In other words, the server not only achieves precise control of the light field projection, but also records the complete and real-time data of each light reflection on the scene, which lays a foundation for subsequent image reconstruction and crack analysis based on these intensity data.

[0033] For example, during a bridge pier surface inspection, engineers installed a digital micromirror array and spatial light modulator on a mobile platform and sent control signals to a backend server via a wireless network. The server sequentially controlled the switching of hundreds of coded light patterns. With each frame, the detector measured the reflected light intensity across the entire field of view in real time and uploaded a record such as "Coded pattern 37, reflected intensity 125.3 mW; timestamp 2025-04-25, 14:32:07.123." The resulting "full-field reflection intensity data" generated by the server resembles a dense "optical fingerprint" that can be used to reconstruct subtle texture variations in the concrete surface, allowing precise identification of crack location and shape.

[0034] In one possible embodiment, obtaining full-field reflection intensity data of a concrete surface after being irradiated by a digital micromirror array and a spatial light modulator specifically includes: loading a coding pattern into the digital micromirror array and the spatial light modulator according to a preset sequence; controlling a synchronization controller to transmit a trigger signal to the digital micromirror array and the spatial light modulator to control the digital micromirror array and the spatial light modulator to irradiate the concrete surface according to the coding pattern; obtaining a reflection intensity value and an illumination timestamp of the concrete surface after being irradiated by each coding pattern; and performing simple smoothing and de-peaking on the reflection intensity value, the illumination timestamp, and the coding pattern to obtain full-field reflection intensity data.

[0035] Specifically, first, the system will load a set of predefined grayscale or phase patterns into the DMD / SLM device in sequence according to a pre-designed light field coding sequence; then, a unified trigger signal will be sent through the synchronization controller to ensure that the two devices start and stop irradiation at the same time, so as to achieve precise timing coordination. For the full-field reflected light intensity after irradiation with each coding pattern, the detector will measure and record the corresponding timestamp in real time, and then send back these original intensity values, corresponding coding numbers and time tags. Finally, in order to eliminate the noise introduced by sudden light intensity jitter and device response delay, the system will perform a simple smoothing process on the reflection intensity sequence and timestamp, and eliminate spike anomalies to obtain stable and consistent full-field reflection intensity data.

[0036] For example, during a bridge pier crack inspection, engineers prepare 100 random light field patterns, numbered 1 to 100. The system first loads pattern 1 onto the DMD, and the SLM simultaneously switches to the corresponding state. After the synchronization controller issues a trigger signal, the detector measures a reflection intensity of 132.7 mW and timestamps it with "2025-04-25 09:15:02.356." The detector then switches to pattern 2 and repeats the measurement. This cycle continues until pattern 100 is reached. After acquisition, the software performs a smoothing filter on these 100 sets of (pattern number, intensity, and time) records to remove abnormal spikes that may be caused by mechanical micro-vibration or light source jitter, ultimately producing a continuous, clean intensity-time data table. This pre-processed full-field reflection intensity data serves as a key measurement input for the subsequent compressed sensing reconstruction process, ensuring higher accuracy and consistency in the reconstructed concrete surface reflectivity image.

[0037] S120: Determine a coding pattern matrix for the digital micromirror array and the spatial light modulator to illuminate the concrete surface.

[0038] Specifically, the server first reads the detailed configuration of each encoding mode from the pattern library—including grayscale distribution, phase modulation scheme, and trigger timing identifiers. Then, according to the preset scanning or testing strategy, these modes are arranged into a matrix structure by row. Each row corresponds to a coding mode, and each column records the specific control parameters of the mode at different times or different hardware channels. In this way, the server can send the entire coding mode matrix to the DMD / SLM controller at once with a single command, ensuring that the subsequent light field projection is both accurate and efficient.

[0039] For example, suppose the pattern library contains eight random or orthogonal codes, each associated with three grayscale settings and two sets of timing signals. The server arranges this 8×5 configuration table into a matrix: row 1 records grayscale modes A, B, and C for mode 1 and their corresponding trigger times t1 and t2; row 2 records mode 2, and so on, up to row 8. When projection begins, the server only needs to issue this matrix once, and the DMD and SLM controllers automatically switch to mode 1, then mode 2, and so on, sequentially, according to the matrix row sequence to complete the coded illumination of the entire area, thus maintaining reproducibility and timing consistency in the complex light field modulation process.

[0040] In one possible implementation, determining a coding pattern matrix for illuminating a concrete surface using a digital micromirror array and a spatial light modulator specifically includes: obtaining a coding pattern corresponding to each sequence in a preset sequence; generating a mapping table based on the coding pattern and the reflection intensity value, the mapping table including the light field distribution grayscale level and trigger timing corresponding to each coding pattern; and outputting the coding pattern matrix in chronological order according to the mapping table.

[0041] Specifically, first, the server will read the definition of each encoding mode in the "preset sequence", which may come from laboratory design or on-site optimization, and each mode contains a specific pixel gray scale distribution or phase modulation scheme. Then, the server matches these modes with the corresponding reflection intensity values collected previously, generating a mapping table that records not only the gray level configuration of a certain mode, but also the exact timing when the mode is triggered. Finally, the server organizes all modes into a matrix form according to the time sequence recorded in the mapping table, and outputs them to the controller at once to ensure that subsequent light field switching is neither missed nor out of order.

[0042] For example, suppose in a field test, three encoding modes are designed: mode A is used to highlight rough texture, mode B to enhance reflection highlights, and mode C to enhance crack edge contrast. The measured reflection intensities are 110.5 mW, 92.3 mW and 118.7 mW, and the corresponding trigger timestamps are 10:00:01.250, 10:00:02.100 and 10:00:03.450. The server integrates this information into the mapping table, for example: A→gray scale distribution [50, 200,...]; trigger time 10:00:01.250. B→gray scale distribution [30, 180,...]; trigger time 10:00:02.100. C→gray scale distribution [70, 220,...]; trigger time 10:00:03.450. Finally, the server outputs a three-row encoding mode matrix in this time sequence, and the controller can accurately drive the DMD / SLM in the order of A→B→C to achieve high consistency of dynamic illumination.

[0043] S130, according to the full-field reflection intensity data and the encoding mode matrix, a two-dimensional reflectivity image of the concrete surface is obtained by using a compressed sensing reconstruction model.

[0044] Specifically, the server sends the "full-field reflection intensity data" (i.e. the overall reflection value measured under each light field encoding mode) and the "encoding mode matrix" (which records the spatial distribution and trigger sequence of each mode) into the compressed sensing reconstruction model. This model takes advantage of the sparse nature of concrete surface cracks in spatial distribution, and through optimization algorithm under the dual constraints of measurement value and encoding mode, iteratively finds the pixel layout that best fits the true reflectivity distribution. In other words, although only a small amount of overall intensity data is collected, the compressed sensing algorithm can still rely on the pre-defined lighting structure and combine the prior sparse assumption to restore a high-resolution two-dimensional reflectivity image, clearly showing the location and direction of the cracks.

[0045] For example, assuming 80 different coding patterns are projected on-site, the server collects 80 reflection intensity values ​​and then inputs them, along with an 80×40,000 coding matrix (corresponding to a 200×200 pixel resolution), into the reconstruction model. The model iterates through several rounds, gradually adjusting the reflectivity estimate for each pixel until both the measurement error and sparsity constraints are satisfied. The resulting 200×200 grayscale image shows crack locations as distinct low-reflectivity bands. Compared to traditional point-by-point scanning or fixed illumination, this approach not only significantly reduces the number of measurements but also maintains sensitivity to micron-level crack details in complex lighting environments, providing a high-quality baseline image for subsequent crack extraction and structural assessment.

[0046] In one possible implementation, a compressed sensing reconstruction model is used to obtain a two-dimensional reflectivity image of the concrete surface based on the full-field reflection intensity data and the coding pattern matrix, specifically including: determining a measurement vector based on the full-field reflection intensity data and the coding pattern matrix; normalizing and background-subtracting the measurement vector to obtain a vector to be input; standardizing the coding pattern matrix to obtain a matrix to be input; and inputting the vector to be input and the matrix to be input into the compressed sensing reconstruction model to generate a two-dimensional reflectivity image.

[0047] Specifically, the system first extracts a measurement vector based on the entire reflection intensity values ​​corresponding to all coding patterns. This is an ordered record of the overall response after each light projection. Next, to eliminate offsets between different patterns caused by light source fluctuations or surface diffuse reflection differences, the measurement vector must be normalized (bringing the data into the same dimensional range) and the background value must be subtracted (eliminating ambient light and fixed reflection components) to obtain an "input vector" to ensure that subsequent algorithms focus on the crack signal itself. Normalizing the original coding pattern matrix is ​​to place each light field mode on the same weight platform during reconstruction, preventing certain high-grayscale or high-power modes from excessively influencing the results. After the above preprocessing, the input vector and input matrix can be simultaneously imported into the compressed sensing reconstruction model, guiding the model to find the optimal balance between minimizing measurement error and sparse priors, and ultimately outputting a two-dimensional image representing the local reflectivity changes on the concrete surface.

[0048] S140 , performing adaptive threshold segmentation and morphological refinement on the two-dimensional reflectivity image to obtain apparent crack distribution data.

[0049] Specifically, the system first performs a grayscale histogram and local contrast analysis on the entire reflectance image, automatically calculating the segmentation threshold that best distinguishes cracks (usually low-reflection bands) from normal concrete surfaces (high-reflection areas), rather than using a fixed brightness threshold. This allows for dynamic adaptation based on the reflective characteristics of different locations and lighting conditions, ensuring that even tiny or weak-contrast cracks can be identified. After obtaining the initial binary image, the server then performs an opening operation to remove noise (such as isolated speckle misidentification) and a closing operation to fill in small fractures. Finally, a skeletonization operation compresses the crack width to a single-pixel centerline, ensuring that each crack appears as a clear, continuous line segment.

[0050] For example, consider a 400×400 pixel reflectance image in which the grayscale values ​​of the crack region are primarily concentrated between 10 and 30, while the background is between 100 and 200. The system first uses the image histogram and local contrast distribution to derive an adaptive threshold of 45, marking pixels with reflectance values ​​below 45 as cracks. While the resulting binary image roughly depicts the crack shape, it may also contain noise points of varying sizes and fractures. The server then uses an opening operation to remove isolated points with a diameter less than 3 pixels and a closing operation to connect fracture segments within 5 pixels of each other. Finally, skeletonization unifies all crack widths into 1-pixel-wide skeleton lines. The system then outputs a coordinate sequence and connectivity topology that accurately reflects the direction and morphology of each crack, which can be directly used as "apparent crack distribution data" for subsequent statistics, positioning, or visualization.

[0051] In one possible implementation, adaptive threshold segmentation and morphological refinement are performed on the two-dimensional reflectivity image to obtain apparent crack distribution data, specifically including: statistically analyzing the grayscale histogram features and local contrast distribution features of the two-dimensional reflectivity image; determining an initial segmentation threshold based on the grayscale histogram features and local contrast distribution features; generating a preliminary binary image based on the two-dimensional reflectivity image and the initial segmentation threshold; performing opening and closing operations and skeletonization and refinement on the preliminary binary image to determine the skeleton line coordinates and the connected topological structure; and generating apparent crack distribution data based on the skeleton line coordinates and the connected topological structure.

[0052] Specifically, the system first performs two statistical analyses on the entire image: first, it calculates a global grayscale histogram to determine the frequency of occurrence of pixels at each grayscale level; second, it evaluates small-scale contrast variations in different regions of the image, identifying locations where local brightness changes are most pronounced. Based on these two statistical results, the algorithm dynamically selects the grayscale threshold that best distinguishes cracks from the background, rather than simply applying a fixed value. Next, the system uses this threshold to binarize the image, marking areas with grayscale levels below the threshold as potential cracks, creating a preliminary black-and-white image.

[0053] After obtaining the preliminary binary image, in order to eliminate noise and fill the crack interruptions, the system sequentially performs morphological opening operations (removing isolated small spots) and closing operations (connecting similar fractured segments), and finally performs skeletonization on the remaining crack areas, compressing the strip cracks of varying widths into a single-pixel wide center line. In this step, the system records the pixel coordinates of each center line and their connectivity with each other (that is, which line segments are connected at intersections or branches). For example, in a 500×500 pixel reflectivity image, the grayscale histogram may show that the crack pixels are concentrated between 20-40 grayscale, while the background is mostly between 100-200; the algorithm automatically sets the threshold to 60 based on this, and extracts all points in the image below 60 as binary foregrounds. After morphological filtering, the originally discontinuous crack bands are spliced ​​and extracted into center lines, and the system derives a set of information such as "(x1,y1)→(x2,y2)→…→(x n ,y n )” skeleton coordinate sequence and their connection topology at bifurcation nodes ultimately form concise and complete crack distribution data, which is convenient for subsequent length measurement, direction statistics and visualization.

[0054] In one possible implementation, apparent crack distribution data is generated based on skeleton line coordinates and a connected topological structure, specifically including: determining skeleton pixel points and the adjacency relationship of skeleton pixel points as a graph structure according to the skeleton line coordinates; determining the endpoints, bifurcation nodes, and intermediate nodes of the graph structure; splitting the graph structure into a crack network through connected components, and each crack network traverses from the endpoint along the intermediate node to the bifurcation node or the next endpoint; extracting a branch coordinate sequence based on the crack network; calculating the geometric length, local strike angle, and bounding box of the branch coordinate sequence to obtain attribute data; and organizing the apparent crack distribution data according to a preset data format based on the branch coordinate sequence and the attribute data.

[0055] Specifically, the crack skeleton, refined into single-pixel centerlines, is first abstracted into a graph structure consisting of "nodes" and "edges": each pixel is a node in the graph, and if two pixels are adjacent in the image, an edge is connected between them. The system then automatically distinguishes three types of key nodes by counting the number of connections at each node: "endpoints" with only one edge, "intermediate nodes" with two edges, and "fork nodes" with three or more edges. Using a graph algorithm, the entire skeleton is split into several "connected components" (i.e., independent crack networks). For each component, the algorithm starts from the endpoint and traverses along the intermediate nodes to the next fork node or endpoint, thereby extracting a continuous sequence of "branch" pixel coordinates.

[0056] After obtaining the coordinate sequence for each branch, the system further calculates its geometric properties: The actual length of each crack segment is determined by accumulating the physical distances between adjacent coordinate points; the local strike angle is determined based on the spatial positions of the sequence's endpoints; and a minimum bounding rectangle (or convex hull) is generated for each segment for rapid location and visualization. After all branches are numbered according to a unified format, the system packages their coordinate sequences along with corresponding lengths, angles, bounding boxes, and other attributes into structured data (e.g., JSON or GeoJSON), generating the final "apparent crack distribution data." For example, a single inspection result might contain three main cracks (C1, C2, and C3), each further divided into several numbered segments (C1-S1, C1-S2, etc.). Their coordinate lists, lengths (e.g., 12.5 cm, 8.3 cm), orientations (e.g., 45°, 120°), and bounding box information are clearly recorded, facilitating subsequent visualization, statistical analysis, and engineering decision-making.

[0057] In one possible implementation, refer to Figure 2 , Figure 2 Another flow chart of a non-contact concrete apparent crack detection method provided in an embodiment of the present application includes steps S210 to S220, and the above steps are as follows: S210, determining the abnormal position coordinates based on the apparent crack distribution data; S220, sending the abnormal position coordinates to the user device corresponding to the detection personnel.

[0058] Specifically, abnormal locations refer to crack nodes that meet preset risk or quality thresholds, such as crack lengths exceeding a certain protocol standard, widths exceeding the safety upper limit, multiple cross branches, or locations at critical load-bearing parts of the structure. The server will traverse the properties of all crack segments, automatically filter out branches that meet the abnormal conditions, and summarize the coordinates of the starting and ending points or skeleton center points of these branches in the global coordinate system into a list of abnormal locations. For example, in the inspection of a bridge pier, if the length of a crack segment exceeds 15 cm and the estimated width exceeds 0.5 mm, the server will mark the center point of the segment (such as 23.456°N, 113.789°E) as an abnormality, and record the corresponding structural unit, crack ID, and risk level for subsequent precise positioning and disposal.

[0059] After identifying the abnormal coordinates, the server will push the coordinate data to the mobile terminal or the on-site AR device of the detection personnel in real time through the established communication channel. The pushing method can be vibration reminder in the App, SMS alarm or background message synchronization. The message body contains key information such as abnormal coordinates, crack number, risk level and location diagram, so that the detection personnel can quickly locate "point to point" on site. For example, when the bridge daily inspection system detects the abnormality of the pier column crack, the tablet of the on-site technician will immediately pop up a red marker point, showing "abnormal crack ID: C12, location: pier column south side (23.456°N, 113.789°E), risk level: high", and attached with clickable map view or real scene overlay diagram, so that the personnel can immediately go to the point for manual verification or take repair measures.

[0060] The application also provides a non-contact concrete surface crack detection device, which refers to Figure 3 , Figure 3 A module schematic diagram of a non-contact concrete surface crack detection device provided by the embodiment of the application. The device is a server, which includes an acquisition module 31 and a processing module 32. The acquisition module 31 acquires full-field reflection intensity data of a concrete surface after the concrete surface is irradiated by a digital micromirror array and a spatial light modulator. The processing module 32 determines a coding pattern matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface. The processing module 32 obtains a two-dimensional reflectivity image of the concrete surface by using a compressed sensing reconstruction model according to the full-field reflection intensity data and the coding pattern matrix. The processing module 32 performs adaptive threshold segmentation and morphological thinning on the two-dimensional reflectivity image to obtain surface crack distribution data.

[0061] In a possible implementation, the acquisition module 31 acquires full-field reflection intensity data of a concrete surface after the concrete surface is irradiated by a digital micromirror array and a spatial light modulator, specifically including: the processing module 32 loads a coding pattern into the digital micromirror array and the spatial light modulator according to a preset sequence; the processing module 32 controls a synchronization controller to emit a trigger signal to the digital micromirror array and the spatial light modulator, so as to control the digital micromirror array and the spatial light modulator to irradiate the concrete surface according to the coding pattern; the acquisition module 31 acquires reflection intensity values and irradiation time stamps of the concrete surface after being irradiated by each coding pattern; and the processing module 32 performs simple smoothing and de-sharp processing on the reflection intensity values, the irradiation time stamps and the coding pattern to obtain the full-field reflection intensity data.

[0062] In one possible implementation, the processing module 32 determines the coding pattern matrix for illuminating the concrete surface using the digital micromirror array and the spatial light modulator, specifically including: the acquisition module 31 acquires the coding pattern corresponding to each sequence in a preset sequence; the processing module 32 generates a mapping table based on the coding pattern and the reflection intensity value, the mapping table including the light field distribution grayscale and trigger timing corresponding to each coding pattern; and the processing module 32 outputs the coding pattern matrix in chronological order according to the mapping table.

[0063] In one possible implementation, the processing module 32 uses a compressed sensing reconstruction model to obtain a two-dimensional reflectivity image of the concrete surface based on the full-field reflection intensity data and the coding pattern matrix, specifically including: the processing module 32 determines the measurement vector based on the full-field reflection intensity data and the coding pattern matrix; the processing module 32 normalizes and background-subtracts the measurement vector to obtain the vector to be input; the processing module 32 standardizes the coding pattern matrix to obtain the matrix to be input; the processing module 32 inputs the vector to be input and the matrix to be input into the compressed sensing reconstruction model to generate a two-dimensional reflectivity image.

[0064] In one possible embodiment, the processing module 32 performs adaptive threshold segmentation and morphological refinement on the two-dimensional reflectivity image to obtain apparent crack distribution data, specifically including: the processing module 32 counts the grayscale histogram features and local contrast distribution features of the two-dimensional reflectivity image; the processing module 32 determines the initial segmentation threshold based on the grayscale histogram features and the local contrast distribution features; the processing module 32 generates a preliminary binary image based on the two-dimensional reflectivity image and the initial segmentation threshold; the processing module 32 performs opening and closing operations and skeletonization and refinement on the preliminary binary image to determine the skeleton line coordinates and the connected topological structure; the processing module 32 generates the apparent crack distribution data based on the skeleton line coordinates and the connected topological structure.

[0065] In one possible embodiment, the processing module 32 generates apparent crack distribution data based on the skeleton line coordinates and the connected topological structure, specifically including: the processing module 32 determines the skeleton pixels and the adjacency relationship of the skeleton pixels according to the skeleton line coordinates, and abstracts it into a graph structure; the processing module 32 determines the endpoints, bifurcation nodes and intermediate nodes of the graph structure; the processing module 32 splits the graph structure into crack networks through connected components, and each crack network traverses from the endpoint along the intermediate node to the bifurcation node or the next endpoint; the processing module 32 extracts the branch coordinate sequence based on the crack network; the processing module 32 calculates the geometric length, local strike angle and bounding box of the branch coordinate sequence to obtain attribute data; the processing module 32 organizes the apparent crack distribution data according to the preset data format based on the branch coordinate sequence and the attribute data.

[0066] In a possible implementation, the processing module 32 determines the coordinates of the abnormal position according to the apparent crack distribution data; the processing module 32 sends the coordinates of the abnormal position to a user device corresponding to the inspection personnel.

[0067] It should be noted that the above embodiments provide devices that implement their functions using only the division of the above functional modules as examples. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0068] This application also provides an electronic device, referring to Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: at least one processor 41, at least one network interface 44, a user interface 43, a memory 45, and at least one communication bus 42.

[0069] The communication bus 42 is used to realize the connection and communication between these components.

[0070] The user interface 43 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 43 may also include a standard wired interface and a wireless interface.

[0071] The network interface 44 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0072] The processor 41 may include one or more processing cores. Using various interfaces and circuits, the processor 41 connects to various components within the server. It executes instructions, programs, code sets, or instruction sets stored in the memory 45, as well as accesses data stored in the memory 45, to perform various server functions and process data. Optionally, the processor 41 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 41 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing content displayed on the display screen; and the modem handles wireless communications. It is understood that the modem may also be implemented as a separate chip, rather than integrated into the processor 41.

[0073] Among them, the memory 45 may include a random access memory (RAM) or a read-only memory (Read-Only Memory). Optionally, the memory 45 includes a non-transitory computer-readable storage medium. The memory 45 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 45 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 45 may also be optionally at least one storage device located away from the aforementioned processor 41. As Figure 4 As shown, the memory 45 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a non-contact concrete surface crack detection method.

[0074] exist Figure 4In the electronic device shown, the user interface 43 is mainly used to provide an interface for the user to input, and obtain data input by the user; and the processor 41 can be used to call an application program stored in the memory 45 and storing a non-contact concrete apparent crack detection method, which, when executed by one or more processors, causes the electronic device to perform the method of one or more of the above embodiments.

[0075] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all described as a series of action combinations, but those skilled in the art should know that the present application is not limited to the order of the actions described, because according to the present application, certain steps can be performed in other order or at the same time. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0076] The present application also provides a computer-readable storage medium, which stores instructions. When executed by one or more processors, the instructions cause the electronic device to perform the method of one or more of the above embodiments.

[0077] In the above embodiments, the description of each embodiment is focused on, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0078] In several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division, and actual implementation can have another division manner. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some service interface, device or unit, and can be electrical or other forms.

[0079] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e. they can be located in one place, or distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0080] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware, or in the form of a software functional unit.

[0081] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of this application. The aforementioned memory includes various media that can store program code, such as USB flash drives, mobile hard drives, magnetic disks, or optical disks.

[0082] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any variation, use or adaptive change of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A non-contact concrete apparent crack detection method, characterized in that: The method comprises: Acquire full-field reflection intensity data of the concrete surface after being illuminated by a digital micromirror array and a spatial light modulator; determining a coding pattern matrix for illuminating the concrete surface with the digital micromirror array and the spatial light modulator; Obtaining a two-dimensional reflectivity image of the concrete surface using a compressed sensing reconstruction model according to the full-field reflection intensity data and the coding pattern matrix; Adaptive threshold segmentation and morphological refinement are performed on the two-dimensional reflectivity image to obtain apparent crack distribution data.

2. The non-contact concrete apparent crack detection method according to claim 1, characterized in that: The obtaining of full-field reflection intensity data of the concrete surface after being illuminated by the digital micromirror array and the spatial light modulator specifically includes: Loading a coding pattern into the digital micromirror array and the spatial light modulator according to a preset sequence; controlling a synchronization controller to transmit a trigger signal to the digital micromirror array and the spatial light modulator to control the digital micromirror array and the spatial light modulator to illuminate the concrete surface according to the coding pattern; Obtaining a reflection intensity value and an illumination timestamp of the concrete surface after being illuminated by each of the coding patterns; The reflection intensity value, the illumination timestamp, and the encoding pattern are simply smoothed and de-peaked to obtain the full-field reflection intensity data.

3. The non-contact concrete apparent crack detection method according to claim 2, characterized in that: Determining the coding pattern matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface specifically includes: Obtaining a coding mode corresponding to each sequence in the preset sequence; Generate a mapping table according to the coding mode and the reflection intensity value, wherein the mapping table includes a light field distribution grayscale and a trigger timing corresponding to each coding mode; According to the mapping table, the coding mode matrix is ​​output in chronological order.

4. The non-contact concrete apparent crack detection method according to claim 1, characterized in that: The method of obtaining a two-dimensional reflectivity image of the concrete surface using a compressed sensing reconstruction model based on the full-field reflection intensity data and the coding pattern matrix specifically includes: Determining a measurement vector according to the full-field reflection intensity data and the coding pattern matrix; Normalizing and background subtracting the measurement vector to obtain a vector to be input; Normalizing the coding mode matrix to obtain a matrix to be input; The vector to be input and the matrix to be input are input into the compressed sensing reconstruction model to generate the two-dimensional reflectivity image.

5. The non-contact concrete apparent crack detection method according to claim 1, characterized in that: The adaptive threshold segmentation and morphological refinement of the two-dimensional reflectivity image to obtain apparent crack distribution data specifically includes: Counting the grayscale histogram features and local contrast distribution features of the two-dimensional reflectivity image; Determining an initial segmentation threshold according to the grayscale histogram feature and the local contrast distribution feature; generating a preliminary binary image according to the two-dimensional reflectivity image and the initial segmentation threshold; Performing opening and closing operations and skeletonization and thinning on the preliminary binary graph to determine skeleton line coordinates and a connected topological structure; The apparent crack distribution data is generated based on the skeleton line coordinates and the connectivity topology.

6. The non-contact concrete apparent crack detection method according to claim 5, characterized in that: The generating of the apparent crack distribution data based on the skeleton line coordinates and the connected topological structure specifically includes: Determine skeleton pixel points and abstract the adjacency relationship of the skeleton pixel points into a graph structure according to the skeleton line coordinates; Determining endpoints, bifurcation nodes, and intermediate nodes of the graph structure; Splitting the graph structure into crack networks by connected components, each of the crack networks traverses from the endpoint along the intermediate node to the bifurcation node or the next endpoint; Extracting a branch coordinate sequence based on the crack network; Calculating the geometric length, local strike angle, and bounding box of the branch coordinate sequence to obtain attribute data; The apparent crack distribution data is obtained according to the branch coordinate sequence and the attribute data and organized in a preset data format.

7. The non-contact concrete apparent crack detection method according to claim 1, characterized in that: The method further comprises: determining the coordinates of the abnormal position according to the apparent crack distribution data; The abnormal location coordinates are sent to the user equipment corresponding to the detection personnel.

8. A non-contact concrete surface crack detection device, characterized in that: The detection device comprises an acquisition module (31) and a processing module (32), wherein: The acquisition module (31) is used to acquire full-field reflection intensity data of the concrete surface after being illuminated by the digital micromirror array and the spatial light modulator; The processing module (32) is used to determine the coding pattern matrix of the digital micromirror array and the spatial light modulator irradiating the concrete surface; The processing module (32) is further configured to obtain a two-dimensional reflectivity image of the concrete surface using a compressed sensing reconstruction model based on the full-field reflection intensity data and the coding pattern matrix; The processing module (32) is further used to perform adaptive threshold segmentation and morphological refinement on the two-dimensional reflectivity image to obtain apparent crack distribution data.

9. An electronic device, characterized in that: The electronic device comprises a processor (41), a memory (45), a user interface (43) and a network interface (44), wherein the memory (45) is used to store instructions, the user interface (43) and the network interface (44) are both used to communicate with other devices, and the processor (41) is used to execute the instructions stored in the memory (45) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images