Intelligent monitoring method and device for tunnel construction

Abstract

The invention discloses an intelligent monitoring method and device for tunnel construction, and relates to the technical field of tunnel monitoring. Three-dimensional scanning equipment arranged in a tunnel collects three-dimensional point cloud data of a tunnel construction area in real time according to target scanning parameters; and sending the three-dimensional point cloud data to a data processing center in real time through a wired transmission network or a wireless transmission module, executing data processing operation in the data processing center to obtain an analysis result, converting the analysis result into a three-dimensional visual model, marking a risk area by applying a color coding rule, and triggering a real-time alarm according to a risk prediction result. According to the method and the device, the construction adjustment instruction is output to the terminal equipment, and the monitoring operation of tunnel construction is completed, so that the tunnel in construction can be automatically monitored on the basis of no manual measurement, the accuracy of the monitoring result can be ensured, and the safety of tunnel construction can also be improved.

Description

Technical Field

[0001] This invention relates to the field of tunnel monitoring technology, and in particular to an intelligent monitoring method and device for tunnel construction. Background Technology

[0002] As a crucial component of modern transportation infrastructure, tunnel engineering's safe operation directly impacts the safety of people's lives and property. Tunnel surrounding rock monitoring technology has evolved from manual inspection to instrumental monitoring, and from single-point monitoring to distributed monitoring. Early surrounding rock monitoring relied primarily on manual visual inspection and simple measuring tools. With the development of sensor technology, single-point sensors such as stress gauges and displacement gauges were gradually introduced for fixed-point monitoring. In recent years, with the development of the Internet of Things, big data, and artificial intelligence technologies, surrounding rock monitoring has gradually moved towards intelligence, networking, and visualization. Multi-sensor fusion monitoring and intelligent early warning systems have begun to be applied in the field of tunnel engineering, providing more scientific technical means for assessing surrounding rock stability.

[0003] In existing technologies, monitoring during tunnel construction is typically conducted manually. However, in actual monitoring processes, factors such as the skill level of the workers, the accuracy of the measuring instruments, and external environmental interference can lead to safety hazards during tunnel construction. Summary of the Invention

[0004] The main objective of this invention is to propose an intelligent monitoring method and device for tunnel construction, aiming to address the problem that existing technologies typically rely on manual methods for monitoring tunnel construction. However, in actual monitoring processes, factors such as the skill level of the operators, the accuracy of the measuring instruments, and external environmental interference can lead to safety hazards during tunnel construction.

[0005] To achieve the above objectives, in a first aspect, the present invention proposes an intelligent monitoring method for tunnel construction, the method comprising: Three-dimensional point cloud data of the tunnel construction area are collected in real time by a three-dimensional scanning device deployed inside the tunnel, according to the target scanning parameters; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; The three-dimensional point cloud data is transmitted to the data processing center in real time via a wired transmission network or a wireless transmission module; Data processing jobs are performed at the data processing center to obtain analysis results; The analysis results are converted into a three-dimensional visualization model, and risk areas are marked using color coding rules. Real-time alarms are triggered based on risk prediction results, and construction adjustment instructions are output to terminal equipment to complete the monitoring operation of the tunnel construction.

[0006] In one embodiment, prior to the step of acquiring three-dimensional point cloud data of the tunnel construction area in real time according to target scanning parameters using a three-dimensional scanning device arranged inside the tunnel, the method further includes: Based on the tunnel length, a laser scanner with a scanning distance greater than a preset length is used to scan the tunnel, and the current area of ​​the overlapping region of adjacent scanning stations is set according to the tunnel cross-sectional dimensions.

[0007] In one embodiment, the wired transmission network adopts a fiber optic network pre-buried in the tunnel, and the wireless transmission module adopts a 5G communication module.

[0008] In one embodiment, the step of performing a data processing job at the data processing center to obtain analysis results includes: The received 3D point cloud data is subjected to noise removal and multi-site data stitching at the data processing center. Machine learning algorithms are used to identify surface defects in the tunnel and analyze deformation trends from the data after it has been spliced ​​together. Based on the analysis of deformation trends, a construction status assessment model for the tunnel is constructed to predict safety risks, and the analysis results are obtained.

[0009] In one embodiment, the step of identifying tunnel surface defects and analyzing deformation trends from the spliced ​​data based on machine learning algorithms includes: Based on machine learning algorithms, the deformation of the tunnel is calculated by comparing point cloud data at consecutive time points. Convolutional neural networks are used to identify cracks and voids, and the size and location coordinates of the defects are quantified to identify surface defects of the tunnel and analyze deformation trends from the data after splicing.

[0010] In one embodiment, after the step of constructing a tunnel construction status assessment model based on the analyzed deformation trend to predict safety risks and obtaining the analysis results, the method further includes: The deformation rate, defect density, and geological parameters are input into the LSTM time series prediction model, which outputs the collapse probability warning level.

[0011] In one embodiment, the color coding rule is as follows: Areas where the deformation exceeds the threshold are marked in red, areas with voids or defects are marked in yellow, and safe areas are marked in green.

[0012] In one embodiment, the construction adjustment instruction includes: When the predicted probability of collapse is greater than 10%, the command will automatically adjust the steel reinforcement density of the support structure. When the crack width is greater than 5mm, the repair grout is injected.

[0013] In one embodiment, after the step of triggering a real-time alarm based on the risk prediction result and outputting construction adjustment instructions to the terminal device to complete the monitoring operation of the tunnel construction, the method further includes: A dynamic demonstration video containing deformation heatmaps and defect location markers is generated and projected onto the tunnel construction site using AR devices.

[0014] Based on the same technical concept, in a second aspect, the present invention also proposes an intelligent monitoring device for tunnel construction, comprising: The data scanning module is used to collect three-dimensional point cloud data of the tunnel construction area in real time according to the target scanning parameters using a three-dimensional scanning device deployed inside the tunnel; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; The data transmission module is used to transmit the three-dimensional point cloud data to the data processing center in real time via a wired transmission network or a wireless transmission module. The analysis module is used to perform data processing jobs in the data processing center and obtain analysis results; The annotation module is used to convert the analysis results into a three-dimensional visualization model and apply color coding rules to annotate risk areas; The operation execution module is used to trigger real-time alarms based on risk prediction results and output construction adjustment instructions to the terminal equipment to complete the monitoring operation of the tunnel construction.

[0015] The technical solution of this invention uses a 3D scanning device deployed inside the tunnel to collect 3D point cloud data of the tunnel construction area in real time according to target scanning parameters. The 3D point cloud data is then transmitted to a data processing center in real time via a wired transmission network or a wireless transmission module. The data processing center performs data processing operations to obtain analysis results, converts the analysis results into a 3D visualization model, and applies color coding rules to mark risk areas. Based on the risk prediction results, a real-time alarm is triggered, and construction adjustment instructions are output to the terminal device, completing the monitoring operation of tunnel construction. This invention enables automated monitoring of tunnels under construction without the need for manual measurement, ensuring the accuracy of monitoring results and improving the safety of tunnel construction. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0017] Figure 1A flowchart of the surrounding rock monitoring method provided by the present invention; Figure 2 for Figure 1 The flowchart of step S300 in the example is shown below; Figure 3 The flowcharts are for some specific embodiments of the surrounding rock monitoring method exemplified by the present invention.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indicators will also change accordingly.

[0021] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the use of "and / or" or "and / or" throughout the text includes three parallel solutions. For example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0022] In tunnel surrounding rock monitoring, manual monitoring methods are limited by the varying skill levels of operators, the accuracy of measuring instruments, and external environmental interference, resulting in insufficient reliability and real-time performance of monitoring data. Specifically, subjective errors are introduced by manual operation, data drift is caused by environmental factors affecting measuring instruments in the underground environment, and the accuracy of data acquisition is reduced by external vibration interference, thus impacting the timeliness and accuracy of surrounding rock stability assessment.

[0023] This invention proposes an intelligent monitoring method and device for tunnel construction.

[0024] Please see Figures 1 to 3 For ease of understanding, this intelligent monitoring method for tunnel construction includes: S100. Real-time acquisition of three-dimensional point cloud data of the tunnel construction area using a three-dimensional scanning device deployed inside the tunnel, according to target scanning parameters; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; S200: The three-dimensional point cloud data is transmitted to the data processing center in real time via a wired transmission network or a wireless transmission module; S300. Perform data processing operations at the data processing center to obtain analysis results; S400. Convert the analysis results into a three-dimensional visualization model and apply color coding rules to mark the risk areas; S500 triggers a real-time alarm based on the risk prediction results and outputs construction adjustment instructions to the terminal equipment to complete the monitoring operation of the tunnel construction.

[0025] This application provides an intelligent monitoring method for tunnel construction, aiming to solve safety hazards caused by human technical limitations, instrument accuracy, and external environmental interference during tunnel construction monitoring. By constructing an automated and intelligent monitoring process to replace traditional manual monitoring methods, real-time, accurate, and visualized monitoring is achieved. In practical applications, target scanning parameters refer to the set of parameters controlling the operation of the 3D scanning equipment, including target scanning resolution and target scanning speed parameters. Specifically, the target scanning resolution can be set to a range of 0.1 mm to 5 mm, and the target scanning speed parameter can be set to a range of 10 m / s to 30 m / s, for example, by manual input of parameter values ​​by the operator or automatic adjustment of parameter values ​​based on the construction schedule.

[0026] The 3D point cloud data is transmitted to the data processing center in real time via a wired transmission network or a wireless transmission module. The wired transmission network can use twisted-pair cables for data transmission, while the wireless transmission module can use Wi-Fi communication technology, such as deploying network switches within the tunnel for wired transmission or installing wireless repeaters for wireless transmission. Furthermore, data processing operations performed at the data processing center can include noise filtering of the point cloud data and multi-site data alignment; for example, applying statistical outlier removal algorithms to remove noisy points, or using feature point matching algorithms for data stitching. The analysis results are then converted into a 3D visualization model, and risk areas are marked using color coding rules. These rules can be defined as high-risk areas marked in purple, medium-risk areas in orange, and low-risk areas in blue, with color brightness dynamically adjusted according to the risk level. Based on the risk prediction results, a real-time alarm is triggered when the analysis shows that the risk level exceeds a preset threshold, and construction adjustment instructions are output to the terminal equipment; for example, when the deformation exceeds a safety threshold, instructions are output to adjust tunneling parameters or add temporary support measures. This embodiment effectively avoids subjective errors and environmental interference in manual monitoring through the synergistic effect of the above-mentioned technical means, ensuring the reliability of monitoring data and the timeliness of response.

[0027] The intelligent monitoring method for tunnel construction replaces traditional manual monitoring by constructing an automated and intelligent monitoring process, effectively addressing safety hazards caused by human operation, insufficient equipment accuracy, and environmental interference. Specifically, 3D scanning equipment collects 3D point cloud data of the tunnel construction area in real time according to preset target scanning parameters, including target scanning resolution and target scanning speed, ensuring the relevance and adaptability of data acquisition. The scanning parameters are dynamically adjusted based on tunnel construction conditions to avoid blind spots or data distortion caused by fixed parameters at different geological or construction stages, thereby improving the reliability of the raw data. Furthermore, the 3D point cloud data is transmitted to the data processing center in real time via wired or wireless transmission networks, ensuring the continuity and environmental adaptability of data transmission. The transmission method is flexibly selected based on tunnel site conditions, overcoming the delays caused by manual recording and transmission, ensuring that the point cloud data arrives at the processing center stably and efficiently in complex underground environments.

[0028] Data processing is conducted in the data processing center to achieve centralized and intelligent data processing. Computing resources are used to automatically analyze point cloud data, replacing the subjectivity and inefficiency of manual interpretation and quickly generating accurate risk assessment results. The analysis results are converted into a 3D visualization model, and risk areas are marked using color coding rules to enhance the intuitive expression of risk information. The 3D visualization model generated based on the analysis results and color-coded allows risk areas to be presented intuitively, facilitating immediate understanding of the tunnel's condition by construction personnel. Real-time alarms are triggered based on risk prediction results, and construction adjustment instructions are sent to terminal equipment, establishing a closed-loop response mechanism. The automatic triggering of alarms and the push of adjustment instructions based on risk prediction results achieve seamless integration from monitoring to intervention, avoiding the lag of manual decision-making and ensuring timely handling of safety hazards.

[0029] As a preferred implementation, the 3D scanning device can specifically be a Leica RTC360 3D laser scanner, with a target scanning resolution set to 3mm and a target scanning speed set to 100,000 points / second; data transmission is carried out via wired transmission using Cat6a network cables laid within the tunnel; the data processing center consists of a server cluster equipped with multi-core CPUs and GPUs; the 3D visualization model is generated using the Three.js library, and color coding rules are applied to differentiate risk areas; when the risk prediction results reach a preset threshold, the system automatically activates an audible and visual alarm and pushes optimization suggestions to the tablet computers of construction management personnel.

[0030] Therefore, this invention eliminates subjective errors and delays caused by manual operation through automated data acquisition, real-time transmission, centralized processing, intuitive visualization, and closed-loop response mechanisms, thereby improving the real-time performance and accuracy of the monitoring process. At the same time, the dynamic matching of parameter settings for construction conditions and the stable data transmission path effectively overcome the influence of insufficient accuracy of measuring instruments and external environmental interference, ensuring that safety hazards are identified and dealt with in a timely manner, thus guaranteeing the safe progress of tunnel construction.

[0031] In this embodiment, a 3D scanning device deployed inside the tunnel collects 3D point cloud data of the tunnel construction area in real time according to the target scanning parameters. The 3D point cloud data is then transmitted to the data processing center in real time via a wired transmission network or a wireless transmission module. The data processing center performs data processing operations to obtain analysis results, converts the analysis results into a 3D visualization model, and applies color coding rules to mark risk areas. Real-time alarms are triggered based on risk prediction results, and construction adjustment instructions are output to the terminal device, completing the monitoring operation of tunnel construction. This invention enables automated monitoring of tunnels under construction without the need for manual measurement, ensuring the accuracy of monitoring results and improving the safety of tunnel construction.

[0032] In one embodiment, prior to step S100, the method further includes: S600: Using a laser scanner with a scanning distance greater than a preset length, scan the tunnel according to its length, and obtain the current area of ​​the overlapping region of adjacent scanning stations based on the tunnel cross-sectional dimensions.

[0033] Among them, a laser scanner with a scanning distance greater than a preset length means that the maximum effective scanning distance of the laser scanning device must be greater than a preset tunnel length threshold. This can be achieved using a laser scanner based on the time-of-flight principle or a laser scanner based on the triangulation principle. The purpose is to avoid coverage blind spots caused by insufficient scanning distance and ensure that a single scan can efficiently cover key areas. Obtaining the tunnel cross-sectional dimensions refers to extracting the geometric parameters of the tunnel cross-section through scanning data. This can be achieved by fitting elliptical or circular cross-sections using point cloud data. The purpose is to provide accurate geometric basis for setting overlapping areas, so that the layout parameters can adapt to tunnel structures with different cross-sectional shapes. Setting the current area of ​​the overlapping area between adjacent scanning stations refers to determining the size of the overlapping area between adjacent scanning positions. This can be achieved using an adaptive algorithm based on the density of overlapping point clouds. The purpose is to ensure sufficient overlapping point density and avoid point cloud misalignment and resource waste.

[0034] Specifically, the solution proposed in this application first selects a laser scanner with a scanning distance greater than a preset length based on the tunnel length to ensure that the equipment capability matches the tunnel scale; then scans the tunnel to obtain accurate cross-sectional dimension data, providing a geometric basis for setting overlapping areas; finally, based on this dimension, dynamically sets the overlapping area of ​​adjacent scanning stations to ensure the continuity and accuracy of point cloud data stitching, thereby effectively solving the data stitching problem caused by unreasonable overlapping area settings.

[0035] The laser scanner can use a phase-type laser scanning device to scan the tunnel in the pre-scanning stage; when obtaining the tunnel cross-sectional dimensions, the cross-sectional contour can be fitted by point cloud data processing software; when setting the overlapping area of ​​adjacent scanning stations, an adaptive algorithm based on point cloud density can be used to dynamically adjust it to ensure the continuity of subsequent point cloud data stitching.

[0036] The above solution effectively avoids the problem of gaps or misalignments in point cloud data splicing caused by unreasonable setting of overlapping areas of adjacent scanning stations, thereby improving the accuracy of tunnel deformation trend analysis and the reliability of risk prediction.

[0037] In one embodiment, the wired transmission network adopts a fiber optic network pre-buried in the tunnel, and the wireless transmission module adopts a 5G communication module.

[0038] Specifically, the fiber optic network pre-buried in the tunnel refers to the fiber optic communication lines laid during the early planning stage of tunnel construction. These lines can be implemented using single-mode or multi-mode fiber optics. The purpose is to avoid physical damage to the transmission lines caused by mechanical operations or environmental changes during construction, and to ensure data integrity by utilizing the inherent high anti-interference capability of optical fibers. Among them, the 5G communication module can be understood as a wireless transmission unit based on fifth-generation mobile communication technology. It can be implemented using communication chipsets that support the Sub-6GHz frequency band. The purpose is to effectively overcome the signal attenuation problem in the enclosed space of the tunnel by leveraging the low latency and high reliability advantages of 5G technology.

[0039] Specifically, the solution in this application establishes a complementary data transmission mechanism by setting the wired transmission network as a pre-buried fiber optic network within the tunnel and configuring the wireless transmission module as a 5G communication module. The pre-buried fiber optic network is physically deployed during the initial stages of tunnel construction, ensuring the lines are not interfered with by subsequent construction activities, while providing high bandwidth to support the high-speed transmission of massive point cloud data. The 5G communication module provides supplementary transmission capabilities in mobile monitoring scenarios or temporary areas where the pre-buried lines are difficult to cover. When the system detects a decline in signal quality in the fiber optic network due to changes in the construction environment, it automatically switches to the 5G communication module to maintain data flow continuity, thereby ensuring that the 3D point cloud data can be stably and in real-time transmitted to the data processing center, providing a reliable data foundation for subsequent analysis.

[0040] The fiber optic network pre-embedded within the tunnel is fixedly laid along the tunnel lining structure, using single-mode fiber as the transmission medium. The wireless transmission module employs a communication module integrating 5G NR technology and is installed on a mobile monitoring device. When drilling and blasting operations are carried out in the construction area, the system detects enhanced local electromagnetic interference causing fluctuations in the fiber optic signal and automatically switches the data transmission path to the 5G communication module, ensuring uninterrupted transmission of point cloud data to the data processing center.

[0041] In this embodiment, through an example, data loss caused by electromagnetic interference, transmission delay caused by signal attenuation, and physical damage to temporary lines caused by construction activities in the tunnel construction environment are effectively avoided, ensuring the stable and reliable transmission of three-dimensional point cloud data, thereby guaranteeing the timely response of the risk warning system and the accurate output of construction adjustment instructions.

[0042] In one embodiment, step S300 includes: S310. The received three-dimensional point cloud data is subjected to noise removal and multi-site data stitching at the data processing center. S320. Based on machine learning algorithms, identify surface defects in the tunnel and analyze deformation trends from the data after splicing. S330. Based on the analyzed deformation trend, construct the tunnel's construction status assessment model to predict safety risks and obtain the analysis results.

[0043] In practical applications, noise removal refers to eliminating false data points introduced by environmental interference and equipment errors during the scanning process. This can be achieved using statistical outlier removal algorithms or radius filtering techniques, aiming to filter out random noise to ensure data purity and lay a reliable foundation for subsequent analysis. Multi-site data stitching refers to integrating point cloud data collected from different locations into a continuous and complete tunnel model. This can be achieved using iterative nearest-point algorithms or feature-point-based registration methods, aiming to overcome the monitoring blind spots caused by the limited coverage of a single scan and making overall deformation analysis of long tunnels possible. Specifically, machine learning algorithms utilize the automatic learning capabilities of point cloud data for pattern recognition, which can be achieved using support vector machines or random forest classifiers. The aim is to efficiently detect surface defects and quantify deformation characteristics, avoiding subjective oversights from human experience. Constructing a construction status assessment model refers to establishing a mathematical framework that transforms deformation data into risk indicators. This can be achieved using decision trees or Bayesian networks, aiming to achieve a scientific closed loop from data to early warning, making safety risk prediction more forward-looking and targeted.

[0044] Specifically, this embodiment first removes noise from the point cloud data to effectively improve data quality and ensure the reliability of input for subsequent processing. Based on this, multi-site data stitching integrates scattered scan data into a unified model, eliminating spatial discontinuity and providing a continuous data foundation for overall analysis. Furthermore, based on machine learning algorithms, surface defects are simultaneously identified and deformation trends are analyzed from the stitched data. The algorithm's automatic processing capability for large amounts of point cloud data accurately quantifies crack size and location coordinates, while simultaneously calculating the deformation rate at continuous time points. Finally, a construction status assessment model is constructed based on the deformation trends, combining deformation data with geological parameters to transform it into quantifiable risk indicators, forming a complete technical chain from data acquisition to risk warning, thereby ensuring the scientific validity and accuracy of safety risk prediction.

[0045] The data processing center is equipped with high-performance servers as processing units. The noise removal module uses a statistical outlier removal algorithm to process point cloud data, effectively filtering out false points caused by environmental interference. The multi-site data stitching module integrates data from different scanning sites into a continuous tunnel model through feature point matching technology. The machine learning module deploys a random forest classifier to automatically identify cracks and voids from the stitched data and calculate the deformation at adjacent time points. The construction status assessment module uses a decision tree model to input deformation trend data and output the safety risk level, realizing automated risk prediction.

[0046] In this embodiment, through an example, noise interference in point cloud data can be effectively eliminated, data from multiple sites can be accurately stitched together to form a complete tunnel model, the accuracy of tunnel surface defect identification is significantly improved, the deformation trend analysis process is more objective and reliable, and the accuracy and timeliness of safety risk prediction are effectively guaranteed, thereby providing a scientific basis for tunnel construction monitoring.

[0047] In one embodiment, step S320 includes: Based on machine learning algorithms, the deformation of the tunnel is calculated by comparing point cloud data at consecutive time points. Convolutional neural networks are used to identify cracks and voids, and the size and location coordinates of the defects are quantified to identify surface defects of the tunnel and analyze deformation trends from the data after splicing.

[0048] Among these methods, calculating tunnel deformation by comparing point cloud data at consecutive time points refers to quantifying the displacement changes of the tunnel structure by analyzing the differences in point cloud data at different time points. This can be achieved using point cloud registration algorithms or time series differencing techniques, aiming to capture the dynamic evolution of tunnel deformation and avoid the static limitations of a single scan. Identifying cracks and cavities using convolutional neural networks refers to automatically extracting spatial features from point cloud data using deep learning models to detect abnormal areas. This can be achieved using 3D convolutional neural networks or dedicated point cloud network architectures, aiming to improve the detection rate and anti-interference ability of small defects in complex environments. Quantifying defect size and location coordinates refers to converting the identification results into specific geometric parameters and spatial coordinates. This can be achieved using point cloud clustering analysis or coordinate mapping methods, aiming to provide quantifiable defect assessment data and lay the foundation for subsequent risk analysis.

[0049] Specifically, in this embodiment, the tunnel deformation is first calculated by comparing point cloud data at consecutive time points to establish a time-series deformation data stream, thereby revealing the gradual change pattern of the tunnel structure. Next, a convolutional neural network is used to extract spatial features from the point cloud data, identifying cracks and voids, and leveraging the feature learning capabilities of deep learning to reduce environmental noise interference. Finally, the defect size and location coordinates are quantified, transforming the qualitative identification results into precise numerical parameters. This combination of dynamic analysis in the time dimension and intelligent identification in the spatial dimension enables deformation trend analysis to reflect a continuous evolution process, provides robust defect identification, and offers reliable input for risk assessment through quantified output, forming a closed-loop data processing chain.

[0050] In the data processing center, point cloud registration algorithms are used to process point cloud data from adjacent scanning cycles to calculate the displacement vector of the tunnel wall. Convolutional neural networks use an encoder-decoder structure to extract features from the point cloud data and automatically identify the contour regions of cracks and cavities. The defect size is calculated through the geometric properties of the point cloud clustering region, and the position coordinates are mapped based on the tunnel global coordinate system, outputting a quantized result containing the defect bounding box and center coordinates.

[0051] In this embodiment, the dynamic changes in tunnel deformation can be effectively captured through examples, noise interference in defect identification can be reduced, and the size and location of defects can be accurately quantified, thereby improving the accuracy and real-time performance of safety risk prediction.

[0052] In one embodiment, after step S330, the method further includes: S340. Input deformation rate, defect density and geological parameters into LSTM time series prediction model, and output collapse probability warning level.

[0053] Among them, deformation rate refers to the deformation of the tunnel structure per unit time, which can be calculated by differential calculation of point cloud data at continuous time points, with the aim of capturing the dynamic trend of deformation; defect density refers to the distribution density of defects on the tunnel surface per unit area, which can be calculated by statistically analyzing the number density of cracks and cavities using image segmentation algorithms, with the aim of quantifying the severity of defects; geological parameters refer to physical quantities reflecting the characteristics of the surrounding rock, which can include parameters such as rock type and degree of joint development, with the aim of providing background information on geological stability; LSTM time series prediction model is a long short-term memory neural network, which can be implemented using a deep learning framework, with the aim of processing time series data and learning historical dependencies; collapse probability warning level refers to the output that classifies the prediction results into different risk levels, which can be implemented using discrete level labels, with the aim of providing actionable warning signals.

[0054] Specifically, in this embodiment, deformation rate, defect density, and geological parameters are input into the LSTM time-series prediction model as a set of temporal features. This model models the long-term dependencies of historical data, enabling continuous simulation of tunnel deformation and defect evolution. Deformation rate, as the core input feature, can capture the dynamic trend of structural changes in real time and identify critical points of accelerated or abrupt deformation. Defect density quantifies the distribution of surface defects and, combined with geological parameters, reflects the overall stability of the surrounding rock; these three factors together constitute a comprehensive health assessment index. The LSTM time-series prediction model transforms discrete monitoring data into continuous risk probability output, outputting collapse probability warning levels. This allows the system to trigger a response mechanism based on clearly defined level thresholds, thereby upgrading the static analysis method to a dynamic quantitative assessment system.

[0055] At the data processing center, the deformation rate data, defect density statistics, and geological parameter database generated in the previous steps are input into the LSTM time series prediction model. This model is implemented using a standard deep learning architecture and outputs a collapse probability warning level. For example, when the model identifies that the deformation rate is continuously rising and the defect density exceeds the benchmark value, it automatically raises the warning level to a high level and triggers real-time alarms and construction adjustment instructions.

[0056] In this embodiment, a continuous quantitative assessment of tunnel collapse risk is achieved through an example, which significantly improves the timeliness of risk warning and the objectivity of decision-making basis, and effectively solves the problems of response lag and insufficient accuracy in dynamic risk prediction.

[0057] In one embodiment, the color coding rule is as follows: Areas where the deformation exceeds the threshold are marked in red, areas with voids or defects are marked in yellow, and safe areas are marked in green.

[0058] Among them, the deformation exceeding the threshold area refers to the area on the tunnel surface where the deformation exceeds the preset safety threshold. It can be achieved by using a fixed threshold range based on historical monitoring data or engineering specifications, with the aim of quickly identifying high-risk areas with excessive structural deformation. The area with voids and defects can be understood as the area on the tunnel surface with defects such as cracks or voids. It can be achieved by using the defect location coordinates detected by point cloud data image recognition technology, with the aim of distinguishing the types of defects that need attention but are not urgent. The safe area specifically refers to the area where both the deformation and defects meet the safety standards. It can be achieved by using the default threshold range, with the aim of providing an intuitive safety benchmark reference.

[0059] Specifically, in this embodiment, by mapping the risk prediction results to a standardized color coding mechanism, risk areas are dynamically marked in the three-dimensional visualization model, so that areas with deformation exceeding the threshold, void defect areas, and safe areas are presented in red, yellow, and green, respectively, thereby forming a clear visual hierarchy, enhancing information transmission efficiency, and improving the decision-making speed of construction personnel in complex environments.

[0060] In the 3D visualization model, the system automatically colors the tunnel surface based on the analysis results. When the deformation of a certain area exceeds the threshold, the area is marked in red; when a void defect is detected, the area is marked in yellow; and the remaining areas are marked in green, so that construction personnel can intuitively identify the risk level.

[0061] In this embodiment, the example demonstrates how construction workers can quickly and accurately distinguish between different risk types, respond promptly to high-risk areas, and effectively reduce safety hazards during tunnel construction.

[0062] In one embodiment, the construction adjustment instruction includes: When the predicted probability of collapse is greater than 10%, the command will automatically adjust the steel reinforcement density of the support structure. When the crack width is greater than 5mm, the repair grout is injected.

[0063] In practical applications, the predicted collapse probability refers to a risk quantification index output by a construction status assessment model, which can be achieved using time-series prediction algorithms or statistical models. Its purpose is to objectively assess the possibility of tunnel structural instability. The 10% threshold refers to a set risk trigger point, which can be dynamically adjusted according to engineering safety specifications and geological parameters. Its purpose is to initiate intervention measures in the early stages of risk accumulation. Automatic adjustment of the steel reinforcement density of the support structure refers to a mechanical adjustment process that requires no manual intervention. It can be achieved using hydraulically driven steel reinforcement positioning devices or automated construction equipment. Its purpose is to instantly enhance the strength of the surrounding rock support. Crack width refers to the dimensional measurement result of defects on the tunnel surface, which can be obtained through point cloud data analysis or image recognition technology. Its purpose is to accurately quantify the severity of defects. The 5mm threshold refers to the critical size for crack repair, which can be determined based on material deformation characteristics and construction standards. Its purpose is to prevent small defects from evolving into structural damage. Injection of repair grout refers to an automatically executed repair operation, which can be achieved using a pumping system in conjunction with pre-embedded grouting pipelines. Its purpose is to promptly fill gaps to block the crack development path.

[0064] Specifically, in this embodiment, by using the risk prediction results as input signals, when the collapse probability exceeds a preset threshold of 10%, the system automatically generates instructions to drive the support equipment to adjust the steel reinforcement density, thereby immediately enhancing the support strength at the critical point of structural safety. Simultaneously, when the crack width exceeds 5mm, the system triggers an instruction to initiate grout injection, precisely repairing surface defects. This automatic response mechanism based on quantified thresholds establishes a closed-loop correlation between risk prediction data and engineering actions, ensuring that the process from monitoring and analysis to construction intervention has a clear execution sequence and triggering conditions, effectively avoiding delays and biases inherent in manual assessment.

[0065] When the collapse probability output by the data processing center is 12%, the system automatically sends instructions to the support construction equipment, which increases the steel reinforcement density by adjusting the steel reinforcement spacing. At the same time, when the crack width measurement value marked in the three-dimensional visualization model is 6mm, the system controls the grout pump to start and inject repair grout through the preset grouting holes in the tunnel to repair the defect.

[0066] In this embodiment, the automated generation and execution of construction adjustment instructions are realized through an example, which significantly reduces risk response time, eliminates the subjectivity of human decision-making, and thus prevents structural instability and defect expansion in a timely manner during tunnel construction, effectively improving the controllability of construction safety.

[0067] In one embodiment, after step S500, the method further includes: S700 generates a dynamic demonstration video containing deformation heatmaps and defect location markers, and projects it onto the tunnel construction site via AR devices.

[0068] Specifically, deformation heatmaps refer to the visual representation of deformation distribution calculated based on historical point cloud data. This can be achieved using color gradient mapping, such as transitioning from blue to red using a color spectrum to represent the increase in deformation. Its purpose is to intuitively reflect the spatiotemporal evolution trend of tunnel surface deformation. Defect location markers are markers that precisely identify the location coordinates and dimensions of cracks and cavities. These can be achieved using vector graphics in a 3D coordinate system or highlighted areas in a point cloud. Their purpose is to clearly present the spatial distribution of risk areas. Dynamic demonstration videos transform static 3D visualization models into continuously changing dynamic forms. This can be achieved using real-time rendering engines or video encoding technology. Its purpose is to dynamically display the evolution process of risk areas. AR devices refer to augmented reality devices, which can be implemented using head-mounted displays or mobile terminals in conjunction with AR application frameworks. Their purpose is to seamlessly overlay virtual content onto the real construction environment.

[0069] Specifically, in this embodiment, a dynamic demonstration video containing deformation heatmaps and defect location markers is first generated based on historical point cloud data. This video presents the dynamic changes in tunnel deformation trends and defect locations through a continuous frame sequence. Subsequently, the video content is superimposed onto the actual field of view at the tunnel construction site in real time using an AR device, ensuring precise alignment between the virtual risk markers and the real tunnel structure. Because the dynamic demonstration video integrates the spatiotemporal evolution characteristics of the deformation heatmap and the spatial coordinate information of the defect location markers, construction personnel can directly perceive the dynamic characteristics of the risk area in the real working environment without switching viewing angles during the AR device projection process. This avoids the information fragmentation problem of traditional screen displays and ensures that risk information is organically integrated with the construction operation environment.

[0070] AR devices can use head-mounted augmented reality displays. These devices capture real-world scenes of tunnel construction sites through built-in sensors and overlay dynamic demonstration video content in real time. The dynamic demonstration video can be generated by a 3D rendering engine called by a data processing center. The deformation heat map uses a gradient color spectrum to map the deformation distribution, and the defect location markers mark the precise locations of cracks and cavities in the form of 3D coordinate points.

[0071] In this embodiment, the example enables construction personnel to obtain the specific location and deformation dynamics of the risk area in real time and intuitively at the tunnel construction site, effectively reducing information transmission delays and on-site misunderstandings, thereby improving the timeliness of construction safety response and operational accuracy.

[0072] Based on the same technical concept, in a second aspect, the present invention also proposes an intelligent monitoring device for tunnel construction, comprising: The data scanning module is used to collect three-dimensional point cloud data of the tunnel construction area in real time according to the target scanning parameters using a three-dimensional scanning device deployed inside the tunnel; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; The data transmission module is used to transmit the three-dimensional point cloud data to the data processing center in real time via a wired transmission network or a wireless transmission module. The analysis module is used to perform data processing jobs in the data processing center and obtain analysis results; The annotation module is used to convert the analysis results into a three-dimensional visualization model and apply color coding rules to annotate risk areas; The operation execution module is used to trigger real-time alarms based on risk prediction results and output construction adjustment instructions to the terminal equipment to complete the monitoring operation of the tunnel construction.

[0073] By combining real-time 3D point cloud data acquisition with automated analysis and processing in a closed-loop manner, this approach overcomes the problems of insufficient reliability and real-time performance of monitoring data caused by differences in operator skill levels, limitations in the accuracy of measuring instruments, and external environmental interference in manual monitoring. This achieves the goal of improving the automation level of tunnel construction monitoring and ensuring the timeliness of risk identification and the accuracy of response. Specifically, the data scanning module achieves high-precision geometric information capture based on target scanning parameters, avoiding omissions from manual visual inspection; the data transmission module ensures that point cloud data arrives at the processing center instantly through a stable transmission path, reducing delays caused by environmental interference; the analysis module uses automated algorithm processing to replace manual interpretation, improving the objectivity of risk assessment; the annotation module intuitively presents risk areas through color coding rules, eliminating ambiguity in manual interpretation; and the operation execution module automatically triggers alarms and outputs adjustment instructions based on risk prediction results, forming a seamless connection mechanism from monitoring to intervention. Therefore, this solution effectively avoids the subjective errors and response delays of manual operation, ensuring real-time early warning and timely elimination of safety hazards in tunnel construction.

[0074] The above description is merely an exemplary embodiment of the present invention and does not limit the scope of the present invention. Any equivalent structural transformations made based on the technical concept of the present invention and the contents of the specification and drawings of the present invention, or direct / indirect applications in other related technical fields, are included within the protection scope of the present invention.

Claims

1. A method for intelligent monitoring of tunnel construction, characterized in that, The method includes: Three-dimensional point cloud data of the tunnel construction area are collected in real time by a three-dimensional scanning device deployed inside the tunnel, according to the target scanning parameters; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; The three-dimensional point cloud data is transmitted to the data processing center in real time via a wired transmission network or a wireless transmission module; Data processing jobs are performed at the data processing center to obtain analysis results; The analysis results are converted into a three-dimensional visualization model, and risk areas are marked using color coding rules. Real-time alarms are triggered based on risk prediction results, and construction adjustment instructions are output to terminal equipment to complete the monitoring operation of the tunnel construction.

2. The intelligent monitoring method for tunnel construction as described in claim 1, characterized in that, Before the step of acquiring three-dimensional point cloud data of the tunnel construction area in real time according to the target scanning parameters using a three-dimensional scanning device arranged inside the tunnel, the method further includes: Based on the tunnel length, a laser scanner with a scanning distance greater than a preset length is used to scan the tunnel, and the current area of ​​the overlapping region of adjacent scanning stations is set according to the tunnel cross-sectional dimensions.

3. The intelligent monitoring method for tunnel construction as described in claim 2, characterized in that, The wired transmission network uses a fiber optic network pre-buried in the tunnel, and the wireless transmission module uses a 5G communication module.

4. The intelligent monitoring method for tunnel construction as described in claim 1, characterized in that, The steps of performing data processing jobs at the data processing center to obtain analysis results include: The received 3D point cloud data is subjected to noise removal and multi-site data stitching at the data processing center. Machine learning algorithms are used to identify surface defects in the tunnel and analyze deformation trends from the data after it has been spliced ​​together. Based on the analysis of deformation trends, a construction status assessment model for the tunnel is constructed to predict safety risks, and the analysis results are obtained.

5. The intelligent monitoring method for tunnel construction as described in claim 4, characterized in that, The steps of identifying tunnel surface defects and analyzing deformation trends from the spliced ​​data based on machine learning algorithms include: Based on machine learning algorithms, the deformation of the tunnel is calculated by comparing point cloud data at consecutive time points. Convolutional neural networks are used to identify cracks and voids, and the size and location coordinates of the defects are quantified to identify surface defects of the tunnel and analyze deformation trends from the data after splicing.

6. The intelligent monitoring method for tunnel construction as described in claim 5, characterized in that, After the step of constructing the tunnel's construction status assessment model based on the analyzed deformation trend to predict safety risks and obtaining the analysis results, the method further includes: The deformation rate, defect density, and geological parameters are input into the LSTM time series prediction model, which outputs the collapse probability warning level.

7. The intelligent monitoring method for tunnel construction as described in any one of claims 1 to 6, characterized in that, The color coding rule is as follows: Areas where the deformation exceeds the threshold are marked in red, areas with voids or defects are marked in yellow, and safe areas are marked in green.

8. The intelligent monitoring method for tunnel construction as described in any one of claims 1 to 6, characterized in that, The construction adjustment instructions include: When the predicted probability of collapse is greater than 10%, the command will automatically adjust the steel reinforcement density of the support structure. When the crack width is greater than 5mm, the repair grout is injected.

9. The intelligent monitoring method for tunnel construction as described in any one of claims 1 to 6, characterized in that, After the steps of triggering a real-time alarm based on the risk prediction results and outputting construction adjustment instructions to the terminal device to complete the monitoring operation of the tunnel construction, the method further includes: A dynamic demonstration video containing deformation heatmaps and defect location markers is generated and projected onto the tunnel construction site using AR devices.

10. An intelligent monitoring device for tunnel construction, characterized in that, include: The data scanning module is used to collect three-dimensional point cloud data of the tunnel construction area in real time according to the target scanning parameters using a three-dimensional scanning device deployed inside the tunnel; wherein, the target scanning parameters include the target scanning resolution and target scanning speed parameters of the three-dimensional scanning device; The data transmission module is used to transmit the three-dimensional point cloud data to the data processing center in real time via a wired transmission network or a wireless transmission module. The analysis module is used to perform data processing jobs in the data processing center and obtain analysis results; The annotation module is used to convert the analysis results into a three-dimensional visualization model and apply color coding rules to annotate risk areas; The operation execution module is used to trigger real-time alarms based on risk prediction results and output construction adjustment instructions to the terminal equipment to complete the monitoring operation of the tunnel construction.

Images