A ground vacuum preloading construction supervision system based on image recognition technology

Abstract

The application provides a kind of ground vacuum preloading construction supervision system based on image recognition technology, including engineering data acquisition equipment, construction positioning equipment, auxiliary construction equipment, geotextile and sealing membrane laying construction monitoring module, drainage board construction monitoring module, vacuum booster system installation construction monitoring module, construction data recording module, and contains each construction equipment arrangement and marking method and coordinate reference identification algorithm and the support algorithm process required by each module.The application carries out color marking on traditional construction materials and elements, and monitors the compliance of each construction link using AI image recognition technology, to achieve automatic supervision of construction, save the workload required for engineering management, and guide on-site construction personnel to complete each link construction according to requirements, improve the safety and economic benefits of dam foundation treatment construction under the condition of reducing personnel investment and labor cost.

Description

Technical Field

[0001] This invention belongs to the field of intelligent construction management, specifically relating to a foundation vacuum preloading construction monitoring system based on image recognition technology. Background Technology

[0002] Vacuum preloading is a crucial step in ensuring the stability and safety of dam structures. However, the effectiveness of processes such as geotextile laying, drainage board installation, vacuum sealing, and probe installation significantly impacts the reliability and quality of the dam foundation treatment. Furthermore, some aspects are concealed works, making inspection and supervision difficult after completion, potentially jeopardizing the overall safety of the project. Therefore, the industry needs a simple, efficient, and automated vacuum preloading construction monitoring system to provide more detailed real-time monitoring of construction processes. Summary of the Invention

[0003] To address the problems existing in the prior art, this invention provides a foundation vacuum preloading construction monitoring system based on image recognition technology to solve the problems of the prior art.

[0004] This invention is implemented as follows: a foundation vacuum preloading construction monitoring system based on image recognition technology, including an engineering data acquisition device capable of data exchange, a construction positioning device, auxiliary construction equipment, a geotextile and sealing membrane laying construction monitoring module and algorithm module, a drainage board insertion construction monitoring module and algorithm module, a vacuum pressurization system installation construction monitoring module and algorithm module, a construction data recording module, and includes modules for the layout and marking of various construction equipment, as well as a coordinate reference recognition algorithm module, a high-definition camera group layout algorithm module, a marker layout algorithm module, and a basic image recognition algorithm group module that can be called.

[0005] The engineering information acquisition equipment group includes a high-definition camera group, a camera lifting rod, and a 3D laser scanner, which are used to acquire high-definition video images and engineering model information on site. The acquired image information will be used by various engineering construction supervision modules for image recognition and data analysis of the construction process; the acquired model information will be used to connect various engineering data for data querying in corresponding functional modules.

[0006] The high-definition camera group layout algorithm module is used to obtain the layout method of the high-definition camera group in the vacuum preloading construction site. The algorithm will combine the camera performance parameters to use the minimum number of cameras while ensuring clear and reliable image data.

[0007] The construction positioning equipment includes a construction area positioning network composed of positioning benchmarks, positioning base stations, and positioning tags. It is used to assist the AI ​​image recognition function in recognizing the construction process images of each vacuum preloading process, so as to increase the image recognition accuracy, reduce the image recognition difficulty, improve the running efficiency of each module program, and reduce energy consumption.

[0008] The benchmark placement algorithm module is used to specify and constrain the placement interval and number of positioning benchmarks required for the project, and to provide a highly recognizable benchmark positioning network for the coordinate reference recognition algorithm, so that the coordinate reference recognition algorithm can obtain the positioning coordinates of each construction material and equipment through multi-angle image recognition reference calculation.

[0009] The construction equipment and material marking module uses different colored construction coatings that are non-toxic, harmless, waterproof, and corrosion-resistant to mark the key equipment and materials in different construction stages of vacuum preloading in a corresponding way, so as to improve the image recognition efficiency of the module and algorithm and reduce the recognition difficulty and cost of the algorithm.

[0010] The basic image recognition algorithm is a set of basic image recognition algorithms that can be called from outside by various functional modules in the system platform. The basic image recognition algorithm module has various basic image analysis and recognition functions, identifies various equipment and materials related to vacuum preloading construction from images collected on site, and assists each module in realizing the required basic image recognition functions and obtaining relevant data.

[0011] The coordinate reference recognition algorithm module uses the known coordinate positioning benchmarks set up at the construction site as positioning references. Then, through high-definition images taken from multiple angles, it compares and calculates the spatial position coordinates of each construction equipment and material in the images to obtain the calculated coordinate values ​​of each equipment and material. This provides a basis for each module to judge the compliance of the installation and construction of the on-site process.

[0012] The auxiliary construction equipment includes a smart safety helmet and a laser positioning device. Through voice and laser positioning markers, it assists on-site personnel in carrying out each process of vacuum preloading, while collecting and feeding back on-site construction information. The collected data serves as the basis for image recognition and program command issuance by various functional modules, and also serves as information for construction process records and receipts, which are associated with the corresponding three-dimensional engineering model.

[0013] The geotextile and sealing membrane laying construction monitoring module provides an interactive program interface and calls geotextile and sealing membrane laying construction monitoring algorithm, coordinate reference recognition algorithm, foundation image recognition algorithm group and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the geotextile and sealing membrane laying process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervision personnel for on-site processing.

[0014] The geotextile and sealing membrane laying construction monitoring algorithm module, according to the requirements and commands of the corresponding module, calls the algorithm and resources to identify and detect the compliance of various construction equipment and materials involved in the geotextile and sealing membrane laying construction process, and feeds back and records the corresponding information.

[0015] The drainage board installation monitoring module provides an interactive program interface and calls the drainage board installation monitoring algorithm, coordinate reference recognition algorithm, basic image recognition algorithm group and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the drainage board installation process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervisors for on-site processing.

[0016] The drainage board installation construction monitoring algorithm module, according to the corresponding module requirements and command call algorithm and resource pairs, identifies and detects the compliance of various construction equipment and materials involved in the drainage board installation construction process through pixel and image judgment and coordinate calculation, and feeds back and records the corresponding information.

[0017] The vacuum booster system installation and construction monitoring module provides an interactive program interface and calls upon the vacuum booster system installation and construction monitoring algorithm, coordinate reference recognition algorithm, basic image recognition algorithm group, and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the vacuum booster system installation and construction procedures in the vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervisors for on-site handling.

[0018] The vacuum booster system installation and construction monitoring algorithm module, according to the corresponding module requirements and command call algorithm and resource pairs, identifies and detects the compliance of various construction equipment and materials involved in the vacuum booster system installation and construction process through pixel and image judgment and coordinate calculation methods, and feeds back and records the corresponding information.

[0019] The construction data recording module includes a data reading submodule, a data linking submodule, a data display submodule, and a report generation submodule. It is used to collect and display the construction monitoring and process data records of vacuum preloading construction by various construction monitoring modules and algorithms, so as to facilitate the project operation and maintenance management personnel to manage the project quality and progress and query historical data.

[0020] Furthermore, in the engineering data acquisition equipment, the high-definition camera group consists of a monitoring network composed of multiple cameras arranged at the vacuum preloading construction site, and positioning tags are installed on each camera; it can acquire high-definition images of each stage of vacuum preloading construction for subsequent identification and judgment by various construction monitoring functional modules;

[0021] The camera lifting rod is used to support each high-definition camera. The rod is equipped with a transmission device and a signal receiving device, which can adjust the height of the camera according to the requirements of the coordinate reference recognition algorithm and collect data information from the required height. The maximum height of the camera lifting rod is related to the size of the vacuum preloading construction site and the accuracy of the camera.

[0022] The 3D scanner is deployed at the vacuum preloading construction site or carried by the construction and installation personnel. It receives feedback signals through laser ranging technology, converts the actual spatial information into a 3D model, and obtains the on-site model of the vacuum preloading construction site and equipment. This model is used to carry the data of subsequent construction monitoring modules and provides an intuitive data query and operation command release interface.

[0023] In the construction data recording module, the data reading submodule reads and organizes the on-site construction data obtained by each construction monitoring module according to the vacuum preloading construction sequence, and classifies and organizes various data during the vacuum preloading construction process to facilitate data association and display.

[0024] The associated display submodule is used to associate the on-site data collected from each engineering process and the construction results data after being sorted and classified by the data reading submodule with the corresponding construction BIM models obtained by on-site scanning and design modeling, and to provide a visual program interface for three-dimensional visualization display of various construction images, models and record data.

[0025] The report generation submodule is used to extract data from the information content sorted and categorized by the data reading submodule, organize the construction data and records of each vacuum preloading process according to the construction procedures, key points of construction monitoring, and key points of the overall construction plan, and automatically generate a vacuum preloading construction monitoring work report for this project based on the report template file framework.

[0026] Furthermore, the algorithm flow of the high-definition camera group deployment algorithm module is as follows:

[0027] S1. Read the outer contour line of the area to be vacuum preloaded from the corresponding construction details or BIM model;

[0028] S2. Calculate the number and spacing of cameras along the outer contour line using the following formula:

[0029]

[0030] In the formula The cameras are spaced out along the outer contour line of the construction site. This is the shape adjustment coefficient for the vacuum preloading area; This is the maximum pixel count of the camera; This refers to the horizontal resolution coefficient. This refers to the lens's light-gathering size; This is the minimum illumination resolution coefficient for the camera;

[0031]

[0032] In the formula The spacing of cameras within the construction area; The cameras are spaced out along the outer contour line of the construction site. This refers to the camera's wide-angle coefficient.

[0033] S3. Based on the camera arrangement intervals along the outer contour line of the construction site and the camera arrangement intervals within the construction area, the construction site is divided into grids to generate a high-definition camera group layout plan.

[0034] S4. Generate a camera layout plan and count the number of cameras and camera lifting poles for use in subsequent module processes and algorithms.

[0035] Furthermore, in the engineering data acquisition equipment, the positioning pole consists of a pole body, a pole head, and a positioning tag fixing component. It is installed in the vacuum preloading construction area according to the pole layout algorithm requirements. The pole head is painted with a conspicuous color that is clearly different from the construction ground, geotextile, vacuum membrane, and various construction equipment and personnel to facilitate camera recognition and algorithm program recognition and positioning of image content. At the same time, a positioning tag that can communicate with the positioning base station is installed to provide basic parameters for each module to call the coordinate reference recognition algorithm for calculation.

[0036] The positioning base station uses UWB positioning technology to locate the coordinates of the positioning tags in the vacuum preloading construction equipment and positioning poles. The positioning base station is deployed in the vacuum preloading construction site. It receives tag signals through the wireless network in the project area and calculates the distance and angle of each device and the positioning tag device in the positioning pole according to the arrival time, arrival time difference and arrival angle algorithm, and obtains the corresponding coordinate information.

[0037] The positioning tags are installed in the drainage board insertion machine, the positioning pole vacuum preloading construction equipment and auxiliary materials. They can respond to the wireless signals emitted by the positioning base station and obtain the spatial coordinates of each positioning tag and the corresponding equipment through the information feedback time difference. They are used for the coordinate positioning of each equipment and material and the display of the model in three-dimensional space.

[0038] The algorithm flow of the benchmark placement algorithm module is as follows:

[0039] S1. Read or input actual parameter data of the vacuum preloading construction site and materials of the project, as well as data from engineering drawings and models;

[0040] S2. Calculate the installation interval of the positioning markers. The calculation formula is as follows:

[0041]

[0042] In the formula The spacing for positioning marker installation; Adjust the coefficients according to the required positioning coordinate accuracy; The area of ​​the vacuum preloading construction site; This refers to the minimum area of ​​a single piece of geotextile and sealing membrane used; This refers to the number of drain pipe connectors; This is the sum of the number of horizontal and vertical pipes in the pressurization pipeline system; Boost system accuracy calibration coefficient;

[0043] S3. Based on the calculated installation interval of the positioning markers, the vacuum preloading construction site is divided into grids, and a positioning marker installation layout diagram is generated.

[0044] S4. The generated results guide the construction site in completing the installation and layout of each positioning marker.

[0045] Furthermore, in the construction equipment and material marking module, the geotextile and sealing membrane marking method is used to color-mark the woven geotextile, non-woven geotextile, and vacuum sealing membrane used in vacuum preloading construction; the selected paint must have a significant color difference from the laying layer where the geotextile or sealing membrane is located and the soil in the construction area, so as to improve the efficiency and accuracy of automatic image recognition.

[0046] The drainage board marking method is used to mark the drainage boards required for the vacuum preloading construction process by coloring them. The drainage channels on the outside and inside of the drainage board need to be colored differently, and there should be obvious differences between the outside and inside of the drainage board and the woven geotextile laid in the current process. At the same time, the length scale and numbers will be marked on the drainage board to automatically identify the insertion depth of each drainage board.

[0047] The vacuum pipeline marking method is used to color-mark the vacuum system pipelines during vacuum preloading construction. The inner and outer sides of the vacuum pipelines and the connectors with the drainage board are colored differently, and it is ensured that the color of each construction material is significantly different from the woven geotextile laid in the current process. This enables the use of algorithms to identify misaligned vacuum pipeline layouts, misaligned drainage board connectors, and various vacuum system layout construction problems.

[0048] The vacuum probe and pressure gauge marking method is used to color-mark the vacuum probe and pressure gauge monitoring equipment required in vacuum preloading construction; the paint selected for the vacuum probe and pressure gauge should have a more obvious color difference from the laying layer where the geotextile or sealing membrane is located and the soil in the construction area; so that the relevant image recognition algorithm can more efficiently identify the relevant monitoring equipment from the image, and facilitate the recognition and recording by the algorithm and program functions;

[0049] The marking method for the positioning pole and settlement meter is used to color-mark the head of the positioning pole and the outer shell of the settlement meter; when the positioning pole is installed, it is inserted into the vacuum preloading construction site and the exposed head is colored.

[0050] Furthermore, the algorithm flow of the coordinate reference recognition algorithm module is as follows:

[0051] S1. Receive commands from various functional modules requesting coordinates of construction equipment and materials;

[0052] S2. Call the basic image recognition algorithm to analyze and identify the target pixel color and the presented pattern, and initially search for the target construction equipment and materials that need to be located in the image, as well as the nearest 1-3 sets of positioning benchmarks within the target range;

[0053] S3. By analyzing the patterns captured by each camera, sort them according to the target pixel color and the size of the pattern in the captured image, and sort each camera in reverse order according to its distance from the target device and material.

[0054] S4. Set two camera height levels, high and low. Based on the order of camera distance from the target, alternately set the camera height and adjust the height of each camera to collect image information of the target construction equipment and materials from different heights and angles, providing a basis for subsequent coordinate calculations.

[0055] S5. Using the current camera coordinates and the coordinates of the nearest positioning benchmark to the target construction equipment and materials in the image, calculate the coordinates of the target construction equipment and materials in the images captured by each camera. The calculation formula is as follows:

[0056]

[0057]

[0058]

[0059] Y, Z axis coordinates The calculation formula is the same;

[0060] In the formula To obtain the initial X-axis coordinate values ​​of the target construction equipment and materials based on the images captured by each camera; The X-axis coordinate of the nearest positioning benchmark in the image captured by the current camera that is clearly observable and identifiable of the target construction equipment and materials; This refers to the difference in X-axis coordinates between the nearest positioning benchmark and the target construction equipment and materials, calculated based on the images captured by the current camera. This is the camera adjustment factor; Adjust the weighting coefficients for each camera position; The maximum angle between the target plane and the plane containing the observed target; This is the distance between the current camera and the nearest positioning marker for the target construction equipment and materials; Adjustment parameters are calculated for the weights; The X-axis coordinates of the target construction equipment and materials are obtained using the algorithm; N is the total number of lenses used to capture images of this target construction equipment and materials. The coordinate calculation is similar;

[0061] S6. Obtain the space for each target construction equipment and material ( After obtaining the coordinates, the spatial location of the corresponding equipment and materials is marked and displayed in the 3D model, and the coordinate data is fed back to each functional module and algorithm as the basis for subsequent modules to judge the compliance of construction procedures and monitor construction.

[0062] Furthermore, the smart safety helmet has a built-in camera, positioning tag, earphone, microphone, interactive button, and signal transceiver; it can receive operation instructions issued by the system platform through the earphone; it can provide feedback on its own status and issue commands through the microphone and interactive button; it can collect and record on-site information according to the needs of each module and algorithm through the camera; and it can provide the wearer's coordinate information in real time through the built-in positioning tag and on-site positioning base station, providing a data foundation for each functional module.

[0063] The laser positioning device has a built-in laser transmitter, signal transceiver, positioning tag, and transmission structure. At the construction site, it uses the positioning base station and the internal positioning tag to obtain the spatial coordinates of the laser positioning device, receives control commands issued by each construction module, and marks the installation positions of vacuum preloading construction equipment and materials for each process at the construction site through the transmission structure and laser transmitter, guiding on-site personnel to carry out construction.

[0064] Furthermore, the algorithm flow of the geotextile and sealing membrane laying construction monitoring algorithm module is as follows:

[0065] S1. Read the basic information of the vacuum preloading construction of the project, including the vacuum preloading design and construction model information and the drawing information obtained by analysis;

[0066] S2. Read various types of on-site collected data and call algorithms for recognition, including high-definition image data collected by the on-site camera group and three-dimensional scanning of the engineering site model, and use the basic image recognition algorithm group to identify the overall construction area of ​​the engineering vacuum preloading and the location of each control point in the image;

[0067] S3. Receive the input command to start the construction monitoring of the positioning poles from the program interface of the geotextile and sealing membrane laying monitoring module, call the basic image recognition algorithm group to monitor and identify each positioning pole in the on-site images collected during the construction process in real time, and mark them.

[0068] S4. Make a judgment. If the judgment finds that the current positioning benchmark coordinate data does not match the drawing and model information, the corresponding benchmark position and coordinate information will be reported to the supervisor through the smart safety helmet and wait for modification. After the modification is completed, the judgment in step S4 will be repeated. If the judgment finds that each positioning benchmark matches the drawing and model information, then continue to step S5.

[0069] S5. Receive the input command to start geotextile laying monitoring from the geotextile and sealing membrane laying monitoring module program interface, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the laying and splicing of each geotextile.

[0070] S6. Monitor and determine in real time whether there are any debris unrelated to the construction on the construction site and on the geotextile during the geotextile laying process; if there are debris during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for on-site care. After the processing is completed, return to step S6 to re-determine; if it is determined that there are no wrinkles during the geotextile laying process, continue to step S7.

[0071] S7. Monitor and determine in real time whether there are any wrinkles or deformations in the geotextile during the laying process; if there are wrinkles in the geotextile during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for processing. After processing is completed, return to step S6 to re-determine; if it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S8.

[0072] S8. Monitor and determine in real time whether there is any damage to the geotextile during the geotextile laying process. If there is any damage to the geotextile during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for processing. After processing is completed, return to step S6 to make a new judgment. If it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S9.

[0073] S9. After each step of the geotextile laying construction is completed, receive the input command from the geotextile and sealing membrane laying construction monitoring module program interface indicating that the geotextile laying construction monitoring is complete, and end the real-time monitoring of the geotextile laying construction by the above algorithm.

[0074] S10. Organize and record the early warning data during the geotextile laying process in this procedure, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0075] Furthermore, the algorithm flow of the drainage board installation monitoring algorithm module is as follows:

[0076] S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, geotextile laying construction status and process processing data.

[0077] S2. Read and detect the positioning coordinates on the drainage board installation equipment, and verify whether the coordinate positions of each positioning tag in the algorithm program are accurate and consistent with the actual site conditions.

[0078] S3. Analyze the construction design drawings or model file data to obtain the coordinates of the insertion positions in each drainage board construction design scheme.

[0079] S4. Receive the drainage board installation construction monitoring start command input by the drainage board installation construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the installation of each drainage board.

[0080] S5. Real-time monitoring uses the coordinate information provided by the positioning tag of the drainage board installation equipment to initially determine whether the current drainage board installation coordinates are consistent with the design and construction documents. At the same time, the coordinate reference recognition algorithm is called to further determine the accuracy of the current drainage board installation coordinates. If the current drainage board installation coordinates are inconsistent with the design and construction plan information, the corresponding construction information is reported to the supervisor and awaits processing. After processing is completed, return to step S5 to re-determine; if the current drainage board installation coordinates are consistent with the design and construction plan information, continue to step S6.

[0081] S6. Monitor and determine in real time whether the drainage board is shifted during the insertion process. Use the basic image recognition algorithm group to identify the drainage board image in the high-definition image during the insertion process and make a judgment. If the judgment finds that the drainage board has shifted during the insertion process, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the drainage board is inserted again and the judgment in step S6 is repeated. If no shift is found in the current drainage board during the insertion process, continue to step S7.

[0082] S7. Monitor and determine in real time whether the drainage board is damaged during the insertion process. Call the basic image recognition algorithm group to analyze the images collected on site and determine whether there are any color differences that the algorithm is sensitive to during the entire installation process of the drainage board, so as to determine the damage of the drainage board. If the drainage board is found to be damaged during the insertion process, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation of the drainage board and the judgment in step S7 are repeated. If no damage is found during the installation process of the drainage board, proceed to step S8.

[0083] S8. After the current drainage board is installed, determine whether the external leakage of the drainage board meets the requirements. By calling the basic image recognition algorithm group to identify the drainage board graphic and the scale value marked on the drainage board from the on-site image, the judgment is made. If the external leakage scale value is not within the range specified in the construction requirements, it is considered as not meeting the requirements. The corresponding construction information is reported to the supervisor and awaits processing. After processing is completed, the judgment in step S8 is repeated. If the external leakage scale value is within the range specified in the construction requirements, it is considered as meeting the requirements. Then, return to step S5 to start the construction monitoring program algorithm loop for the next drainage board. If the installation of all drainage boards has been completed, continue to step S9.

[0084] S9. When the on-site personnel confirm that the installation of each drainage board is completed and there are no abnormalities, they shall accept the input command that the installation of the drainage board is completed from the program interface of the drainage board installation monitoring module, and end the real-time monitoring of the drainage board installation by the above algorithm.

[0085] S10. Organize and record the early warning data and key step video image data during the installation of drainage boards in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0086] Furthermore, the algorithm flow of the vacuum booster system installation and construction monitoring algorithm module is as follows:

[0087] S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, and the operation status and process processing data of preceding construction.

[0088] S2. Analyze the construction design drawings or model file data to obtain the coordinates of the vacuum piping system layout, the coordinates of the connection joints with the drainage board, and the installation coordinates of the main connecting pipe fittings.

[0089] S3. Receive the vacuum booster system installation start command input by the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring of the installation and construction of each pipeline and connector of the vacuum booster system during the installation and construction process;

[0090] S4. Monitor the vacuum pipeline construction process in real time, call the basic image recognition algorithm group to analyze the images collected on site, and use the basic image recognition algorithm group and coordinate reference recognition algorithm to identify the coordinates of the first, middle and last sections of each vacuum pipeline and compare them with the construction design information. If the position and direction coordinates of each pipeline are found to be off from the design plan during construction, the corresponding construction information is reported to the supervisor and awaits processing. After processing, the installation of the vacuum pipeline section and the judgment in step S4 are carried out again. If no damage is found during the installation of the drainage board, proceed to step S5.

[0091] S5. Monitor and determine in real time whether the vacuum pipeline and connectors are damaged during installation. Analyze the images collected on-site using the basic image recognition algorithm group. If damage to the vacuum pipeline is found during installation, notify the supervisory personnel of the corresponding construction information and wait for processing. After processing, re-install the vacuum pipeline and repeat the judgment in step S5. If no damage is found during the installation of the vacuum pipeline, continue to step S6.

[0092] S6. Monitor in real time whether the connection between the vacuum pipeline and the drainage board of various types of joints is reliable. If there is any abnormality in the connection image or connection point coordinates between the vacuum pipeline and the drainage board, it is considered that the connection between the vacuum pipeline and the drainage board is unreliable. The corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation construction of the pipeline connection and the judgment in step S6 are carried out again. If there is no abnormality in the connection image or connection point coordinates between the vacuum pipeline and the drainage board, it is considered that the connection between the vacuum pipeline and the drainage board is reliable, and the process continues to step S7.

[0093] S7. Receive the start command for installation and construction monitoring of the under-membrane vacuum probe from the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring of the installation and construction of each pipeline and connector of the vacuum booster system.

[0094] S8. Monitor the construction process of the vacuum pipeline system in real time, and call the basic image recognition algorithm group to analyze the images collected on site; if the installation coordinates of the probe do not match the design information, the corresponding construction information will be reported to the supervisor and wait for processing. After processing, the installation of the current probe equipment and the judgment of step S8 will be carried out again; if the coordinates of the current probe equipment are accurate, continue to step S9.

[0095] S9. After the installation of each probe device is completed, according to the different probe characteristics and data exchange methods, the initial degree of each probe is read through the data exchange equipment or high-definition image recognition to determine whether the initial degree of each probe is abnormal, so as to avoid probe distortion and data fraud. If the initial reading of the probe is abnormal, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation of the current probe device and the judgment in step S9 are repeated. If the initial reading of the current probe is not abnormal, the process continues to step S10.

[0096] S10. After the vacuum booster system and the probe under the membrane are installed, receive the vacuum booster system installation completion command input by the vacuum booster system installation and construction monitoring module, and end the real-time monitoring of the vacuum system construction by the above algorithm.

[0097] S11. Organize and record the early warning data and key step video image data during the construction of the vacuum booster system in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0098] The advantages and technical effects of this invention are as follows: This invention provides a foundation vacuum preloading construction supervision system based on image recognition technology. It uses color marking to identify traditional construction materials and elements and uses AI image recognition technology to monitor the compliance of each construction stage, thereby achieving automated supervision of construction, saving the workload required for project management, and guiding on-site construction personnel to complete each stage of construction according to requirements. While reducing personnel input and labor costs, it improves the safety and economic benefits of dam foundation treatment construction. Attached Figure Description

[0099] Figure 1 The overall architecture diagram of the foundation vacuum preloading construction monitoring system based on image recognition technology provided by the present invention.

[0100] Figure 2 The execution flowchart of the geotextile and sealing membrane laying construction monitoring algorithm provided by the present invention.

[0101] Figure 3 The execution flowchart of the drainage board installation monitoring algorithm provided by the present invention is shown.

[0102] Figure 4 The execution flowchart of the vacuum booster system installation and construction monitoring algorithm provided by the present invention is shown. Detailed Implementation

[0103] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0104] This invention provides a foundation vacuum preloading construction monitoring system based on image recognition technology, referenced Figure 1 It includes engineering data acquisition equipment for data exchange, construction positioning equipment, auxiliary construction equipment, geotextile and sealing membrane laying monitoring module, drainage board installation monitoring module, vacuum pressurization system installation monitoring module, and construction data recording module. It also includes the layout and marking methods for each construction device, coordinate reference identification algorithms, and the supporting algorithm flows required for each module. A detailed description of each functional module follows:

[0105] (I) Engineering Data Acquisition Equipment

[0106] The engineering information acquisition equipment includes a high-definition camera array, a camera lifting pole, and a 3D laser scanner. It is used to acquire high-definition video images and engineering model information from the site. The acquired image information will be used by various engineering construction supervision modules for image recognition and data analysis of the construction process; the acquired model information will be used to connect various engineering data for data querying in corresponding functional modules.

[0107] (1.1) High-definition camera group

[0108] The high-definition camera group consists of a monitoring network composed of multiple cameras deployed at the vacuum preloading construction site, with positioning tags installed on each camera. Starting from the cleaning process of the vacuum preloading construction site, the high-definition camera group is deployed around the construction site and in the site using camera lifting rods, according to the deployment method, to collect high-definition images of each stage of the vacuum preloading construction for subsequent identification and judgment by various construction monitoring functional modules.

[0109] (1.2) Camera lifting pole

[0110] The camera lifting pole supports each high-definition camera. The pole contains a transmission device and a signal receiving device, which can adjust the camera height according to the coordinate reference recognition algorithm requirements, collecting data from the necessary height. The maximum height of the camera lifting pole is related to the size of the vacuum preloading construction site and the camera's accuracy.

[0111] (1.3) 3D scanner

[0112] The 3D scanner, either deployed at the vacuum preloading construction site or carried by construction and installation personnel, uses laser ranging technology to receive feedback signals and convert actual spatial information into a 3D model, acquiring a site model of the vacuum preloading construction site and equipment. This model serves to carry data from subsequent construction monitoring modules and provides an intuitive interface for data querying and issuing operation commands.

[0113] (II) Algorithm for Deploying High-Definition Camera Groups

[0114] The high-definition camera group deployment algorithm analyzes and calculates based on the characteristics and requirements of the technical approach of this invention. It is used to determine the deployment method of the high-definition camera group in the vacuum preloading construction site. The algorithm combines camera performance parameters to use the minimum number of cameras while ensuring clear and reliable image data. It also considers the requirements of the subsequent benchmark deployment algorithm to deploy cameras at different heights and shooting angles. The high-definition camera deployment is mainly divided into two types: deployment along the outer contour line of the construction area and deployment within the construction area. The specific algorithm flow is as follows:

[0115] S1. Read the outer contour line of the area to be vacuum preloaded from the corresponding construction details or BIM model.

[0116] S2. Calculate the number and spacing of cameras along the outer contour line using the following formula:

[0117]

[0118] In the formula The cameras are spaced out along the outer contour line of the construction site. This is the shape adjustment coefficient for the vacuum preloading area; This is the maximum pixel count of the camera; This refers to the horizontal resolution coefficient. This refers to the lens's light-gathering size; This represents the minimum illumination resolution coefficient of the camera. The number of sides of the outer contour of the vacuum preloading site determines the shape of the construction site. A circle has the highest value of 1. The more sides, the more complex the shape of the construction site. The lower the value, the greater the value; in general engineering practice The value varies from approximately 30 to 60 meters depending on the performance of the camera used. This can be seen from the formula. The value is related to the accuracy parameters of each camera; the higher the camera accuracy, the better. The higher the value, the larger the interval between cameras. Given a fixed perimeter of the outer contour of the construction site, fewer cameras are needed.

[0119]

[0120] In the formula The spacing of cameras within the construction area; The cameras are spaced out along the outer contour line of the construction site. This represents the camera's wide-angle coefficient. The value is related to the size of the camera's capture range; the larger the range, the higher the value. In general engineering practice... The value varies depending on the performance of the camera used, and is approximately in the range of 50-80 meters.

[0121] S3. Based on the camera arrangement intervals along the outer contour line of the construction site and the camera arrangement intervals within the construction area, divide the construction site into grids and generate a high-definition camera group layout plan.

[0122] S4. Generate a camera layout plan and count the number of cameras and camera lifting poles for use in subsequent module processes and algorithms.

[0123] (iii) Construction positioning equipment

[0124] The construction positioning equipment includes positioning benchmarks, positioning base stations, and positioning tags. This construction area positioning network assists the AI ​​image recognition function in recognizing images of each stage of vacuum preloading, thereby increasing image recognition accuracy, reducing image recognition difficulty, improving the efficiency of each module's program operation, and reducing energy consumption.

[0125] (3.1) Positioning benchmark

[0126] The positioning markers consist of a pole body, a pole head, and positioning tag fixing components. They are installed in the vacuum preloading construction area according to the marker placement algorithm requirements. The pole head is painted with a conspicuous color that clearly distinguishes it from the construction ground, geotextile, vacuum membrane, construction equipment, and personnel, facilitating camera recognition and algorithmic identification and positioning of image content. Simultaneously, positioning tags that can communicate with the positioning base station are installed to periodically receive and calibrate the position coordinates of each positioning marker, providing basic parameters for each module to call the coordinate reference recognition algorithm for calculations. During installation, the positioning markers are inserted into the vacuum preloading construction site, with the pole head protruding from the construction ground, generally 5cm above the ground surface.

[0127] (3.2) Positioning base station

[0128] The positioning base station uses UWB positioning technology to locate the coordinates of the positioning tags in the vacuum preloading construction equipment and positioning poles. The positioning base station is deployed in the vacuum preloading construction site. It receives tag signals through the wireless network in the project area and calculates the distance and angle of each device and the positioning tag device in the positioning pole according to the arrival time, arrival time difference and arrival angle algorithm, and obtains the corresponding coordinate information.

[0129] (3.3) Location tags

[0130] The positioning tags, installed in vacuum preloading construction equipment and auxiliary materials such as drainage board insertion machines and positioning poles, can respond to wireless signals emitted by positioning base stations. Through the information feedback time difference, they obtain the spatial coordinates of each positioning tag and its corresponding equipment, used for coordinate positioning of each piece of equipment and material, and for spatial display of the corresponding equipment model in the construction 3D model. When installed with the positioning pole, the positioning tags are installed below the head of the positioning pole, and after installation, they are higher than the construction ground level.

[0131] (iv) Benchmarking Algorithm

[0132] The benchmark placement algorithm is used to specify and constrain the spacing and number of positioning benchmarks required for the project, providing a highly recognizable benchmark positioning network for the coordinate reference recognition algorithm. This allows the coordinate reference recognition algorithm to obtain the positioning coordinates of each construction material and equipment through multi-angle image recognition and reference calculation. The specific calculation process of the benchmark placement algorithm is as follows:

[0133] S1. Read or input actual parameter data of the vacuum preloading construction site and materials of the project, as well as data from engineering drawings and models.

[0134] S2. Calculate the installation interval of the positioning markers. The calculation formula is as follows:

[0135]

[0136] In the formula The spacing for positioning marker installation; Adjust the coefficients according to the required positioning coordinate accuracy; The area of ​​the vacuum preloading construction site; This refers to the minimum area of ​​a single piece of geotextile and sealing membrane used; This refers to the number of drain pipe connectors; This is the sum of the number of horizontal and vertical pipes in the pressurization pipeline system; The accuracy calibration coefficient of the booster system. The installation interval is related to the precision and difficulty of the engineering construction and the accuracy parameters of the camera; the higher the required value, the smaller the installation interval. The installation interval is related to the difficulty and precision requirements of the booster system in the project; the higher the required value, the smaller the installation interval. If the booster system is not used in the vacuum preloading construction plan, then... The value is 0; The theoretical value is greater than 0m, but in general engineering cases, the minimum actual interval is not less than 5m; the common interval is 10-20m.

[0137] S3. Based on the calculated installation interval of the positioning markers, the vacuum preloading construction site is divided into grids, and a layout diagram of the positioning marker installation is generated.

[0138] S4. The generated results guide the construction site in completing the installation and layout of each positioning marker.

[0139] (v) Marking method for construction equipment and materials

[0140] The aforementioned marking method for construction equipment and materials uses non-toxic, harmless, waterproof, and corrosion-resistant colored construction coatings to mark key equipment and materials in different construction stages during vacuum preloading in a corresponding way, thereby improving the image recognition efficiency of the module and algorithm and reducing the difficulty and cost of algorithm recognition.

[0141] (5.1) Marking methods for geotextiles and sealing films

[0142] The marking method for geotextiles and sealing membranes is used to color-mark the woven geotextiles, non-woven geotextiles, and vacuum sealing membranes used in vacuum preloading construction. The selected paint must have a significant color difference from the layer where the geotextile or sealing membrane is laid and the soil in the construction area. This method ensures that when the geotextile or sealing membrane is damaged, a clear color difference will be displayed in the video image, improving the efficiency and accuracy of automatic image recognition. Taking RGB colors as an example, the color difference of the selected paint for the geotextile and sealing membrane should meet the following calculation formula:

[0143]

[0144] In the formula The RGB parameters are for the coloring of the geotextile and sealing membrane to be laid; This refers to the RGB parameters of the color of the underlying geotextile, sealing membrane, or construction site in the current process, with each RGB parameter ranging from 0 to 255. This formula constrains the selection of paint colors for the geotextile and sealing membrane, ensuring a significant color difference with the underlying materials and other construction equipment materials, which facilitates effective image recognition by various algorithms.

[0145] (5.2) Drainage board marking method

[0146] The drainage board marking method is used to color-mark the drainage boards required for the installation process during vacuum preloading. The outer and inner drainage channels of the drainage board must be painted with different colors, and there must be a clear difference between the outer and inner sides of the drainage board and the woven geotextile laid in the current process. This method ensures that when the drainage board breaks, deforms, twists, or misaligns, the inner paint will be exposed. Simultaneously, the color of the outer side will be affected by the mechanical effects of twisting and compression, thus creating a clear difference between the normally installed drainage board and the image, improving the efficiency and accuracy of automatic image recognition. The drainage board will also be marked with length markings and numbers for automatic identification of the insertion depth of each drainage board. The color difference of the selected paint for the drainage board, taking RGB colors as an example, should meet the following calculation formula:

[0147]

[0148]

[0149] Both of the above formulas must be satisfied simultaneously, and they are related by an "AND" expression. In the formula... RGB parameters for coloring the outer side of the drainage board; RGB parameters for coloring the outer side of the drainage board; The RGB parameters represent the colors of the underlying geotextile, sealing membrane, or construction site used in the current process, with each RGB parameter ranging from 0 to 255. The formula shows that the color difference requirements for the inner and outer sides of the drainage board are higher. Therefore, this formula is used to constrain the selection of pigments for the inner and outer coatings of the drainage pipe. When damage to the drainage board occurs, a significant color difference will appear in the high-definition monitoring image, facilitating effective image recognition by various algorithms.

[0150] (5.3) Vacuum pipe marking method

[0151] The vacuum pipeline marking method is used to color-mark the vacuum system pipelines during vacuum preloading construction. The inner and outer sides of the vacuum pipelines, as well as the connectors to the drainage boards, are painted with different colors, ensuring that the colors of each construction material are significantly different from the woven geotextile laid in the current process. This method allows for clear color differences in the high-definition images captured when the vacuum pipeline is damaged, facilitating program identification. Furthermore, the overall layout of the vacuum pipelines and the connectors to the drainage boards generally follow a strong regularity according to the construction drawings. By color-marking and comparing with relevant drawings, algorithms can identify various vacuum system layout construction problems, such as misaligned vacuum pipeline layouts, misaligned or missing drainage board connectors. For the vacuum pipelines and drainage board connectors, the selected paint color difference, using RGB colors as an example, should meet the following calculation formula:

[0152]

[0153]

[0154]

[0155] The above three formulas must be satisfied simultaneously, and are related by an "AND" expression. In the formula... RGB parameters for coloring the outside of the vacuum pipe; RGB parameters for coloring the inside of the vacuum pipe; RGB parameters for coloring the drainage board connector; The RGB parameters represent the colors of the underlying geotextile, sealing membrane, or construction site used in the current process, with each RGB parameter ranging from 0 to 255. The formula shows that the color difference requirements for the inner and outer sides of the vacuum pipe and the connection to the drainage board are higher, which facilitates effective image recognition by various algorithms.

[0156] (5.4) Marking method for vacuum probe and pressure gauge

[0157] The vacuum probe and pressure gauge marking method is used to color-code monitoring equipment such as vacuum probes and pressure gauges required in vacuum preloading construction. The paint selected for the vacuum probes and pressure gauges must have a more pronounced color difference from the layer where the geotextile or sealing membrane is laid and the soil in the construction area. This allows the relevant image recognition algorithm to more efficiently identify the relevant monitoring equipment from the image, facilitating identification and recording by the algorithm and program functions. Taking RGB colors as an example, the color difference of the selected paint should meet the following calculation formula:

[0158]

[0159] In the formula The RGB parameters of the monitoring equipment to be installed in the current process, such as vacuum probes or pressure gauges; The RGB parameters represent the colors of the underlying geotextile, sealing membrane, or construction site in the current process, with each RGB parameter ranging from 0 to 255. The formula shows that the monitoring equipment to be installed needs to have a significant color difference from the underlying construction materials or soil color in the current process, which is beneficial for effective image recognition by the various algorithms.

[0160] (5.5) Marking method for positioning benchmarks and settlement gauges

[0161] The marking method for the positioning markers and settlement gauges is used to color-mark the heads of the positioning markers and the outer shells of the settlement gauges. When the positioning markers are installed, they are inserted into the vacuum preloading construction site, and the exposed heads are painted. The color difference of the paint selected for the positioning marker heads and the outer shells of the settlement gauges, using RGB colors as an example, should meet the following calculation formula:

[0162]

[0163] In the formula RGB parameters for coloring the head of the positioning marker or the outer casing of the settlement meter; This formula defines the RGB parameters for the color of the underlying geotextile, sealing membrane, or construction site in the current process, with each RGB parameter ranging from 0 to 255. This formula constrains the selection of the paint color for the positioning marker head or the settlement meter casing, ensuring a significant color difference between the underlying construction materials or the site soil, which facilitates effective image recognition by various algorithms.

[0164] (vi) Basic Image Recognition Algorithm Group

[0165] The aforementioned basic image recognition algorithm is a set of basic image recognition algorithms that can be externally invoked by various functional modules within the system platform. This set of algorithms must possess various basic image analysis and recognition functions, including pixel recognition, pixel classification, image extraction, image feature value analysis, and data comparison for acquired images. Each functional module, by invoking the corresponding basic image recognition algorithm, uses pixel recognition and pattern analysis to identify the equipment and materials related to vacuum preloading construction from the acquired images, thus assisting each module in achieving the required basic image recognition functions and acquiring relevant data.

[0166] (vii) Coordinate reference recognition algorithm

[0167] The coordinate reference recognition algorithm utilizes positioning benchmarks with known coordinates set up at the construction site as positioning references. Then, through high-definition images captured from multiple angles, it compares and calculates the spatial coordinates of various construction equipment and materials in the images to obtain the calculated coordinate values ​​of each piece of equipment and material. This provides a basis for each module's function to determine the compliance of on-site installation and construction procedures. The specific algorithm flow is as follows:

[0168] S1: Receive commands from various functional modules requesting coordinates of construction equipment and materials.

[0169] S2. Call the basic image recognition algorithm to analyze and identify the target pixel color and the presented pattern, and initially search for the target construction equipment and materials that need to be located in the image, as well as the nearest 1-3 sets of positioning benchmarks within the target range.

[0170] S3. By analyzing the patterns captured by each camera, the target pixel color and the size of the pattern in the captured image are sorted. The camera with the larger pixel area in the captured image is considered to be the closest to the target. Based on this rule, the cameras are sorted in reverse order according to their distance from the target device and material. That is, the closest to the target is sorted as 1, and the farthest from the target is sorted as the largest.

[0171] S4. Set two camera height levels, high and low. Based on the order of camera distance from the target, alternately set the camera height: when the camera is ranked 1, the camera height level is high; when the camera is ranked 2, the camera height level is low; when the camera is ranked 3, the camera height level is high, and so on. Send a signal to the camera lifting rod in the engineering data acquisition equipment to adjust the height of each camera. Collect image information of the target construction equipment and materials from different heights and angles to provide a basis for subsequent coordinate calculation. At the same time, if a camera cannot collect pixel or graphic feature information of the target construction equipment and materials during image acquisition, it is considered that the camera is in an obstructed state, and the height will be adjusted first to achieve image acquisition.

[0172] S5. Using the current camera coordinates and the coordinates of the nearest positioning benchmark to the target construction equipment and materials in the image, calculate the coordinates of the target construction equipment and materials in the images captured by each camera. The calculation formula is as follows:

[0173]

[0174]

[0175]

[0176] Y, Z axis coordinates The calculation formula is the same.

[0177] In the formula To obtain the initial X-axis coordinate values ​​of the target construction equipment and materials based on the images captured by each camera; The X-axis coordinate of the nearest positioning benchmark in the image captured by the current camera that is clearly observable and identifiable of the target construction equipment and materials; This refers to the difference in X-axis coordinates between the nearest positioning benchmark and the target construction equipment and materials, calculated based on the images captured by the current camera. This is the camera adjustment factor; Adjust the weighting coefficients for each camera position; The maximum angle between the target plane and the plane containing the observed target; This is the distance between the current camera and the nearest positioning marker for the target construction equipment and materials; Adjustment parameters are calculated for the weights; The X-axis coordinates of the target construction equipment and materials are obtained using the algorithm; N is the total number of lenses used to capture images of this target construction equipment and materials. The same principle applies to coordinate calculation.

[0178] The formula shows that... This was obtained by adjusting and calculating the X-axis difference between the positioning benchmark and the target in the image captured by the camera. The values ​​are derived from the coordinates fed back by the positioning tag in the positioning benchmark; It is related to the angle between the direct angle of the camera's central axis and the XY plane where the benchmark is located, as well as parameters such as the camera's focal length and performance. The selected plane must include the coordinate axis directions to be calculated. Generally, since the positioning pole and the observation target are close to the ground (i.e., the XY plane), the XY plane is selected for the calculation of the X and Y axes, and the larger angle between the Z axis and the XZ plane and the YZ plane is selected for the calculation of the Z axis. Therefore, the image data collected by the camera at a lower height needs to be used when calculating the Z axis to improve the calculation accuracy. Adjust the parameters for weight calculation, and adjust them through weight calculation to make... The calculated value is 1; The calculation formula shows that the closer the camera is to the target, the larger the angle between the camera and the plane containing the target and the coordinate axis to be calculated. The higher the weight, the greater the total number of shots captured. The value of N is related to the camera's acquisition accuracy performance parameters and the accuracy requirements of the target construction equipment and materials. The weaker the camera's performance, the higher the target acquisition accuracy requirement, and the higher the value of N. This can be seen from the formula. The value is the weighted sum of the preliminary calculated values ​​of the coordinates of the target construction equipment and materials obtained by each camera; The same principle applies to coordinate calculation.

[0179] S6. Obtain the space for each target construction equipment and material ( After obtaining the coordinates, the spatial location of the corresponding equipment and materials is marked and displayed in the 3D model, and the coordinate data is fed back to each functional module and algorithm as the basis for subsequent modules to judge the compliance of construction procedures and monitor construction.

[0180] (viii) Auxiliary construction equipment

[0181] The auxiliary construction equipment includes smart safety helmets and laser positioning devices. These devices assist on-site personnel in performing each stage of vacuum preloading through voice and laser positioning markers, while simultaneously collecting and feeding back on-site construction information. The collected data serves as the basis for image recognition and program command issuance by various functional modules, and also provides information for construction process records and receipts, which are linked to the corresponding 3D engineering model.

[0182] (8.1) Smart safety helmet

[0183] The smart safety helmet incorporates a built-in camera, positioning tag, earphone, microphone, interactive button, and signal transceiver. It receives operation commands from the system platform via the earphone. The microphone and interactive button provide feedback on its own status and issue commands. The camera collects and records on-site information according to the needs of each module and algorithm. Through the built-in positioning tag and on-site positioning base station, it can provide the wearer's coordinates in real time, providing a data foundation for each functional module.

[0184] (8.2) Laser positioning device

[0185] The laser positioning device integrates a laser transmitter, signal transceiver, positioning tag, and transmission structure. It can obtain spatial coordinates at the construction site using a positioning base station and internal positioning tags. It receives control commands from various construction modules. Through the transmission structure and laser transmitter, it marks the installation locations of vacuum preloading equipment and materials for each construction process, guiding on-site personnel during construction.

[0186] (ix) Monitoring module for geotextile and sealing membrane laying construction

[0187] The geotextile and sealing membrane laying construction monitoring module provides an interactive program interface and calls resources such as geotextile and sealing membrane laying construction monitoring algorithms, coordinate reference recognition algorithms, foundation image recognition algorithm groups, and high-definition camera groups. According to the corresponding construction process and progress, it monitors and manages the geotextile and sealing membrane laying process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervision personnel for on-site handling.

[0188] (x) Monitoring algorithm for geotextile and sealing membrane laying construction

[0189] The aforementioned geotextile and sealing membrane laying construction monitoring algorithm can, according to the needs and commands of the corresponding modules, invoke algorithms and resources to identify and detect the compliance of various aspects of construction equipment and materials involved in the geotextile and sealing membrane laying construction process, and provide feedback and record the corresponding information. (Reference) Figure 2 The specific algorithm flow is as follows:

[0190] S1. Read the basic information of the vacuum preloading construction of the project, including the vacuum preloading design and construction model information and the drawing information obtained by analysis.

[0191] S2. Read various types of on-site collected data and call algorithms for recognition, including high-definition image data collected by the on-site camera group and three-dimensional scanning of the engineering site model, and use the basic image recognition algorithm group to identify the overall construction area of ​​the engineering vacuum preloading and the location of each control point in the image.

[0192] S3. Receive the input command from the geotextile and sealing membrane laying construction monitoring module program interface to start the monitoring of positioning markers, call the basic image recognition algorithm group to monitor and identify each positioning marker in the on-site images collected during the construction process in real time, and mark them.

[0193] S4. During the installation of each positioning marker, the coordinate values ​​of the positioning markers are read through the positioning base station and positioning tag, and compared with the design coordinate values ​​of each positioning marker in the design drawings and model information. If it is found that the current positioning marker coordinate data does not match the drawing and model information, the corresponding marker position and coordinate information will be reported to the supervisor through the smart safety helmet and wait for modification. After the modification is completed, the judgment in step S4 will be repeated. If it is determined that each positioning marker matches the drawing and model information, then proceed to step S5.

[0194] S5. Receive the input command to start geotextile laying monitoring from the geotextile and sealing membrane laying monitoring module program interface, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the laying and splicing of each geotextile.

[0195] S6. Real-time monitoring and judgment of whether there are any unrelated debris on the construction site and on the geotextile during the geotextile laying process, to prevent damage to the geotextile during laying. When debris is found, the corresponding part in the high-definition image collected on site will show obvious color difference and pattern with the geotextile. Therefore, basic image recognition algorithms can be used to identify the remaining debris, and coordinate reference recognition algorithms can be used for positioning and marking. If debris is found during the laying process, the algorithm will notify the supervisor of the construction risk information through a smart safety helmet and wait for on-site handling. After the handling is completed, return to step S6 to re-judge; if it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S7.

[0196] S7. Real-time monitoring and judgment of whether there are wrinkles or deformations in the geotextile during the laying process. When wrinkles and deformations occur, the corresponding parts of the geotextile will show obvious shadow color difference patterns in the high-definition images collected on site, and the patterns at the splicing points will be distorted. Therefore, basic image recognition algorithms can be used to identify the wrinkled parts, and coordinate reference recognition algorithms can be used for positioning and marking. If there are geotextile wrinkles during the laying process, the algorithm will notify the supervisory personnel of the existing construction risk information through the smart safety helmet and wait for processing. After processing is completed, return to step S6 to re-judge; if it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S8.

[0197] S8. Real-time monitoring and judgment of whether there is geotextile damage during the geotextile laying process. When damage occurs, the corresponding part of the geotextile will show a significant color difference with the underlying soil in the high-definition image collected on site. Therefore, the damaged part can be identified using a basic image recognition algorithm and located and marked using a coordinate reference recognition algorithm. If geotextile damage is found during the laying process, the algorithm will notify the supervisor of the construction risk information through a smart safety helmet and wait for processing. After processing, it will return to step S6 for re-judgment; if it is determined that there are no wrinkles in the geotextile during the laying process, it will continue to step S9.

[0198] S9. After each step of the geotextile laying construction is completed, receive the input command from the geotextile and sealing membrane laying construction monitoring module program interface indicating that the geotextile laying construction monitoring is complete, and end the real-time monitoring of the geotextile laying construction by the above algorithm.

[0199] S10. Organize and record the early warning data during the geotextile laying process in this procedure, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0200] (xi) Drainage board installation construction monitoring module

[0201] The drainage board installation monitoring module provides an interactive program interface and calls resources such as drainage board installation monitoring algorithm, coordinate reference recognition algorithm, basic image recognition algorithm group and high-definition camera group. According to the corresponding construction process and progress, it monitors and manages the drainage board installation process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervision personnel for on-site handling.

[0202] (xii) Algorithm for monitoring the installation of drainage boards

[0203] The drainage board installation construction monitoring algorithm can, according to the corresponding module requirements and command calls, call algorithm and resource pairs to identify and detect the compliance of various aspects of the construction equipment and materials involved in the drainage board installation process through pixel and image judgment and coordinate calculation, and then feed back and record the corresponding information. (Reference) Figure 3 The specific algorithm flow is as follows:

[0204] S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, geotextile laying construction status and process processing data, etc.

[0205] S2. Read and detect the positioning coordinates on the drainage board installation equipment. If manual installation is used, positioning tags need to be installed on the drainage board installation auxiliary gun to obtain the installation coordinate data of each drainage board; if a board insertion machine is used, positioning tags need to be installed at the insertion and removal points and the head to obtain the installation coordinate data of each drainage board. At the same time, verify whether the coordinate positions responded by each positioning tag in the algorithm program are accurate and consistent with the actual site conditions.

[0206] S3. Analyze the construction design drawings or model file data to obtain the coordinates of the insertion positions in each drainage board construction design scheme.

[0207] S4. Receive the drainage board installation monitoring start command input by the drainage board installation monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the installation of each drainage board.

[0208] S5. Real-time monitoring uses the coordinate information provided by the positioning tags of the drainage board installation equipment to initially determine whether the current drainage board installation coordinates are consistent with the design and construction documents. Simultaneously, a coordinate reference recognition algorithm is invoked to further determine the accuracy of the current drainage board installation coordinates by referencing the positioning tags in the drainage board installation area. The feedback information from the construction equipment positioning tags is also verified. If the current drainage board installation coordinates are inconsistent with the design and construction plan information, the corresponding construction information is reported to the supervisory personnel and awaits processing. After processing, the process returns to step S5 for re-evaluation. If the current drainage board installation coordinates are consistent with the design and construction plan information, proceed to step S6.

[0209] S6. Monitor and determine in real time whether the drainage board shifts during insertion. Utilize the basic image recognition algorithm to acquire and identify the characteristic color pixels and graphics of the drainage board in the high-definition images collected on-site. After identifying the drainage board, monitor the insertion process and record the insertion process with cameras at key time points. Using the basic image recognition algorithm, identify the drainage board image in the high-definition images during insertion, determining whether the corresponding graphic of the drainage board remains perpendicular to the ground and whether a large area of ​​tilting occurs during the process. If displacement is detected during insertion, the corresponding construction information is reported to the supervisor and awaits processing. After processing, the insertion of the drainage board and the judgment in step S6 are repeated. If no displacement is detected during the current insertion process, proceed to step S7.

[0210] S7. Monitor and determine in real time whether the drainage board is damaged during insertion. Analyze the images collected on-site using the basic image recognition algorithm. Due to the significant color difference between the outer and inner coatings of the drainage board, the basic image recognition algorithm identifies the characteristic colors and other color differences within the drainage board in the on-site image. This determines whether any algorithm-sensitive color differences occurred during the entire installation process, thus assessing the drainage board's damage. If damage is detected during insertion, the corresponding construction information is reported to the supervisor and awaits processing. After processing, the drainage board insertion is repeated, and the assessment in step S7 is repeated. If no damage is found during insertion, proceed to step S8.

[0211] S8. After the current drainage board is installed, determine whether the external leakage of the drainage board meets the requirements. By calling the basic image recognition algorithm group to identify the drainage board graphic and the scale value marked on the drainage board from the on-site image, the judgment is made. If the external leakage scale value is not within the range specified in the construction requirements, it is considered as not meeting the requirements. The corresponding construction information is reported to the supervisor and awaits processing. After processing is completed, the judgment in step S8 is repeated. If the external leakage scale value is within the range specified in the construction requirements, it is considered as meeting the requirements. Then, return to step S5 to start the construction monitoring program algorithm loop for the next drainage board. If the installation of all drainage boards has been completed, continue to step S9.

[0212] S9. When the on-site personnel confirm that the installation of each drainage board is completed and there are no abnormalities, they shall accept the input command from the drainage board installation monitoring module program interface indicating that the installation of the drainage board is complete, and end the real-time monitoring of the drainage board installation by the above algorithm.

[0213] S10. Organize and record early warning data, key step video images, and other data during the installation of drainage boards in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0214] (xiii) Vacuum booster system installation and construction monitoring module

[0215] The vacuum booster system installation and construction monitoring module provides an interactive program interface and calls resources such as vacuum booster system installation and construction monitoring algorithms, coordinate reference recognition algorithms, basic image recognition algorithm groups, and high-definition camera groups. According to the corresponding construction process and progress, it monitors and manages the vacuum booster system installation and construction procedures in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervisors for on-site handling.

[0216] (XIV) Monitoring Algorithm for Vacuum Booster System Installation

[0217] The vacuum booster system installation and construction monitoring algorithm can, according to the corresponding module requirements and command calls, invoke algorithms and resources to identify and detect the compliance of various aspects of the construction equipment and materials involved in the vacuum booster system installation and construction process through pixel and image judgment and coordinate calculation, and then feed back and record the corresponding information. (Reference) Figure 4 The specific algorithm flow is as follows:

[0218] S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, and the operation status and process processing data of preceding construction.

[0219] S2. Analyze the construction design drawings or model file data to obtain the coordinates of the vacuum piping system layout, the coordinates of the connection joints with the drainage board, and the installation coordinates of the main connecting pipe fittings.

[0220] S3. Receive the vacuum booster system installation start command input from the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the installation and construction of each pipeline and connector of the vacuum booster system.

[0221] S4. Monitor the vacuum pipeline construction process in real time. Analyze the images collected on-site using the basic image recognition algorithm. Due to the significant color difference between the vacuum system pipeline, geotextile, and the soil at the construction site, the basic image recognition algorithm can identify and mark the pipeline in high-definition images. Then, the coordinate reference recognition algorithm is used to identify the coordinates of the first, middle, and last segments of each vacuum pipeline and compare them with the construction design information. If the position and direction coordinates of each pipeline are found to deviate from the design plan during construction, the corresponding construction information is reported to the supervisors and awaits processing. After processing, the installation of that section of vacuum pipeline is restarted, and step S4 is repeated. If no damage is found during the installation of the drainage board, proceed to step S5.

[0222] S5. Real-time monitoring and judgment of whether the vacuum pipeline and connectors are damaged during installation. The basic image recognition algorithm group is used to analyze the images collected on-site. Due to the significant color difference between the outer and inner coatings of the vacuum pipeline, the basic image recognition algorithm identifies characteristic colors and other color differences between the on-site images and the interior of the vacuum pipeline to determine whether any algorithm-sensitive color differences have occurred during the current installation process, thus assessing the extent of damage to the vacuum pipeline. If damage is detected during installation, the corresponding construction information is reported to the supervisor and awaits processing. After processing, the installation of the vacuum pipeline is repeated, and the judgment in step S5 is repeated. If no damage is found during the installation process, proceed to step S6.

[0223] S6. Real-time monitoring of the reliability of the connection points between the vacuum pipeline and the drainage board at various joints. This is performed after the current installation of the vacuum pipeline and drainage board is completed. Due to significant color differences between the various joints of the vacuum system pipeline and the drainage board, a basic image recognition algorithm can be used to extract high-definition images of the connection points and compare them with standard installation images. Simultaneously, a coordinate reference recognition algorithm is used to verify the installation coordinates of each connection point against the construction design data to determine if they are correct. If the connection image or connection point coordinates of the vacuum pipeline and drainage board are abnormal, the connection is considered unreliable. The corresponding construction information is then reported to the supervisor and awaits processing. After processing, the installation of the pipeline connection point and the judgment in step S6 are repeated. If the connection image or connection point coordinates of the vacuum pipeline and drainage board are normal, the connection is considered reliable, and step S7 continues.

[0224] S7. Receive the start command for installation and construction monitoring of the under-membrane vacuum probe from the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the installation and construction of each pipeline and connector of the vacuum booster system.

[0225] S8. Monitor the construction process of the vacuum pipeline system in real time. Analyze the images collected on-site using the basic image recognition algorithm group. Utilizing the basic image recognition algorithm group and coordinate reference recognition algorithm, calculate the installation coordinates of various probes such as vacuum probes, pressure gauges, and settlement markers during construction and installation by analyzing the color difference of equipment materials and the coordinates of positioning benchmarks. Verify and judge these coordinates against the design data. If the installation coordinates of the probes do not match the design information, notify the supervisor and wait for processing. After processing, re-install the current probe equipment and repeat the judgment in step S8. If the current probe equipment coordinates are accurate, continue to step S9.

[0226] S9. After the installation of each probe device is completed, based on the characteristics of different probes and the data exchange method, read the initial degree of each probe through the data exchange equipment or high-definition image recognition to determine whether the initial degree of each probe is abnormal, and to avoid probe distortion and data fraud. If the initial reading of the probe is abnormal, the corresponding construction information will be reported to the supervisor and await processing. After processing, the installation of the current probe device and the judgment in step S9 will be carried out again. If the initial reading of the current probe is not abnormal, proceed to step S10.

[0227] S10. After the vacuum booster system and the probe under the membrane are installed, receive the vacuum booster system installation completion command input from the vacuum booster system installation and construction monitoring module, and end the real-time monitoring of the vacuum system construction by the above algorithm.

[0228] S11. Organize and record early warning data, key step video images, and other data during the construction of the vacuum booster system in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

[0229] (xv) Construction data recording module

[0230] The construction data recording module includes a data reading submodule, a data linking submodule, a data display submodule, and a report generation submodule. It is used to collect and display the construction monitoring and process data records of vacuum preloading construction from various construction monitoring modules and algorithms. This facilitates project operation and maintenance management personnel in managing project quality and progress, as well as querying historical data, providing a basis for project management and traceability data for subsequent project operation, maintenance, and renovation.

[0231] (15.1) Data Reading Submodule

[0232] The data reading submodule reads and organizes the on-site construction data obtained by each construction monitoring module according to the vacuum preloading construction sequence. This includes key construction and installation images, positioning markers, drainage boards, vacuum pipelines, and other equipment and materials video data at critical construction time nodes under each process. Each construction monitoring module uses corresponding algorithms to identify and judge the construction process images, the handling methods, handling times, handling personnel, and corresponding video and image data, etc., and classifies and organizes the above data to facilitate data association and display.

[0233] (15.2) Related Display Submodule

[0234] The associated display submodule is used to associate the on-site collected data and construction result data of each engineering process after being sorted and classified by the data reading submodule with the construction BIM model obtained by on-site scanning and design modeling, and to provide a visual program interface for three-dimensional visualization display of various construction images, models and record data.

[0235] (15.3) Report generation submodule

[0236] The report generation submodule is used to extract data from the information content sorted and classified by the data reading submodule, organize the construction data and records of each vacuum preloading process according to the construction procedures, key points of construction monitoring, overall construction plan and other key points, and automatically generate the project vacuum preloading construction monitoring work report according to the report template file framework.

[0237] The following example, using vacuum preloading construction of a hydropower project, illustrates the specific implementation steps of a foundation vacuum preloading construction monitoring method based on image recognition technology according to the present invention:

[0238] Step 1: Before the actual application of the project, complete the research and development of the various construction monitoring modules and supporting algorithms involved in the invention, and complete the relevant tests of image recognition and coordinate calculation of the algorithms, vacuum preloading equipment and construction materials.

[0239] Step 2: Deploy and install the application function system platform, which includes the various construction monitoring modules and supporting algorithms involved in this invention, on the data center server or the professional server at the vacuum preloading construction site of the corresponding project.

[0240] Step 3: According to the construction plan, at the site where vacuum preloading construction is to be carried out, install construction positioning equipment such as positioning base stations and positioning poles according to the requirements of the benchmark layout algorithm, and install positioning tags on the corresponding construction equipment according to the construction procedures.

[0241] Step 4: According to the construction plan, install cameras and lifting rods at the site where vacuum preloading is to be carried out, according to the high-definition camera group layout algorithm, and use a 3D scanner to scan the 3D model of the vacuum preloading site as the project progress plan and data needs.

[0242] Step 5: Based on the project schedule, distribute smart safety helmets to all construction workers before the start of vacuum preloading construction and install laser positioning devices around the construction site.

[0243] Step 6: On the server, calibrate and test all equipment involved in vacuum preloading monitoring using the functions and algorithms of each construction monitoring module on the system platform. This includes: testing the functions and information feedback of each camera and lifting rod in the engineering data acquisition equipment; reading and calibrating information from each construction positioning device; calibrating the positioning and voice feedback of the smart safety helmet; and testing the automatic remote operation of the laser positioning device. This ensures that the functions of each construction monitoring module can effectively control and interact with various existing construction monitoring auxiliary equipment.

[0244] Step 7: According to the marking method of construction equipment and materials, and in accordance with the actual vacuum preloading construction plan, before the start of the corresponding construction process, the geotextile and sealing membrane, drainage board, vacuum pipe, vacuum probe and pressure gauge, positioning mark and settlement meter are painted in the manner and color required by this invention.

[0245] Step 8: According to the vacuum preloading construction plan, receive the order to start the laying of geotextile and sealing membrane. Use the geotextile and sealing membrane laying monitoring module and algorithm to judge the compliance of the construction process. If the construction details are found to be inconsistent with the relevant requirements, the corresponding construction information will be reported to the supervisor and await processing and recording.

[0246] Step 9: According to the vacuum preloading construction plan, receive the construction start command for the installation of drainage boards, and use the soil installation drainage board construction monitoring module and algorithm to judge the compliance of the construction process. If it is found that the construction details do not meet the relevant requirements, the corresponding construction information will be reported to the supervisor and await processing and recording.

[0247] Step 10: According to the vacuum preloading construction plan, receive the construction start command for the installation of drainage boards, and use the vacuum pressurization system to install the construction monitoring module and algorithm to judge the compliance of the construction process. If the construction details are found to be inconsistent with the relevant requirements, the corresponding construction information will be reported to the supervisor and await processing and recording.

[0248] Step 11: After the completion of each construction process, the construction data recording module will collect key construction and installation images, positioning markers, drainage boards, vacuum pipelines, and other equipment and materials at critical construction time points under each process. The construction monitoring module will also collect and record the results of image recognition and judgment of each construction process, the handling methods, handling times, handling personnel, and corresponding video and image data. This will facilitate the project operation and maintenance management personnel in managing project quality and progress, as well as querying historical data. It will provide a basis for project management and provide traceability data for the operation, maintenance, and renovation of the project in the future.

[0249] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A foundation vacuum preloading construction monitoring system based on image recognition technology, characterized in that, It includes engineering data acquisition equipment capable of data exchange, construction positioning equipment, auxiliary construction equipment, geotextile and sealing membrane laying construction monitoring module and algorithm module, drainage board installation construction monitoring module and algorithm module, vacuum pressurization system installation construction monitoring module and algorithm module, construction data recording module, and modules for the layout and marking of various construction equipment, as well as coordinate reference recognition algorithm module, high-definition camera group layout algorithm module, benchmark layout algorithm module, and callable basic image recognition algorithm group module; The engineering information acquisition equipment group includes a high-definition camera group, a camera lifting rod, and a 3D laser scanner, which are used to acquire high-definition video images and engineering model information on site. The acquired image information will be used by various engineering construction supervision modules for image recognition and data analysis of the construction process; the acquired model information will be used to connect various engineering data for data querying in corresponding functional modules. The high-definition camera group layout algorithm module is used to obtain the layout method of the high-definition camera group in the vacuum preloading construction site. The algorithm will combine the camera performance parameters to use the minimum number of cameras while ensuring clear and reliable image data. The construction positioning equipment includes a construction area positioning network composed of positioning benchmarks, positioning base stations, and positioning tags. It is used to assist the AI ​​image recognition function in recognizing the construction process images of each vacuum preloading process, so as to increase the image recognition accuracy, reduce the image recognition difficulty, improve the running efficiency of each module program, and reduce energy consumption. The benchmark placement algorithm module is used to specify and constrain the placement interval and number of positioning benchmarks required for the project, and to provide a highly recognizable benchmark positioning network for the coordinate reference recognition algorithm, so that the coordinate reference recognition algorithm can obtain the positioning coordinates of each construction material and equipment through multi-angle image recognition reference calculation. The construction equipment and material marking module uses different colored construction coatings that are non-toxic, harmless, waterproof, and corrosion-resistant to mark the key equipment and materials in different construction stages of vacuum preloading in a corresponding way, so as to improve the image recognition efficiency of the module and algorithm and reduce the recognition difficulty and cost of the algorithm. The basic image recognition algorithm is a set of basic image recognition algorithms that can be called from outside by various functional modules in the system platform. The basic image recognition algorithm module has various basic image analysis and recognition functions, identifies various equipment and materials related to vacuum preloading construction from images collected on site, and assists each module in realizing the required basic image recognition functions and obtaining relevant data. The coordinate reference recognition algorithm module uses the known coordinate positioning benchmarks set up at the construction site as positioning references. Then, through high-definition images taken from multiple angles, it compares and calculates the spatial position coordinates of each construction equipment and material in the images to obtain the calculated coordinate values ​​of each equipment and material. This provides a basis for each module to judge the compliance of the installation and construction of the on-site process. The auxiliary construction equipment includes a smart safety helmet and a laser positioning device. Through voice and laser positioning markers, it assists on-site personnel in carrying out each process of vacuum preloading, while collecting and feeding back on-site construction information. The collected data serves as the basis for image recognition and program command issuance by various functional modules, and also serves as information for construction process records and receipts, which are associated with the corresponding three-dimensional engineering model. The geotextile and sealing membrane laying construction monitoring module provides an interactive program interface and calls geotextile and sealing membrane laying construction monitoring algorithm, coordinate reference recognition algorithm, foundation image recognition algorithm group and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the geotextile and sealing membrane laying process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervision personnel for on-site processing. The geotextile and sealing membrane laying construction monitoring algorithm module, according to the requirements and commands of the corresponding module, calls the algorithm and resources to identify and detect the compliance of various construction equipment and materials involved in the geotextile and sealing membrane laying construction process, and feeds back and records the corresponding information. The drainage board installation monitoring module provides an interactive program interface and calls the drainage board installation monitoring algorithm, coordinate reference recognition algorithm, basic image recognition algorithm group and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the drainage board installation process in vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervisors for on-site processing. The drainage board installation construction monitoring algorithm module, according to the corresponding module requirements and command call algorithm and resource pairs, identifies and detects the compliance of various construction equipment and materials involved in the drainage board installation construction process through pixel and image judgment and coordinate calculation, and feeds back and records the corresponding information. The vacuum booster system installation and construction monitoring module provides an interactive program interface and calls upon the vacuum booster system installation and construction monitoring algorithm, coordinate reference recognition algorithm, basic image recognition algorithm group, and high-definition camera group resources. According to the corresponding construction process and progress, it monitors and manages the vacuum booster system installation and construction procedures in the vacuum preloading construction, and automatically publishes abnormal construction information that does not meet the requirements to the construction data recording module, and simultaneously sends messages to the project supervisors for on-site handling. The vacuum booster system installation and construction monitoring algorithm module, according to the corresponding module requirements and command call algorithm and resource pairs, identifies and detects the compliance of various construction equipment and materials involved in the vacuum booster system installation and construction process through pixel and image judgment and coordinate calculation methods, and feeds back and records the corresponding information. The construction data recording module includes a data reading submodule, a data linking submodule, a data display submodule, and a report generation submodule. It is used to collect and display the construction monitoring and process data records of vacuum preloading construction by various construction monitoring modules and algorithms, so as to facilitate the project operation and maintenance management personnel to manage the project quality and progress and query historical data.

2. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The engineering data acquisition equipment includes a high-definition camera group, which consists of a monitoring network of multiple cameras arranged at the vacuum preloading construction site, and each camera is equipped with a positioning tag; it can acquire high-definition images of each stage of vacuum preloading construction for subsequent identification and judgment by various construction monitoring function modules; The camera lifting rod is used to support each high-definition camera. The rod is equipped with a transmission device and a signal receiving device, which can adjust the height of the camera according to the requirements of the coordinate reference recognition algorithm and collect data information from the required height. The maximum height of the camera lifting rod is related to the size of the vacuum preloading construction site and the accuracy of the camera. The 3D scanner is deployed at the vacuum preloading construction site or carried by the construction and installation personnel. It receives feedback signals through laser ranging technology, converts the actual spatial information into a 3D model, and obtains the on-site model of the vacuum preloading construction site and equipment. This model is used to carry the data of subsequent construction monitoring modules and provides an intuitive data query and operation command release interface. In the construction data recording module, the data reading submodule reads and organizes the on-site construction data obtained by each construction monitoring module according to the vacuum preloading construction sequence, and classifies and organizes various data during the vacuum preloading construction process to facilitate data association and display. The associated display submodule is used to associate the on-site data collected from each engineering process and the construction results data after being sorted and classified by the data reading submodule with the corresponding construction BIM models obtained by on-site scanning and design modeling, and to provide a visual program interface for three-dimensional visualization display of various construction images, models and record data. The report generation submodule is used to extract data from the information content sorted and categorized by the data reading submodule, organize the construction data and records of each vacuum preloading process according to the construction procedures, key points of construction monitoring, and key points of the overall construction plan, and automatically generate a vacuum preloading construction monitoring work report for this project based on the report template file framework.

3. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The algorithm flow of the high-definition camera group deployment algorithm module is as follows: S1. Read the outer contour line of the area to be vacuum preloaded from the corresponding construction details or BIM model; S2. Calculate the number and spacing of cameras along the outer contour line using the following formula: In the formula The cameras are spaced out along the outer contour line of the construction site. This is the shape adjustment coefficient for the vacuum preloading area; This is the maximum pixel count of the camera; This refers to the horizontal resolution coefficient. This refers to the lens's light-gathering size; This is the minimum illumination resolution coefficient for the camera; In the formula The spacing of cameras within the construction area; The cameras are spaced out along the outer contour line of the construction site. This refers to the camera's wide-angle coefficient. S3. Based on the camera arrangement intervals along the outer contour line of the construction site and the camera arrangement intervals within the construction area, the construction site is divided into grids to generate a high-definition camera group layout plan. S4. Generate a camera layout plan and count the number of cameras and camera lifting poles for use in subsequent module processes and algorithms.

4. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, In the engineering data acquisition equipment, the positioning pole consists of a pole body, a pole head, and a positioning tag fixing component. It is installed in the vacuum preloading construction area according to the pole layout algorithm requirements. The pole head is painted with a conspicuous color that is clearly different from the construction ground, geotextile, vacuum membrane, and various construction equipment and personnel to facilitate camera recognition and algorithm program recognition and positioning of image content. At the same time, a positioning tag that can communicate with the positioning base station is installed to provide basic parameters for each module to call the coordinate reference recognition algorithm for calculation. The positioning base station uses UWB positioning technology to locate the coordinates of the positioning tags in the vacuum preloading construction equipment and positioning poles. The positioning base station is deployed in the vacuum preloading construction site. It receives tag signals through the wireless network in the project area and calculates the distance and angle of each device and the positioning tag device in the positioning pole according to the arrival time, arrival time difference and arrival angle algorithm, and obtains the corresponding coordinate information. The positioning tags are installed in the drainage board insertion machine, the positioning pole vacuum preloading construction equipment and auxiliary materials. They can respond to the wireless signals emitted by the positioning base station and obtain the spatial coordinates of each positioning tag and the corresponding equipment through the information feedback time difference. They are used for the coordinate positioning of each equipment and material and the display of the model in three-dimensional space. The algorithm flow of the benchmark placement algorithm module is as follows: S1. Read or input actual parameter data of the vacuum preloading construction site and materials of the project, as well as data from engineering drawings and models; S2. Calculate the installation interval of the positioning markers. The calculation formula is as follows: In the formula The spacing for positioning marker installation; Adjust the coefficients according to the required positioning coordinate accuracy; The area of ​​the vacuum preloading construction site; This refers to the minimum area of ​​a single piece of geotextile and sealing membrane used; This refers to the number of drain pipe connectors; This is the sum of the number of horizontal and vertical pipes in the pressurization pipeline system; Boost system accuracy calibration coefficient; S3. Based on the calculated installation interval of the positioning markers, the vacuum preloading construction site is divided into grids, and a positioning marker installation layout diagram is generated. S4. The generated results guide the construction site in completing the installation and layout of each positioning marker.

5. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, In the construction equipment and material marking module, the geotextile and sealing membrane marking method is used to mark the woven geotextile, non-woven geotextile, and vacuum sealing membrane used in vacuum preloading construction with color coating; the selected coating must have a significant color difference from the laying layer where the geotextile or sealing membrane is located and the soil in the construction area to improve the efficiency and accuracy of automatic image recognition. The drainage board marking method is used to mark the drainage boards required for the vacuum preloading construction process by coloring them. The drainage channels on the outside and inside of the drainage board need to be colored differently, and there should be obvious differences between the outside and inside of the drainage board and the woven geotextile laid in the current process. At the same time, the length scale and numbers will be marked on the drainage board to automatically identify the insertion depth of each drainage board. The vacuum pipeline marking method is used to color-mark the vacuum system pipelines during vacuum preloading construction. The inner and outer sides of the vacuum pipelines and the connectors with the drainage board are colored differently, and it is ensured that the color of each construction material is significantly different from the woven geotextile laid in the current process. This enables the use of algorithms to identify misaligned vacuum pipeline layouts, misaligned drainage board connectors, and various vacuum system layout construction problems. The vacuum probe and pressure gauge marking method is used to color-mark the vacuum probe and pressure gauge monitoring equipment required in vacuum preloading construction; the paint selected for the vacuum probe and pressure gauge should have a more obvious color difference from the laying layer where the geotextile or sealing membrane is located and the soil in the construction area; so that the relevant image recognition algorithm can more efficiently identify the relevant monitoring equipment from the image, and facilitate the recognition and recording by the algorithm and program functions; The marking method for the positioning pole and settlement meter is used to color-mark the head of the positioning pole and the outer shell of the settlement meter; when the positioning pole is installed, it is inserted into the vacuum preloading construction site and the exposed head is colored.

6. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The algorithm flow of the coordinate reference recognition algorithm module is as follows: S1. Receive commands from various functional modules requesting coordinates of construction equipment and materials; S2. Call the basic image recognition algorithm to analyze and identify the target pixel color and the presented pattern, and initially search for the target construction equipment and materials that need to be located in the image, as well as the nearest 1-3 sets of positioning benchmarks within the target range; S3. By analyzing the patterns captured by each camera, sort them according to the target pixel color and the size of the pattern in the captured image, and sort each camera in reverse order according to its distance from the target device and material. S4. Set two camera height levels, high and low. Based on the order of camera distance from the target, alternately set the camera height and adjust the height of each camera to collect image information of the target construction equipment and materials from different heights and angles, providing a basis for subsequent coordinate calculations. S5. Using the current camera coordinates and the coordinates of the nearest positioning benchmark to the target construction equipment and materials in the image, calculate the coordinates of the target construction equipment and materials in the images captured by each camera. The calculation formula is as follows: Y, Z axis coordinates The calculation formula is the same; In the formula To obtain the initial X-axis coordinate values ​​of the target construction equipment and materials based on the images captured by each camera; The X-axis coordinate of the nearest positioning benchmark in the image captured by the current camera that is clearly observable and identifiable of the target construction equipment and materials; This refers to the difference in X-axis coordinates between the nearest positioning benchmark and the target construction equipment and materials, calculated based on the images captured by the current camera. This is the camera adjustment factor; Adjust the weighting coefficients for each camera position; The maximum angle between the target plane and the plane containing the observed target; This is the distance between the current camera and the nearest positioning marker for the target construction equipment and materials; Adjustment parameters are calculated for the weights; The X-axis coordinates of the target construction equipment and materials are obtained using the algorithm; N is the total number of lenses used to capture images of this target construction equipment and materials. The coordinate calculation is similar; S6. Obtain the space for each target construction equipment and material ( After obtaining the coordinates, the spatial location of the corresponding equipment and materials is marked and displayed in the 3D model, and the coordinate data is fed back to each functional module and algorithm as the basis for subsequent modules to judge the compliance of construction procedures and monitor construction.

7. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The auxiliary construction equipment includes a smart safety helmet with a built-in camera, positioning tag, earphone, microphone, interactive button, and signal transceiver. It can receive operation instructions issued by the system platform through the earphone; it can provide feedback on its own status and issue commands through the microphone and interactive button; it can collect and record on-site information according to the needs of each module and algorithm through the camera; and it can provide the wearer's coordinate information in real time through the built-in positioning tag and on-site positioning base station, providing a data foundation for each functional module. The laser positioning device has a built-in laser transmitter, signal transceiver, positioning tag, and transmission structure. At the construction site, it uses the positioning base station and the internal positioning tag to obtain the spatial coordinates of the laser positioning device, receives control commands issued by each construction module, and marks the installation positions of vacuum preloading construction equipment and materials for each process at the construction site through the transmission structure and laser transmitter, guiding on-site personnel to carry out construction.

8. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The algorithm flow of the geotextile and sealing membrane laying construction monitoring algorithm module is as follows: S1. Read the basic information of the vacuum preloading construction of the project, including the vacuum preloading design and construction model information and the drawing information obtained by analysis; S2. Read various types of on-site collected data and call algorithms for recognition, including high-definition image data collected by the on-site camera group and three-dimensional scanning of the engineering site model, and use the basic image recognition algorithm group to identify the overall construction area of ​​the engineering vacuum preloading and the location of each control point in the image; S3. Receive the input command to start the construction monitoring of the positioning poles from the program interface of the geotextile and sealing membrane laying monitoring module, call the basic image recognition algorithm group to monitor and identify each positioning pole in the on-site images collected during the construction process in real time, and mark them. S4. Make a judgment. If the judgment finds that the current positioning benchmark coordinate data does not match the drawing and model information, the corresponding benchmark position and coordinate information will be reported to the supervisor through the smart safety helmet and wait for modification. After the modification is completed, the judgment in step S4 will be repeated. If the judgment finds that each positioning benchmark matches the drawing and model information, then continue to step S5. S5. Receive the input command to start geotextile laying monitoring from the geotextile and sealing membrane laying monitoring module program interface, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the laying and splicing of each geotextile. S6. Monitor and determine in real time whether there are any debris unrelated to the construction on the construction site and on the geotextile during the geotextile laying process; if there are debris during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for on-site care. After the processing is completed, return to step S6 to re-determine; if it is determined that there are no wrinkles during the geotextile laying process, continue to step S7. S7. Monitor and determine in real time whether there are any wrinkles or deformations in the geotextile during the laying process; if there are wrinkles in the geotextile during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for processing. After processing is completed, return to step S6 to re-determine; if it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S8. S8. Monitor and determine in real time whether there is any damage to the geotextile during the geotextile laying process. If there is any damage to the geotextile during the laying process, the algorithm will notify the supervisor of the construction risk information through the smart safety helmet and wait for processing. After processing is completed, return to step S6 to make a new judgment. If it is determined that there are no wrinkles in the geotextile during the laying process, continue to step S9. S9. After each step of the geotextile laying construction is completed, receive the input command from the geotextile and sealing membrane laying construction monitoring module program interface indicating that the geotextile laying construction monitoring is complete, and end the real-time monitoring of the geotextile laying construction by the above algorithm. S10. Organize and record the early warning data during the geotextile laying process in this procedure, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

9. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The algorithm flow of the drainage board installation construction monitoring algorithm module is as follows: S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, geotextile laying construction status and process processing data. S2. Read and detect the positioning coordinates on the drainage board installation equipment, and verify whether the coordinate positions of each positioning tag in the algorithm program are accurate and consistent with the actual site conditions. S3. Analyze the construction design drawings or model file data to obtain the coordinates of the insertion positions in each drainage board construction design scheme. S4. Receive the drainage board installation construction monitoring start command input by the drainage board installation construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring during the installation of each drainage board. S5. Real-time monitoring uses the coordinate information provided by the positioning tag of the drainage board installation equipment to initially determine whether the current drainage board installation coordinates are consistent with the design and construction documents. At the same time, the coordinate reference recognition algorithm is called to further determine the accuracy of the current drainage board installation coordinates. If the current drainage board installation coordinates are inconsistent with the design and construction plan information, the corresponding construction information is reported to the supervisor and awaits processing. After processing is completed, return to step S5 to re-determine; if the current drainage board installation coordinates are consistent with the design and construction plan information, continue to step S6. S6. Monitor and determine in real time whether the drainage board is shifted during the insertion process. Use the basic image recognition algorithm group to identify the drainage board image in the high-definition image during the insertion process and make a judgment. If the judgment finds that the drainage board has shifted during the insertion process, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the drainage board is inserted again and the judgment in step S6 is repeated. If no shift is found in the current drainage board during the insertion process, continue to step S7. S7. Monitor and determine in real time whether the drainage board is damaged during the insertion process. Call the basic image recognition algorithm group to analyze the images collected on site and determine whether there are any color differences that the algorithm is sensitive to during the entire installation process of the drainage board, so as to determine the damage of the drainage board. If the drainage board is found to be damaged during the insertion process, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation of the drainage board and the judgment in step S7 are repeated. If no damage is found during the installation process of the drainage board, proceed to step S8. S8. After the current drainage board is installed, determine whether the external leakage of the drainage board meets the requirements. By calling the basic image recognition algorithm group to identify the drainage board graphic and the scale value marked on the drainage board from the on-site image, the judgment is made. If the external leakage scale value is not within the range specified in the construction requirements, it is considered as not meeting the requirements. The corresponding construction information is reported to the supervisor and awaits processing. After processing is completed, the judgment in step S8 is repeated. If the external leakage scale value is within the range specified in the construction requirements, it is considered as meeting the requirements. Then, return to step S5 to start the construction monitoring program algorithm loop for the next drainage board. If the installation of all drainage boards has been completed, continue to step S9. S9. When the on-site personnel confirm that the installation of each drainage board is completed and there are no abnormalities, they shall accept the input command that the installation of the drainage board is completed from the program interface of the drainage board installation monitoring module, and end the real-time monitoring of the drainage board installation by the above algorithm. S10. Organize and record the early warning data and key step video image data during the installation of drainage boards in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

10. The foundation vacuum preloading construction monitoring system based on image recognition technology according to claim 1, characterized in that, The algorithm flow of the vacuum booster system installation and construction monitoring algorithm module is as follows: S1. Read the basic information of each preceding construction process and project, as well as various construction process information, including engineering design drawings or model information, engineering construction layout, coordinates of vacuum preloading construction control points, coordinates of positioning benchmarks of preceding construction, and the operation status and process processing data of preceding construction. S2. Analyze the construction design drawings or model file data to obtain the coordinates of the vacuum piping system layout, the coordinates of the connection joints with the drainage board, and the installation coordinates of the main connecting pipe fittings. S3. Receive the vacuum booster system installation start command input by the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring of the installation and construction of each pipeline and connector of the vacuum booster system during the installation and construction process; S4. Monitor the vacuum pipeline construction process in real time, call the basic image recognition algorithm group to analyze the images collected on site, and use the basic image recognition algorithm group and coordinate reference recognition algorithm to identify the coordinates of the first, middle and last sections of each vacuum pipeline and compare them with the construction design information. If the position and direction coordinates of each pipeline are found to be off from the design plan during construction, the corresponding construction information is reported to the supervisor and awaits processing. After processing, the installation of the vacuum pipeline section and the judgment in step S4 are carried out again. If no damage is found during the installation of the drainage board, proceed to step S5. S5. Monitor and determine in real time whether the vacuum pipeline and connectors are damaged during installation. Analyze the images collected on-site using the basic image recognition algorithm group. If damage to the vacuum pipeline is found during installation, notify the supervisory personnel of the corresponding construction information and wait for processing. After processing, re-install the vacuum pipeline and repeat the judgment in step S5. If no damage is found during the installation of the vacuum pipeline, continue to step S6. S6. Monitor in real time whether the connection between the vacuum pipeline and the drainage board of various types of joints is reliable. If there is any abnormality in the connection image or connection point coordinates between the vacuum pipeline and the drainage board, it is considered that the connection between the vacuum pipeline and the drainage board is unreliable. The corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation construction of the pipeline connection and the judgment in step S6 are carried out again. If there is no abnormality in the connection image or connection point coordinates between the vacuum pipeline and the drainage board, it is considered that the connection between the vacuum pipeline and the drainage board is reliable, and the process continues to step S7. S7. Receive the start command for installation and construction monitoring of the under-membrane vacuum probe from the vacuum booster system installation and construction monitoring module, and call the basic image recognition algorithm group and coordinate reference recognition algorithm to perform real-time monitoring of the installation and construction of each pipeline and connector of the vacuum booster system. S8. Monitor the construction process of the vacuum pipeline system in real time, and call the basic image recognition algorithm group to analyze the images collected on site; if the installation coordinates of the probe do not match the design information, the corresponding construction information will be reported to the supervisor and wait for processing. After processing, the installation of the current probe equipment and the judgment of step S8 will be carried out again; if the coordinates of the current probe equipment are accurate, continue to step S9. S9. After the installation of each probe device is completed, according to the different probe characteristics and data exchange methods, the initial degree of each probe is read through the data exchange equipment or high-definition image recognition to determine whether the initial degree of each probe is abnormal, so as to avoid probe distortion and data fraud. If the initial reading of the probe is abnormal, the corresponding construction information is reported to the supervisor and the process is awaited. After the process is completed, the installation of the current probe device and the judgment in step S9 are repeated. If the initial reading of the current probe is not abnormal, the process continues to step S10. S10. After the vacuum booster system and the probe under the membrane are installed, receive the vacuum booster system installation completion command input by the vacuum booster system installation and construction monitoring module, and end the real-time monitoring of the vacuum system construction by the above algorithm. S11. Organize and record the early warning data and key step video image data during the construction of the vacuum booster system in this process, and publish the relevant data to the construction data recording module for subsequent data analysis and query.

Images