Automated acquisition method and system for multi-element information of tunnel during construction

Abstract

The present invention relates to the technical field of tunnel monitoring during construction, in particular to an automated acquisition method and system for multi-element information of a tunnel during construction. The method comprises: projecting light sources on the inner wall of a tunnel to form detection points and detection lines, then using measurement cameras to record changes in the pixel positions of a region where the detection points and the detection lines are located in different time periods, and using the detection points and the detection lines as reference benchmarks to integrate point cloud data of the tunnel and correct same, so as to form an accurate four-dimensional model of the tunnel. Thus, the present invention allows for dynamic and omnibearing observation and recording of the structural state and the construction progress of tunnels, thus achieving automated and intelligent collection of tunnel information, saving manpower, and improving tunnel construction efficiency.

Description

A method and system for automatically collecting multi-dimensional information of tunnels during construction Technical Field

[0001] The present invention relates to the technical field of tunnel monitoring during construction, and in particular to a method and system for automatically collecting multivariate information of tunnels during construction. Background Art

[0002] After initial support construction, monitoring measurements are needed to obtain deformation data from the initial support and to control the quality of tunnel construction. Currently, expensive total stations combined with targets are often used periodically during tunnel construction to monitor deformation at fixed points on the tunnel vault and sidewalls.

[0003] On the one hand, this periodic and intermittent method of monitoring deformation inside the tunnel cannot guarantee the real-time update of deformation data and cannot provide timely warning to engineering personnel;

[0004] On the other hand, the monitoring points of the target are discontinuous, with long longitudinal spacing and sparse cross-sectional points. It is impossible to measure the deformation of the continuous section of the tunnel.

[0005] In summary, the traditional tunnel deformation monitoring method has high equipment costs, cannot be monitored in real time, and has a limited monitoring range that cannot fully reflect the deformation parameters of the tunnel inner wall, which is not conducive to engineering personnel's control over construction quality and safety.

[0006] In view of this, in order to improve the above technical problems, the present invention provides a method and system for automatically collecting multi-dimensional information of a tunnel during the construction period.

[0007] Summary of the Invention

[0008] In a first aspect, the present invention provides a multi-element information automatic collection system for a tunnel during construction, comprising a plurality of light sources, wherein the light sources are arranged on the sidewalls of the tunnel and form detection points and / or detection lines on the inner wall of the tunnel;

[0009] A plurality of measuring cameras are arranged around the detection points and / or detection lines, and the measuring cameras are used to collect and record image data of the detection points and / or detection line areas in real time, and the shooting angle range formed by the combination of the plurality of measuring cameras covers all surfaces of the detection points and / or detection line areas;

[0010] a data processor, wherein a signal input terminal of the data processor is connected to the measurement camera, and is configured to receive image data collected by the measurement camera in real time, and respectively calculate the three-dimensional coordinates of the region where the detection points and / or detection lines are located using the detection points and / or detection lines as position references, and calculate the size and shape of the region where the detection points and / or detection lines are located using image data of the region where the detection points and / or detection lines are located obtained from different shooting angles, and then integrate the three-dimensional coordinates, size, and shape information to create low-density point cloud data of the tunnel inner wall;

[0011] An application terminal, wherein the signal input end of the application terminal is connected to the signal output end of the data processor, is used to receive the real-time point cloud data updated and corrected by the data processor, and display the four-dimensional tunnel model established based on the point cloud data. At the same time, based on the feedback of the four-dimensional model, the application terminal displays deformation data of the detection point and / or detection line area, concrete thickness data, tunnel flatness data, anchor number, position data, and tunnel face sketch data.

[0012] Preferably, the light source is a point light source, a line light source, a surface light source or a combination of one or more types.

[0013] In a second aspect, the present invention provides a method for automatically collecting multi-dimensional information of a tunnel during construction. The method adopts the above-mentioned automatic collection system for multi-dimensional information of a tunnel during construction.

[0014] Step 101: Arrange multiple light sources horizontally or vertically on the tunnel sidewalls, and use the light sources to form detection points and / or detection lines on the tunnel vault and sidewalls;

[0015] Step 102: multiple measurement cameras are set up around the detection point and / or detection line, and the multiple measurement cameras are used to capture and collect image data of the entire detection point and / or detection line area in real time from multiple angles at different time periods;

[0016] Step 103: The image data collected in real time by the measurement camera is transmitted to a data processor. The data processor calculates the three-dimensional coordinates of the area where the detection points and / or detection lines are located using the detection points and / or detection lines as position references, and calculates the size and shape of the area where the detection points and / or detection lines are located using image data of the detection points and / or detection lines obtained from different camera angles. The data processor then integrates the three-dimensional coordinates, size, and shape information to create low-density point cloud data of the tunnel inner wall.

[0017] At the same time, an image capture algorithm is used to compare the pixel displacement data and key frame data of the same detection point and / or detection line area at different angles in different time periods frame by frame. The detection point and / or detection line are used as a calibration reference, and the pixel movement data and key frame data of the area where the detection point and / or detection line are located are used to calibrate and update the part of the point cloud data that has moved in real time.

[0018] Step 104: The data processor sends the real-time corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel;

[0019] Step 105: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0020] Step 106: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm, and displays the real-time dynamic deformation data of the tunnel inner wall, the concrete thickness data, and the tunnel flatness data.

[0021] Preferably, the light source in step 101 includes a point light source, and the method for realizing deformation detection of a precise point on the tunnel wall based on the point light source is as follows:

[0022] Step 201: Projecting a point light source onto a point on the tunnel sidewall that needs to be accurately observed to form a detection point;

[0023] Step 202: multiple measurement cameras are set up around the detection point, and the multiple measurement cameras are used to shoot and collect image data of the entire detection point in different time periods from multiple angles;

[0024] Step 203: The image data collected by the measurement camera is transmitted to a data processor. The data processor uses an image capture algorithm to compare the pixel displacement data and key frame data of the same detection point area at different angles in different time periods frame by frame, and calculates the deformation data of the detection point area.

[0025] Step 204: The data processor sends the deformation data of the detection point area and the time axis to the application terminal for displaying the current deformation data of the tunnel sidewall area where the detection point is located.

[0026] Preferably, the light source in step 101 includes a line light source, and the tunnel wall deformation detection is implemented based on the line light source as follows:

[0027] Step 301: multiple line light sources are installed on the tunnel sidewalls along the tunnel length direction to form cross-sectional detection lines with equal or random spacing on the tunnel vault and sidewalls;

[0028] Step 302: multiple measurement cameras are set up around the inspection line, and the multiple measurement cameras are used to capture image data of the entire inspection line area at different time periods from multiple angles;

[0029] Step 303: The image data collected by the measurement camera is transmitted to a data processor, and the three-dimensional coordinates of the area where the detection line is located are calculated using the detection line as a position reference. Image data of the detection line area obtained from different shooting angles are used to calculate the size and shape of the area where the detection line is located. The three-dimensional coordinates, size, and shape information are then integrated to create low-density point cloud data of the tunnel inner wall.

[0030] At the same time, the data processor uses threshold segmentation and centerline extraction algorithms to extract the detection line area, and then fuses and calculates with the original detection line and the historical pixel position data of the area where the detection line is located to obtain pixel movement data and key frame data of the same detection line area at different angles in different time periods; and uses the detection line as a calibration reference, using the pixel movement data and key frame data of the area where the detection line is located to calibrate and update the part of the point cloud data that has moved in real time;

[0031] Step 304: The data processor sends the corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel;

[0032] Step 305: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0033] Step 306: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm, and displays the real-time dynamic deformation data of the vault and sidewalls of the tunnel where the detection line is located.

[0034] Preferably, the light source in step 101 further includes a surface light source, and the tunnel wall deformation detection is implemented based on the surface light source as follows:

[0035] Step 401: multiple surface light sources are installed on the tunnel sidewalls along the tunnel length direction to form a grid detection line on the tunnel vault and sidewalls;

[0036] Step 402: multiple measurement cameras are set up around the grid detection line, and the multiple measurement cameras are used to capture image data of the entire grid detection line area at different time periods from multiple angles;

[0037] Step 403: The image data collected by the measurement camera is transmitted to a data processor, and the three-dimensional coordinates of the area where the grid detection lines are located are calculated using the grid detection lines as a position reference. The size and shape of the area where the detection lines are located are calculated using the image data of the grid detection lines obtained at different shooting angles. The three-dimensional coordinates, size, and shape information are then integrated with the pixel positions of the grid detection lines in the image data to create low-density point cloud data of the tunnel inner wall.

[0038] At the same time, the data processor uses a deep learning image segmentation algorithm to extract the intersection of the grid detection lines as capture points, obtains pixel movement data and key frame data of the same capture point area at different angles and in different time periods, and calculates the pixel displacement data of the detection line area between two adjacent capture points based on the capture point. Then, the three-dimensional coordinates, size, and shape information are integrated to create low-density point cloud data of the tunnel inner wall.

[0039] Step 404: Send the corrected point cloud data to the application terminal and convert it into a four-dimensional model of the tunnel;

[0040] Step 405: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0041] Step 406: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm to display the overall dynamic deformation parameters of the tunnel vault and sidewall space.

[0042] Preferably, the point cloud data corrected in step 104 can also be used to calculate the thickness of the tunnel inner wall concrete. The specific calculation method is as follows:

[0043] Step 501: setting the grid detection line as a reference line;

[0044] Step 502: Calculate the distance between the intersection of the normal vector of each cloud point in the point cloud data and the baseline;

[0045] Step 503: retrieve the cloud point location data and match it with the intersection distance data;

[0046] Step 504: Synchronize and match the intersection distance data to the three-dimensional tunnel model based on the cloud point position data to obtain the concrete thickness at different locations in the tunnel.

[0047] Preferably, the four-dimensional tunnel model generated in step 105 can also be used to measure the flatness of the tunnel inner wall. The specific measurement method is as follows:

[0048] Step 601: using the detection point and / or detection line as a position reference point, and using a centerline capture algorithm to obtain the centerline of the three-dimensional tunnel model;

[0049] Step 602: Expand the entire 3D tunnel model along the center line of the 3D tunnel model to generate an orthogonal image;

[0050] Step 603: using a 3D measurement algorithm to obtain the elevation difference between the cloud point normal vector in the orthogonal image and the detection point and / or detection line on the tunnel inner wall surface;

[0051] Step 604: Collect elevation difference data to obtain the flatness of the tunnel inner wall.

[0052] Preferably, the measurement camera data collection method in step 102 can also be used to record the spray anchor support process and the anchor hole position and quantity. The specific recording method is as follows:

[0053] Step 701: Setting up a measurement camera in the tunnel to record the entire bolt support process to generate video data;

[0054] Step 702: The video data collected by the measurement camera is transmitted to a data processor. The three-dimensional coordinates of the area where the detection points and / or detection lines are located are calculated using the detection points and / or detection lines as position references. The size and shape of the area where the detection points and / or detection lines are located are calculated using image data of the area where the detection points and / or detection lines are located obtained from different shooting angles. The three-dimensional coordinates, size, and shape information are then integrated to create low-density point cloud data of the tunnel inner wall.

[0055] At the same time, the data processor uses image recognition technology to automatically identify the anchor rods and their number in the video data, and locates the anchor rods with the help of detection points and / or detection lines. Then, the anchor rods' position and number information are integrated and corrected with the point cloud data again using the detection points and / or detection lines as the position reference.

[0056] Step 703: Send the corrected point cloud data to the application terminal and convert it into a four-dimensional model of the tunnel;

[0057] Step 704: The application displays the location, quantity, and drilling time of each anchor hole in the four-dimensional tunnel model; at the same time, the video data recorded by the measurement camera can also be exported to form a work video summary.

[0058] Preferably, the data collection method of the measurement camera in step 102 can also be used to generate tunnel face sketch data, and the specific generation method is as follows:

[0059] Step 801: multiple measurement cameras are set up in front of the tunnel face at different angles and positions;

[0060] Step 802: The measurement camera captures the image data of the tunnel face in real time;

[0061] Step 803: The image data collected by the measurement camera is transmitted to a data processor. The data processor uses image recognition technology to analyze the tunnel face image data obtained from different shooting angles to obtain data on the type of tunnel face rock, geotechnical integrity, groundwater discharge status, and rock hardness.

[0062] Step 804: The application terminal receives the image information integrated by the data processor and displays a sketch image of the tunnel face annotated with data including face rock type data, geotechnical integrity data, groundwater discharge status data, and rock hardness data.

[0063] The beneficial effects of the present invention are:

[0064] 1. This invention uses a measurement camera in conjunction with detection points and lines to construct low-density point cloud data of the tunnel and convert it into a four-dimensional model. By comparing the dimensions of the four-dimensional model at different time periods, deformation data of the tunnel's inner wall is derived. This replaces the traditional method of monitoring tunnel displacement using mobile measurement targets with total stations, enabling real-time monitoring and feedback of tunnel wall deformation data, effectively ensuring the continuity, integrity, and timeliness of tunnel monitoring.

[0065] 2. This invention uses a measurement camera, detection points, and detection lines to construct low-density point cloud data of the tunnel. The image information collected in real time by the measurement camera is used to update any points in the point cloud data that have moved. This not only transmits the latest point cloud data to the application terminal for real-time generation of an updated 4D tunnel model, but also avoids the need for repeated reconstruction of the entire tunnel point cloud data, thus ensuring the efficiency of the entire computational process.

[0066] 3. The detection points and / or detection lines in the present invention can not only serve as construction lines to assist the measurement camera in capturing the morphology of the monitored area, but also serve as position reference lines to provide reference standards for the synthesis and updating of point cloud data, the positioning of anchor rods, the flattening of tunnel surfaces, and the measurement of concrete thickness, so that the acquired data can accurately correspond to the actual tunnel conditions, resulting in more accurate results.

[0067] 4. This invention utilizes pixel displacement data to partially update and replace point cloud data. The resulting real-time updated 4D model enables more timely and accurate access to tunnel face sketch information, anchor bolt construction progress, anchor bolt placement, concrete thickness and flatness in any area, and inner wall deformation parameters. All of this information can be collected, integrated, and displayed through a single system, enhancing the automation and intelligence of the entire tunnel construction monitoring system, improving overall tunnel construction efficiency, safety, and data accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] FIG1 is a schematic diagram of the system composition structure of the present invention;

[0069] FIG2 is a flow chart of a method for automatically collecting multi-dimensional information during a tunnel construction period provided by the present invention;

[0070] FIG3 is a schematic diagram of a point light source projection according to the present invention;

[0071] FIG4 is a schematic diagram of a process for detecting deformation of a certain point area in a tunnel using a point light source according to the present invention;

[0072] FIG5 is a schematic diagram of a line light source projection according to the present invention;

[0073] FIG6 is a schematic diagram of a process for detecting deformation of a certain section of a tunnel vault and sidewall area using a linear light source according to the present invention;

[0074] FIG7 is a schematic diagram of a surface light source projection according to the present invention;

[0075] FIG8 is a schematic diagram of a process for detecting deformation of the entire inner wall area of ​​a tunnel using a surface light source according to the present invention;

[0076] FIG9 is a schematic diagram of a process for calculating the thickness of tunnel sidewall concrete using point cloud data according to the present invention;

[0077] FIG10 is a schematic diagram of a process for calculating the flatness of a tunnel sidewall using point cloud data according to the present invention;

[0078] FIG11 is a flow chart of the present invention for real-time monitoring of the anchor bolt installation position, quantity, and installation time;

[0079] FIG12 is a schematic diagram of a process for automatically generating tunnel face sketch data according to the present invention;

[0080] Figure 13 is a real-life picture of the grid detection line laboratory display;

[0081] Figure 14 shows the image acquisition of the tunnel face. DETAILED DESCRIPTION

[0082] In order to deepen the understanding of the present invention, the present invention will be further described in detail below with reference to the examples. The examples are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0083] Example 1:

[0084] As shown in FIG1 , this embodiment 1 proposes an automated multi-information collection system for tunnels during construction, comprising a plurality of light sources arranged on the sidewalls of the tunnel and forming detection points and / or detection lines on the inner wall of the tunnel;

[0085] A plurality of measuring cameras are arranged around the detection points and / or detection lines, and the measuring cameras are used to collect and record image data of the detection points and / or detection line areas in real time, and the shooting angle range formed by the combination of the plurality of measuring cameras covers all surfaces of the detection points and / or detection line areas;

[0086] a data processor, wherein a signal input terminal of the data processor is connected to the measurement camera, and is configured to receive image data collected by the measurement camera in real time, and respectively calculate the three-dimensional coordinates of the region where the detection points and / or detection lines are located using the detection points and / or detection lines as position references, and calculate the size and shape of the region where the detection points and / or detection lines are located using image data of the region where the detection points and / or detection lines are located obtained from different shooting angles, and then integrate the three-dimensional coordinates, size, and shape information to create low-density point cloud data of the tunnel inner wall;

[0087] An application terminal, wherein the signal input end of the application terminal is connected to the signal output end of the data processor, is used to receive the real-time point cloud data updated and corrected by the data processor, and display the four-dimensional tunnel model established based on the point cloud data. At the same time, based on the feedback of the four-dimensional model, the application terminal displays deformation data of the detection point and / or detection line area, concrete thickness data, tunnel flatness data, anchor number, position data, and tunnel face sketch data.

[0088] The light source is a point light source, a line light source, a surface light source or a combination of one or more.

[0089] Tunnel engineers can flexibly select appropriate light sources to monitor tunnel deformation according to different monitoring needs. They can then use multiple measurement cameras, preferably three, to capture images of one or more detection points, or one or more detection lines, from three different angles. The three measurement cameras with different angles can capture all surfaces in the area where the detection lines or detection points are located.

[0090] In this way, image data of the detection points or detection lines in different time periods are collected in all directions, avoiding the disadvantage of a single measurement camera with a fixed shooting angle and losing some deformation data within a limited viewing angle range. It also ensures that the data processor can obtain accurate pixel movement data and key frame data of the detection lines or detection points when analyzing the image data.

[0091] Based on the pixel movement data and key frame data of the detection line or detection point, the cloud points that have moved in the low-density point cloud data are updated and replaced in real time. With the assistance of the application terminal, the real-time corrected point cloud data is converted into a four-dimensional model that is easy to retrieve using three-dimensional software. Based on the four-dimensional model, the basic data and deformation parameters of the tunnel vault and side walls in different time periods can be freely observed, retrieved and displayed.

[0092] Then, by integrating the real-time pixel position data collected by the measuring camera with the point cloud data, a four-dimensional model of the tunnel was established. This not only achieved continuous and real-time monitoring of the overall deformation of the tunnel, but the three-dimensional tunnel model formed by the integration of monitoring data can also provide basic parameters for data measurement in later tunnel projects, such as calculation of tunnel side wall concrete thickness and flatness.

[0093] Example 2:

[0094] As shown in FIG2 , this second embodiment proposes an automated method for collecting multi-dimensional information of tunnels during the construction period. The specific collection method is as follows:

[0095] Step 101: Arrange multiple light sources horizontally or vertically on the tunnel sidewalls, and use the light sources to form detection points and / or detection lines on the tunnel vault and sidewalls;

[0096] Step 102: multiple measurement cameras are set up around the detection point and / or detection line, and the multiple measurement cameras are used to capture and collect image data of the entire detection point and / or detection line area in real time from multiple angles at different time periods;

[0097] Step 103: The image data collected in real time by the measurement camera is transmitted to a data processor. The data processor calculates the three-dimensional coordinates of the area where the detection points and / or detection lines are located using the detection points and / or detection lines as position references, and calculates the size and shape of the area where the detection points and / or detection lines are located using image data of the detection points and / or detection lines obtained from different camera angles. The data processor then integrates the three-dimensional coordinates, size, and shape information to create low-density point cloud data of the tunnel inner wall.

[0098] At the same time, an image capture algorithm is used to compare the pixel displacement data and key frame data of the same detection point and / or detection line area at different angles in different time periods frame by frame. The detection point and / or detection line are used as a calibration reference, and the pixel movement data and key frame data of the area where the detection point and / or detection line are located are used to calibrate and update the part of the point cloud data that has moved in real time.

[0099] Step 104: The data processor sends the real-time corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel;

[0100] Step 105: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0101] Step 106: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm, and displays the real-time dynamic deformation data of the tunnel inner wall, the concrete thickness data, and the tunnel flatness data.

[0102] Using the detection points and / or detection lines, reference points and / or reference lines are constructed on the tunnel inner wall. These reference points and / or reference lines are used as alignment benchmarks to integrate the position data of the detection points and / or detection lines monitored in real time by the measurement camera into the point cloud data.

[0103] In the subsequent monitoring process, it is only necessary to use the measurement camera to monitor the area where multiple detection points or multiple detection lines are located in real time, obtain the pixel displacement data of multiple detection points or multiple detection lines under the recognition of the data processor, and use the pixel displacement data and key frame data to replace and update the changed cloud point parts in the entire point cloud data.

[0104] Therefore, not only can the measuring camera be used to monitor the deformed area in real time to ensure that the obtained four-dimensional tunnel model is always dynamically synchronized; but also the pixel displacement data and key frame data can be used to replace and update the changed cloud points to realize the replacement and reconstruction of part of the four-dimensional model area. While ensuring real-time monitoring, it also greatly saves the computing power of the overall reconstruction of the tunnel model.

[0105] This reduces the traditional method of using a total station to perform deformation analysis on the tunnel inner wall during tunnel construction, which typically involves scanning and monitoring once or twice a week, to a deformation analysis rate in seconds. This greatly improves the timeliness of tunnel monitoring and provides data support for construction workers' safe construction in the tunnel and project quality control.

[0106] As shown in FIG3 and FIG4 , the light source in step 101 includes a point light source. The method for realizing deformation detection of a precise point on the tunnel wall based on the point light source is as follows:

[0107] Step 201: Projecting a point light source onto a point on the tunnel sidewall that needs to be accurately observed to form a detection point;

[0108] Step 202: multiple measurement cameras are set up around the detection point, and the multiple measurement cameras are used to shoot and collect image data of the entire detection point in different time periods from multiple angles;

[0109] Step 203: The image data collected by the measurement camera is transmitted to a data processor. The data processor uses an image capture algorithm to compare the pixel displacement data and key frame data of the same detection point area at different angles in different time periods frame by frame, and calculates the deformation data of the detection point area.

[0110] Step 204: The data processor sends the deformation data of the detection point area and the time axis to the application terminal for displaying the current deformation data of the tunnel sidewall area where the detection point is located.

[0111] As shown in FIG5 and FIG6 , the light source in the above step 101 may also include a line light source. The method for realizing tunnel wall deformation detection based on the line light source is as follows:

[0112] The light source in step 101 includes a line light source, and the method for realizing tunnel wall deformation detection based on the line light source is as follows:

[0113] Step 301: multiple line light sources are installed on the tunnel sidewalls along the tunnel length direction to form cross-sectional detection lines with equal or random spacing on the tunnel vault and sidewalls;

[0114] Step 302: multiple measurement cameras are set up around the inspection line, and the multiple measurement cameras are used to capture image data of the entire inspection line area at different time periods from multiple angles;

[0115] Step 303: The image data collected by the measurement camera is transmitted to a data processor, and the three-dimensional coordinates of the area where the detection line is located are calculated using the detection line as a position reference. Image data of the detection line area obtained from different shooting angles are used to calculate the size and shape of the area where the detection line is located. The three-dimensional coordinates, size, and shape information are then integrated to create low-density point cloud data of the tunnel inner wall.

[0116] At the same time, the data processor uses threshold segmentation and centerline extraction algorithms to extract the detection line area, and then fuses and calculates with the original detection line and the historical pixel position data of the area where the detection line is located to obtain pixel movement data and key frame data of the same detection line area at different angles in different time periods; and uses the detection line as a calibration reference, using the pixel movement data and key frame data of the area where the detection line is located to calibrate and update the part of the point cloud data that has moved in real time;

[0117] Step 304: The data processor sends the corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel;

[0118] Step 305: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0119] Step 306: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm, and displays the real-time dynamic deformation data of the vault and sidewalls of the tunnel where the detection line is located.

[0120] As shown in FIG7 and FIG8 , the light source in step 101 further includes a surface light source. The method for realizing tunnel wall deformation detection based on the surface light source is as follows:

[0121] The light source in step 101 further includes a surface light source. The method for realizing tunnel wall deformation detection based on the surface light source is as follows:

[0122] Step 401: multiple surface light sources are installed on the tunnel sidewalls along the tunnel length direction to form a grid detection line on the tunnel vault and sidewalls;

[0123] Step 402: multiple measurement cameras are set up around the grid detection line, and the multiple measurement cameras are used to capture image data of the entire grid detection line area at different time periods from multiple angles;

[0124] Step 403: The image data collected by the measurement camera is transmitted to a data processor, and the three-dimensional coordinates of the area where the grid detection lines are located are calculated using the grid detection lines as a position reference. The size and shape of the area where the detection lines are located are calculated using the image data of the grid detection lines obtained at different shooting angles. The three-dimensional coordinates, size, and shape information are then integrated with the pixel positions of the grid detection lines in the image data to create low-density point cloud data of the tunnel inner wall.

[0125] At the same time, the data processor uses a deep learning image segmentation algorithm to extract the intersection of the grid detection lines as capture points, obtains pixel movement data and key frame data of the same capture point area at different angles and in different time periods, and calculates the pixel displacement data of the detection line area between two adjacent capture points based on the capture point. Then, the three-dimensional coordinates, size, and shape information are integrated to create low-density point cloud data of the tunnel inner wall.

[0126] Step 404: Send the corrected point cloud data to the application terminal and convert it into a four-dimensional model of the tunnel;

[0127] Step 405: The application terminal displays the continuously corrected and converted four-dimensional tunnel model in the order of the time axis to obtain a dynamic four-dimensional tunnel model;

[0128] Step 406: The application terminal compares the new 3D model with the historical 3D model using an image comparison algorithm to display the overall dynamic deformation parameters of the tunnel vault and sidewall space.

[0129] Example 3:

[0130] As shown in FIG9 , the point cloud data corrected in step 104 can also be used to calculate the thickness of the tunnel inner wall concrete. The specific calculation method is as follows:

[0131] Step 501: setting the grid detection line as a reference line;

[0132] Step 502: Calculate the distance between the intersection of the normal vector of each cloud point in the point cloud data and the baseline;

[0133] Step 503: retrieve the cloud point location data and match it with the intersection distance data;

[0134] Step 504: Synchronize and match the intersection distance data to the three-dimensional tunnel model based on the cloud point position data to obtain the concrete thickness at different locations in the tunnel.

[0135] Example 4:

[0136] As shown in FIG10 , the three-dimensional tunnel model generated in step 105 can be used to measure the flatness of the tunnel inner wall. The specific measurement method is as follows:

[0137] Step 601: using the detection point and / or detection line as a position reference point, and using a centerline capture algorithm to obtain the centerline of the three-dimensional tunnel model;

[0138] Step 602: Expand the entire 3D tunnel model along the center line of the 3D tunnel model to generate an orthogonal image;

[0139] Step 603: using a 3D measurement algorithm to obtain the elevation difference between the cloud point normal vector in the orthogonal image and the detection point and / or detection line on the tunnel inner wall surface;

[0140] Step 604: Collect elevation difference data to obtain the flatness of the tunnel inner wall.

[0141] Therefore, by unfolding the three-dimensional model into a two-dimensional orthogonal image, the three-dimensional surface measurement method is converted into a two-dimensional plane calculation method, which not only saves the computing power required to calculate the elevation difference but also ensures that the calculation results are more accurate.

[0142] Embodiment 5:

[0143] As shown in FIG11 , the data collection method of the measurement camera in step 102 can be used to record the process of shotcrete support and the location and number of anchor holes. The specific recording method is as follows:

[0144] Step 701: Setting up a measurement camera in the tunnel to record the entire bolt support process to generate video data;

[0145] Step 702: The video data collected by the measurement camera is transmitted to a data processor. The three-dimensional coordinates of the area where the detection points and / or detection lines are located are calculated using the detection points and / or detection lines as position references. The size and shape of the area where the detection points and / or detection lines are located are calculated using image data of the area where the detection points and / or detection lines are located obtained from different shooting angles. The three-dimensional coordinates, size, and shape information are then integrated to create low-density point cloud data of the tunnel inner wall.

[0146] At the same time, the data processor uses image recognition technology to automatically identify the anchor rods and their number in the video data, and locates the anchor rods with the help of detection points and / or detection lines. Then, the anchor rods' position and number information are integrated and corrected with the point cloud data again using the detection points and / or detection lines as the position reference.

[0147] Step 703: Send the corrected point cloud data to the application terminal and convert it into a four-dimensional model of the tunnel;

[0148] Step 704: The application displays the location, quantity, and drilling time of each anchor hole in the four-dimensional tunnel model; at the same time, the video data recorded by the measurement camera can also be exported to form a work video summary.

[0149] Example 6:

[0150] As shown in FIG12 , the data collection method of the measurement camera in step 102 can also be used to realize the generation of tunnel face sketch data. The specific generation method is as follows:

[0151] Step 801: multiple measurement cameras are set up in front of the tunnel face at different angles and positions;

[0152] Step 802: The measurement camera captures the image data of the tunnel face in real time;

[0153] Step 803: The image data collected by the measurement camera is transmitted to a data processor. The data processor uses image recognition technology to analyze the tunnel face image data obtained from different shooting angles to obtain data on the type of tunnel face rock, geotechnical integrity, groundwater discharge status, and rock hardness.

[0154] Step 804: The application terminal receives the image information integrated by the data processor and displays a sketch image of the tunnel face annotated with data including face rock type data, geotechnical integrity data, groundwater discharge status data, and rock hardness data.

[0155] From this tunnel face sketch image, engineers can clearly understand the structure, hardness and joint characteristics of each part of the tunnel face, providing data support for the optimization of tunnel face drilling and blasting.

[0156] This method and system for automatically collecting multi-dimensional information during a construction period tunnel establishes detection points and detection lines on the inner sidewall of the tunnel using a light source. The detection points are used as capture points for a measuring camera, facilitating real-time detection of pixel displacements in the tunnel sidewall region at the detection points over different time periods to obtain deformation data for the detection point region, or by connecting the detection lines to obtain deformation data for multiple continuous sections of the tunnel wall. Furthermore, the detection points and detection lines can be used as reference datums to update partial point cloud data based on the real-time pixel positions of the detection points and detection lines collected by the measuring camera, thereby obtaining an accurate four-dimensional tunnel model in real time. Furthermore, the detection points and detection lines, as reference objects, can also assist the measuring camera in accurately locating and recording specific points, such as the positions of anchor bolts. Furthermore, in the process of calculating the concrete thickness and smoothness of the tunnel wall based on the four-dimensional tunnel model, the detection points and detection lines can be used as reference datums to facilitate comparison and calculation of the distance from the normal vector of each cloud point in the point cloud data to the detection points and detection lines to obtain the concrete thickness. Furthermore, the detection points and detection lines can be used to locate the centerline of the tunnel via the detection points and detection lines, thereby unfolding the three-dimensional tunnel model into an orthogonal image around the centerline to facilitate obtaining concrete smoothness data.

[0157] Specifically, this application collects tunnel data through the interaction of light sources and measurement cameras, and establishes an accurate four-dimensional model of the tunnel construction process through a data processor and application terminal. This four-dimensional model can be used to optimize the drilling and blasting design of the tunnel face, calculate concrete thickness and flatness, and record the video of the shotcrete construction process and the position of the anchor rods. The entire monitoring and subsequent measurement are automated, with monitoring systems connected in parallel and data shared between them, effectively improving the automation and intelligence of the entire tunnel construction monitoring system, making the entire tunnel construction more efficient and safer.

[0158] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic multi-information collection system for tunnels during construction, characterized by: It comprises a plurality of light sources, which are arranged on the side walls of the tunnel and form detection points and / or detection lines on the inner wall of the tunnel; A plurality of measuring cameras are arranged around the detection point and / or detection line, and the measuring cameras are used to collect and record image data of the detection point and / or detection line area in real time, and the shooting angle range formed by the combination of the plurality of measuring cameras covers all surfaces of the detection point and / or detection line area; A data processor, wherein a signal input terminal of the data processor is connected to the measuring camera, and is used to receive image data collected by the measuring camera in real time, and respectively calculate the three-dimensional coordinates of the area where the detection point and / or the detection line are located with the detection point and / or the detection line as the position reference, and calculate the size and shape of the area where the detection point and / or the detection line are located with the image data of the area where the detection point and / or the detection line are located obtained from different shooting angles, and then integrate the three-dimensional coordinate, size, and shape information to establish low-density point cloud data of the inner wall of the tunnel; An application terminal, wherein a signal input end of the application terminal is connected to a signal output end of the data processor, and is used to receive real-time point cloud data updated and corrected by the data processor, and to display a four-dimensional model of the tunnel established based on the point cloud data. At the same time, based on the feedback of the four-dimensional model, the application terminal displays deformation data of the detection point and / or detection line area, concrete thickness data, tunnel flatness data, anchor number, position data, and face sketch data.

2. The automatic multi-information collection system for tunnels during construction according to claim 1 is characterized by: The light source is a point light source, a line light source, a surface light source or a combination of one or more thereof.

3. A method for automatically collecting multi-dimensional information of a tunnel during construction, wherein the method adopts the automatic collection system for multi-dimensional information of a tunnel during construction as claimed in any one of claims 1 to 2, and is characterized in that: Step 101: Arrange multiple light sources horizontally or vertically on the side wall of the tunnel, and use the light sources to form detection points and / or detection lines on the vault and side wall of the tunnel; Step 102: multiple measuring cameras are set up around the detection point and / or detection line, and multiple measuring cameras are used to shoot and collect image data of the entire detection point and / or detection line area in different time periods in real time from multiple angles; Step 103: The image data collected by the measuring camera in real time is transmitted to a data processor, and the data processor calculates the three-dimensional coordinates of the area where the detection point and / or the detection line are located using the detection point and / or the detection line as the position reference, and calculates the size and shape of the area where the detection point and / or the detection line are located using the image data of the detection point and / or the detection line area obtained from different shooting angles, and then integrates the three-dimensional coordinate, size, and shape information to establish low-density point cloud data of the inner wall of the tunnel; At the same time, the image capture algorithm is used to compare the pixel displacement data and key frame data of the same detection point and / or detection line area in different time periods at different angles frame by frame; and the detection point and / or detection line are used as the alignment reference, and the pixel movement data and key frame data of the area where the detection point and / or detection line are located are used to calibrate and update the part of the point cloud data that has moved in real time; Step 104: The data processor sends the real-time corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel; Step 105: the application terminal displays the tunnel four-dimensional model obtained by continuous correction and transformation in the order of the time axis to obtain a dynamic tunnel four-dimensional model; Step 106: The application terminal compares the new three-dimensional model with the historical three-dimensional model using an image comparison algorithm, and displays the real-time dynamic deformation data of the tunnel inner wall, the concrete thickness data, and the tunnel flatness data.

4. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3 is characterized in that: The light source in step 101 includes a point light source, and the method for realizing deformation detection of a precise point on the tunnel wall based on the point light source is as follows: Step 201: Projecting a point light source onto a point on the tunnel side wall that needs to be accurately observed to form a detection point; Step 202: multiple measuring cameras are set up around the detection point, and the image data of the entire detection point in different time periods are collected by using multiple measuring cameras in cooperation with each other at multiple angles; Step 203: The image data collected by the measuring camera is transmitted to a data processor, and the data processor compares the pixel displacement data and key frame data of the same detection point area at different angles in different time periods frame by frame using an image capture algorithm, and calculates the deformation data of the detection point area; Step 204: The data processor sends the deformation data of the detection point area and the time axis to the application terminal for displaying the current deformation data of the tunnel side wall area where the detection point is located.

5. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3 is characterized in that: The light source in step 101 includes a line light source, and the method for realizing tunnel wall deformation detection based on the line light source is as follows: Step 301: multiple line light sources are installed on the side walls of the tunnel along the length direction of the tunnel to form cross-sectional detection lines with the same spacing or any spacing on the tunnel vault and the side walls; Step 302: multiple measuring cameras are set up around the inspection line, and the multiple measuring cameras are used to cooperate to capture image data of the entire inspection line area in different time periods at multiple angles; Step 303: The image data collected by the measuring camera is transmitted to the data processor, and the three-dimensional coordinates of the area where the detection line is located are calculated using the detection line as the position reference, and the image data of the detection line area is obtained from different shooting angles to calculate the size and shape of the area where the detection line is located, and then the three-dimensional coordinates, size, and shape information are integrated to establish low-density point cloud data of the inner wall of the tunnel; At the same time, the data processor uses threshold segmentation and centerline extraction algorithms to extract the detection line area, and then fuses and calculates with the original detection line and the historical pixel position data of the detection line area to obtain the pixel movement data and key frame data of the same detection line area at different angles in different time periods; The detection line is used as a calibration reference, and pixel movement data and key frame data of the area where the detection line is located are used to calibrate and update the part of the point cloud data that has moved in real time; Step 304: the data processor sends the corrected point cloud data to the application terminal to convert it into a four-dimensional model of the tunnel; Step 305: the application terminal displays the tunnel four-dimensional model obtained by continuous correction and transformation in the order of the time axis to obtain a dynamic tunnel four-dimensional model; Step 306: The application terminal compares the new three-dimensional model with the historical three-dimensional model using an image comparison algorithm, and displays real-time dynamic deformation data of the vault and side walls of the tunnel where the detection line is located.

6. The method for automatically collecting multi-dimensional information during tunnel construction according to claim 3, characterized in that: The light source in step 101 also includes a surface light source, and the method for realizing tunnel wall deformation detection based on the surface light source is as follows: Step 401: installing a plurality of surface light sources on the side walls of the tunnel along the length direction of the tunnel to form a grid detection line on the tunnel vault and the side walls; Step 402: multiple measuring cameras are set up around the grid detection line, and the multiple measuring cameras are used to cooperate to capture image data of the entire grid detection line area in different time periods at multiple angles; Step 403: The image data collected by the measuring camera is transmitted to the data processor, and the three-dimensional coordinates of the area where the grid detection line is located are calculated using the grid detection line as a position reference, and the size and shape of the area where the detection line is located are calculated using the grid detection line image data obtained from different shooting angles, and then the three-dimensional coordinates, size, and shape information are integrated with the pixel position of the grid detection line in the image data to establish low-density point cloud data of the inner wall of the tunnel; At the same time, the data processor uses a deep learning image segmentation algorithm to extract the intersection of the grid detection line as the capture point, obtains the pixel movement data and key frame data of the same capture point area at different angles in different time periods, and calculates the pixel displacement data of the detection line area between two adjacent capture points based on the capture point; Then, the three-dimensional coordinates, size, and shape information are integrated to establish low-density point cloud data of the inner wall of the tunnel; Step 404: sending the corrected point cloud data to the application terminal to be converted into a four-dimensional model of the tunnel; Step 405: the application terminal displays the tunnel four-dimensional model obtained by continuous correction and transformation in the order of the time axis to obtain a dynamic tunnel four-dimensional model; Step 406: The application terminal compares the new three-dimensional model with the historical three-dimensional model using an image comparison algorithm to display the overall dynamic deformation parameters of the tunnel vault and sidewall space.

7. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3 is characterized by: The corrected point cloud data in step 104 can also be used to calculate the thickness of the concrete wall of the tunnel. The specific calculation method is as follows: Step 501: setting the grid detection line as a reference line; Step 502: Calculate the distance between the intersection of the normal vector of each cloud point in the point cloud data and the baseline; Step 503: retrieve the cloud point location data and match it with the intersection point distance data; Step 504: Synchronize and match the intersection distance data to the three-dimensional model of the tunnel based on the cloud point position data to obtain the concrete thickness at different positions in the tunnel.

8. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3 is characterized by: The four-dimensional tunnel model generated in step 105 can also be used to measure the flatness of the inner wall of the tunnel. The specific measurement method is as follows: Step 601: using the detection point and / or detection line as a position reference point, and using a centerline capture algorithm to obtain a centerline of a three-dimensional tunnel model; Step 602: Expand the entire three-dimensional tunnel model along the center line of the three-dimensional tunnel model to generate an orthogonal image; Step 603: using a three-dimensional measurement algorithm to obtain the elevation difference between the cloud point normal vector in the orthogonal image and the detection point and / or detection line on the tunnel inner wall surface; Step 604: Collect elevation difference data to obtain the flatness of the inner wall of the tunnel.

9. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3, characterized in that: The measurement camera data collection method in step 102 can also be used to record the spray anchor support process and the anchor hole position and quantity. The specific recording method is as follows: Step 701: Setting up a measuring camera in the tunnel to record the entire bolt support process to form video data; Step 702: The video data collected by the measuring camera is transmitted to the data processor, and the three-dimensional coordinates of the area where the detection point and / or the detection line are located are calculated using the detection point and / or the detection line as the position reference, and the size and shape of the area where the detection point and / or the detection line are located are calculated using the image data of the area where the detection point and / or the detection line are located obtained from different shooting angles, and then the three-dimensional coordinate, size, and shape information are integrated to establish low-density point cloud data of the inner wall of the tunnel; At the same time, the data processor uses image recognition technology to automatically identify the anchor rods and their quantity in the video data, and locates the position of the anchor rods with the help of detection points and / or detection lines; then the position and quantity information of the anchor rods are integrated and corrected with the point cloud data again based on the detection points and / or detection lines; Step 703: sending the corrected point cloud data to the application terminal and converting it into a four-dimensional model of the tunnel; Step 704: The application terminal displays the position, quantity and drilling time of each hole of the anchor rod in the four-dimensional model of the tunnel; at the same time, the video data recorded by the measurement camera can also be exported to form a work video summary.

10. The method for automatically collecting multi-dimensional information of tunnels during construction according to claim 3, characterized in that: The data collection method of the measuring camera in step 102 can also be used to realize the generation of tunnel face sketch data, and the specific generation method is as follows: Step 801: multiple measurement cameras are set up in front of the tunnel face at different angles and positions; Step 802: The measuring camera captures the image data of the tunnel face in real time; Step 803: The image data collected by the measurement camera is transmitted to a data processor, and the data processor uses image recognition technology to analyze the tunnel face image data obtained under different shooting angles to obtain tunnel face rock type data, rock and soil integrity data, groundwater discharge status data, and rock hardness data; Step 804: The application terminal receives the image information integrated by the data processor, and displays a sketch image of the tunnel face annotated with data including face rock type data, geotechnical integrity data, groundwater discharge status data, and rock hardness data.

Images