Method and system for automatically collecting multiple information of tunnel during construction period

A system for automatically collecting multiple information of a tunnel during construction period, comprising:

• a plurality of light sources, wherein the light sources are set on a side wall of the tunnel to form detection points and/or detection lines on an inner wall of the tunnel;
• a plurality of measuring cameras are erected around the detection points and/or the detection lines, and the measuring cameras are used for collecting and recording image data of areas of the detection points and/or the detection lines in real time, and a shooting visual angle range formed by a combination of the plurality of the measuring cameras covers all surfaces of the areas of the detection points and/or the detection lines;
• a data processor, wherein a signal input end of the data processor is connected with the measuring cameras, used for receiving the image data collected by the measuring cameras in real time, calculating three-dimensional coordinates of the areas of the detection points and/or the detection lines taking the detection points and/or the detection lines as position references respectively, calculating sizes and shapes of the areas of the detection points and/or the detection lines with the image data of the areas of the detection points and/or the detection lines obtained from different shooting visual angles, and then integrating the three-dimensional coordinates, size and shape information to establish low-density point cloud data of the inner wall of the tunnel;
• and
• an application terminal, wherein a signal input end of the application terminal is connected with a signal output end of the data processor, and is used for receiving real-time point cloud data updated and corrected by the data processor, and displaying a four-dimensional model of the tunnel established according to the point cloud data, and at the same time, feeding back and displaying deformation data, concrete thickness data, tunnel flatness data, anchor rod number, position data and tunnel face sketch data comprising the areas of the detection points and/or the detection lines according to the four-dimensional model;
• wherein a method for automatically collecting the multiple information of the tunnel during the construction period comprises following steps:
• step 101:
• arranging the plurality of the light sources in lateral or longitudinal direction of the side wall of the tunnel, and forming the detection points and/or the detection lines on a crown and the side wall of the tunnel by using the light sources;
• step 102:
• erecting the plurality of the measuring cameras around the detection points and/or the detection lines, and shooting and collecting the image data of a whole area of the detection points and/or the detection lines in different time periods in real time from multiple angles by using a cooperation of the plurality of the measuring cameras;
• step 103:
• transmitting the image data collected by the measuring cameras in real time to the data processor, and by the data processor, calculating the three-dimensional coordinates of the areas of the detection points and/or the detection lines taking the detection points and/or the detection lines as the position references respectively, calculating the sizes and the shapes of the areas of the detection points and/or the detection lines with the image data of the areas of the detection points and/or the detection lines obtained from different shooting visual angles, and then integrating the three-dimensional coordinates, the size and shape information to establish the low-density point cloud data of the inner wall of the tunnel;
• at the same time, using an image capture algorithm to compare pixel displacement data and key frame data of an area of a same detection point and/or detection line at different angles in different periods;
• and taking the detection points and/or the detection lines as an alignment benchmark, and using the pixel displacement data and the key frame data in the areas of the detection points and/or the detection lines, calibrating and updating a part of point clouds generating position movement in the point cloud data in real time;
• step 104:
• sending real-time corrected point cloud data to the application terminal by the data processor to convert into the four-dimensional model of the tunnel;
• step 105:
• displaying a continuously corrected and converted four-dimensional model of the tunnel by the application terminal according to an order of a time axis to obtain a dynamic four-dimensional model of the tunnel;
• and
• step 106:
• using an image comparison algorithm to compare a new three-dimensional model with a historical three-dimensional model by the application terminal, and displaying real-time dynamic deformation data, concrete thickness data and tunnel flatness data of the inner wall of the tunnel;
• the light sources in the step 101 comprise a point light source, and a method for realizing deformation detection of an accurate point on a tunnel wall based on the point light source is as follows:
• step 201:
• projecting the point light source to a certain point needing accurate observation on the side wall of the tunnel to form a detection point;
• step 202:
• erecting the plurality of the measuring cameras around the detection point, and shooting and collecting the image data of a whole detection point in the different time periods from the multiple angles by using the cooperation of the plurality of the measuring cameras;
• step 203:
• transmitting the image data collected by the measuring cameras to the data processor, and using the image capture algorithm to compare the pixel displacement data and the key frame data of the area of the same detection point at the different angles by the data processor in the different time periods frame by frame, and then calculating the deformation data of the area of the detection point;
• and
• step 204:
• sending the deformation data of the area of the detection point and the time axis to the application terminal by the data processor for displaying current deformation data of an area of the side wall of the tunnel of the detection point;
• the light sources in the step 101 comprise a line light source, and a tunnel wall deformation detection method based on the line light source is as follows:
• step 301:
• installing the plurality of line light sources on the side wall of the tunnel along a length direction of the tunnel, and forming section detection lines with a same or arbitrary spacing on the crown and the side wall of the tunnel;
• step 302:
• erecting the plurality of the measuring cameras around the detection lines, and using the cooperation of the plurality of the measuring cameras to shoot image data in the different time periods in a whole area of the detection lines at the multiple angles;
• step 303:
• transmitting the image data collected by the measuring cameras to the data processor, calculating the three-dimensional coordinates of the area of the detection lines taking the detection lines as the position references respectively, calculating a size and a shape of the area of the detection lines with the image data of the area of the detection lines obtained from different shooting visual angles, and then integrating the three-dimensional coordinates, the size and shape information to establish the low-density point cloud data of the inner wall of the tunnel;
• at the same time, extracting the area of the detection lines by a threshold segmentation and centerline extraction algorithm by the data processor, and then fusing with original detection lines and historical pixel position data of the area of the detection lines and calculating to obtain pixel displacement data and key frame data of a same area of the detection lines at the different angles in the different time periods;
• and using the detection lines as an alignment benchmark to calibrate and update the part of the point clouds generating the position movement in the point cloud data in real time by using the pixel displacement data and the key frame data in the area of the detection lines;
• step 304:
• sending corrected point cloud data to the application terminal by the data processor to convert into the four-dimensional model of the tunnel;
• step 305:
• displaying the continuously corrected and converted four-dimensional model of the tunnel by the application terminal according to the order of the time axis to obtain the dynamic four-dimensional model of the tunnel;
• and
• step 306:
• using the image comparison algorithm to compare the new three-dimensional model with the historical three-dimensional model by the application terminal, and displaying the real-time dynamic deformation data of the crown and the side wall of the detection lines in the tunnel;
• the light sources in the step 101 also comprise a surface light source, and a tunnel wall deformation detection method based on the surface light source is as follows:
• step 401:
• installing a plurality of surface light sources on the side wall of the tunnel along the length direction of the tunnel, and forming grid detection lines on the crown and the side wall of the tunnel;
• step 402:
• erecting the plurality of the measuring cameras around the grid detection lines, and using the cooperation of the plurality of the measuring cameras to shoot image data in the different time periods in a whole area of the grid detection lines at the multiple angles;
• step 403:
• transmitting the image data collected by the measuring cameras to the data processor, calculating three-dimensional coordinates of the area of the grid detection lines taking the grid detection lines as the position references respectively, calculating the size and the shape of the area of the detection lines with the image data of the area of the detection lines obtained from different shooting visual angles, and then integrating the three-dimensional coordinates, the size and shape information by a pixel position of the grid detection lines in the image data to establish the low-density point cloud data of the inner wall of the tunnel;
• at the same time, extracting intersections of the grid detection lines as capture points using a deep learning image segmentation algorithm by the data processor, and obtaining pixel displacement data and key frame data of a same capture point area in the different time periods at the different angles, and calculating pixel displacement data of the area of the detection lines between two adjacent capture points based on the capture points;
• and then integrating the three-dimensional coordinates, the size and shape information to establish the 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 convert into the four-dimensional model of the tunnel;
• step 405:
• displaying the continuously corrected and converted four-dimensional model of the tunnel by the application terminal according to the order of the time axis to obtain the dynamic four-dimensional model of the tunnel;
• and
• step 406:
• using the image comparison algorithm to compare the new three-dimensional model with the historical three-dimensional model by the application terminal, and displaying overall dynamic deformation parameters of tunnel crown and side wall space;
• the corrected point cloud data in the step 104 is also capable of being used to calculate concrete thickness of the inner wall of the tunnel, and a specific calculation method is as follows:
• step 501:
• setting the grid detection lines as reference lines;
• step 502:
• calculating an intersection distance between a normal vector of each cloud point in the point cloud data and the reference lines;
• step 503:
• calling out cloud point position data and matching with intersection distance data;
• and
• step 504:
• synchronizing and matching the intersection distance data to a three-dimensional model of the tunnel based on the cloud point position data to obtain concrete thickness at different positions in the tunnel;
• a four-dimensional tunnel model generated in the step 105 is also capable of being used to measure smoothness of the inner wall of the tunnel, and a specific measurement method is as follows:
• step 601:
• taking the detection points and/or the detection lines as position reference points, obtaining a centerline of a three-dimensional tunnel model by a centerline capture algorithm;
• step 602:
• expanding a whole three-dimensional tunnel model along a centerline of the three-dimensional tunnel model to generate an orthogonal image map;
• step 603:
• obtaining an elevation difference between the normal vector of the cloud point in the orthogonal image map and the detection points and/or the detection lines on an inner wall surface of the tunnel by using a three-dimensional measurement algorithm;
• and
• step 604:
• collecting elevation difference data to obtain flatness of the inner wall of the tunnel;
• a data collection method of the measuring cameras in the step 102 is also capable of being used to record a shotcrete and anchor support process and positions and a number of anchor holes, and a specific recording method is as follows:
• step 701:
• erecting the measuring cameras in the tunnel to record a whole anchor support process by video to form video data;
• step 702:
• transmitting the video data collected by the measuring cameras to the data processor, and calculating the three-dimensional coordinates of the areas of the detection points and/or the detection lines taking the detection points and/or the detection lines as the position references respectively, calculating the sizes and the shapes of the areas of the detection points and/or the detection lines with the image data of the areas of the detection points and/or the detection lines obtained from different shooting visual angles, and then integrating the three-dimensional coordinates, the size and shape information to establish the low-density point cloud data of the inner wall of the tunnel;
• at the same time, using an image identification technology to automatically identify anchor rods and a number in the video data by the data processor, and locating positions of the anchor rods with an assistance of the detection points and/or the detection lines;
• then, integrating and correcting position and quantity information of the anchor rods and the point cloud data with the detection points and/or the detection lines as the position references again;
• step 703:
• sending the corrected point cloud data to the application terminal to convert into the four-dimensional model of the tunnel;
• and
• step 704:
• displaying the positions and the number of the anchor rods and drilling time of each hole in the four-dimensional model of the tunnel by the application terminal;
• wherein the video data recorded by the measuring cameras is also capable of being derived to form a working video summary;
• the data collection method of the measuring cameras in the step 102 is also capable of being used to generate the tunnel face sketch data, and a specific generation method is as follows:
• step 801:
• erecting the plurality of the measuring cameras in front of a tunnel face at different angles and positions;
• step 802:
• capturing formation image data of the tunnel face in real time by the measuring cameras;
• step 803:
• transmitting the image data collected by the measuring cameras to the data processor, and using the image identification technology to analyze the image data of the tunnel face obtained from different shooting visual angles by the data processor, so as to obtain type data of tunnel face rocks, integrity data of rock and soil, groundwater outflow state data and rock hardness data;
• and
• step 804:
• receiving image information integrated by the data processor by the application terminal, and displaying a tunnel face sketch image marked with the type data of tunnel face rocks, the integrity data of rock and soil, the groundwater outflow state data and the rock hardness data.