Three-dimensional laser strong earthquake tunnel lining crack detection method and system

Abstract

The application provides a three-dimensional laser strong earthquake tunnel lining crack detection method and system, comprising: detecting a tunnel through a multi-directional focusing three-dimensional laser scanning device to obtain multiple groups of tunnel laser point cloud data; using an optimized federated weighted learning model to respectively process the multiple groups of tunnel laser point cloud data, combining asynchronous and residual test adaptive adjustment models to optimize and adjust the processed tunnel laser point cloud data, and obtaining an optimized and adjusted tunnel laser point cloud data set; accurately detecting tunnel lining cracks according to the optimized and adjusted tunnel laser point cloud data set, and reducing interference factors in the tunnel; collecting high-resolution images of strong earthquake tunnel linings, performing panoramic enlarged image visual intelligent identification, combining strong earthquake tunnel lining crack identification information, and performing tunnel lining crack marking and crack position and size warning.

Description

TECHNICAL FIELD

[0001] The application relates to a precision detection tunnel detection noise reduction technology field, and more particularly to a three-dimensional laser strong earthquake tunnel lining crack detection method and system. BACKGROUND

[0002] Tunnel lining crack detection precision is very critical, tunnel strong earthquake lining crack is difficult to effectively detect, which limits the tunnel lining crack detection technology, how to improve the laser scanning device to detect the tunnel precision and reduce the interference, how to respectively process the tunnel laser point cloud data after optimization and adjustment, how to improve the precise detection of the tunnel lining crack and reduce the interference factors in the tunnel, how to process and accurately identify the strong earthquake tunnel lining image, and how to improve the tunnel lining crack marking and crack position size warning and timeliness, and the like problems remain to be solved, therefore, it is necessary to provide a three-dimensional laser strong earthquake tunnel lining crack detection method and system to at least partially solve the problems in the prior art. SUMMARY

[0003] A series of simplified concepts are introduced in the summary section, which will be further described in detail in the specific embodiment section; the summary section of the application does not mean to try to limit the key features and necessary technical features of the claimed technical solution, nor to try to determine the protection scope of the claimed technical solution.

[0004] To at least partially solve the above problems, the application provides a three-dimensional laser strong earthquake tunnel lining crack detection method, comprising:

[0005] S10, detecting the tunnel by a multi-directional focusing three-dimensional laser scanning device, and acquiring a plurality of groups of tunnel laser point cloud data;

[0006] S20, using an optimized federated weighted learning model to respectively process the plurality of groups of tunnel laser point cloud data, combining asynchronous and residual test adaptive adjustment model, optimizing and adjusting the processed tunnel laser point cloud data, and acquiring an optimized and adjusted tunnel laser point cloud data set;

[0007] S30, accurately detecting the tunnel lining crack according to the optimized and adjusted tunnel laser point cloud data set, and reducing the interference factors in the tunnel;

[0008] S40, collecting a strong earthquake tunnel lining high-resolution image, performing panoramic enlarged image visual intelligent identification, combining strong earthquake tunnel lining crack identification information, and marking the tunnel lining crack and warning the crack position size.

[0009] Preferably, S10 comprises:

[0010] S101, adopt angle-adjusting non-collinear multi-directional laser scanning detection, and construct a multi-directional focusing three-dimensional laser scanning device;

[0011] S102, detect a tunnel through the multi-directional focusing three-dimensional laser scanning device, and acquire multiple groups of tunnel laser point cloud data;

[0012] The multi-directional focusing three-dimensional laser scanning device comprises a laser autonomous positioning module, a laser scanning automatic moving module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module and a laser focal point adjusting module; through multi-directional focusing three-dimensional laser scanning, the influence of interference factors in the tunnel on laser scanning is reduced.

[0013] Preferably, S20 comprises:

[0014] S201, using an optimized federated weighted learning model, multiple groups of tunnel laser point cloud data are processed respectively, and optimized processed tunnel laser point cloud data are acquired;

[0015] S202, the optimized federated weighted learning model is combined with an asynchronous and residual test adaptive adjustment model, the optimized and adjusted processed tunnel laser point cloud data are optimized and adjusted, an optimized and adjusted processed tunnel laser point cloud data set is constructed, and an optimized and adjusted processed tunnel laser point cloud data set is acquired.

[0016] Preferably, S30 comprises:

[0017] S301, according to the optimized and adjusted processed tunnel laser point cloud data set, a tunnel lining crack is accurately detected, and tunnel lining crack detection information is acquired; interference factors in the tunnel are reduced; the interference factors in the tunnel include dust interference factors or exposed steel interference factors, structure joint interference factors;

[0018] S302, the tunnel lining crack detection information is sent to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0019] Preferably, S40 comprises:

[0020] S401, a high-definition camera vision recognition device is set, high-resolution images of a strong earthquake tunnel lining are collected, panoramic enlarged image vision intelligent recognition is performed, and complete information of a strong earthquake tunnel lining crack is acquired;

[0021] S402, according to the tunnel lining crack detection information, combined with strong earthquake tunnel lining crack identification information, tunnel lining crack marking and crack position and size warning are performed.

[0022] The application provides a three-dimensional laser strong earthquake tunnel lining crack detection system, which comprises:

[0023] The multi-directional focusing three-dimensional scanning subsystem acquires multiple sets of tunnel laser point cloud data by detecting the tunnel through the multi-directional focusing three-dimensional laser scanning device.

[0024] The optimized federal asynchronous residual adjustment subsystem uses an optimized federal weighted learning model to respectively process the multiple sets of tunnel laser point cloud data, combines an asynchronous and residual test adaptive adjustment model, optimally adjusts the processed tunnel laser point cloud data, and acquires an optimized and adjusted tunnel laser point cloud dataset.

[0025] The lining crack detection and interference reduction subsystem accurately detects tunnel lining cracks and reduces interference factors in the tunnel according to the optimized and adjusted tunnel laser point cloud dataset.

[0026] The strong earthquake tunnel lining identification and warning subsystem collects high-resolution images of strong earthquake tunnel linings, performs panoramic enlarged image visual intelligent identification, combines strong earthquake tunnel lining crack identification information, and performs tunnel lining crack marking and crack position and size warning.

[0027] Preferably, the multi-directional focusing three-dimensional scanning subsystem comprises:

[0028] The angle-adjusted multi-directional laser detection subsystem adopts angle-adjusted non-collinear multi-directional laser scanning detection to construct the multi-directional focusing three-dimensional laser scanning device.

[0029] The multi-directional focusing laser point cloud subsystem acquires multiple sets of tunnel laser point cloud data by detecting the tunnel through the multi-directional focusing three-dimensional laser scanning device.

[0030] The multi-directional focusing three-dimensional laser scanning device comprises a laser autonomous positioning module, a laser scanning automatic movement module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module, and a laser focal point adjustment module. The multi-directional focusing three-dimensional laser scanning reduces the influence of interference factors in the tunnel on laser scanning.

[0031] Preferably, the optimized federal asynchronous residual adjustment subsystem comprises:

[0032] The optimized federal weighted learning subsystem uses an optimized federal weighted learning model to respectively process the multiple sets of tunnel laser point cloud data and acquire optimized processed tunnel laser point cloud data.

[0033] The asynchronous residual test adaptive subsystem combines the optimized federal weighted learning model with an asynchronous and residual test adaptive adjustment model, optimally adjusts the processed tunnel laser point cloud data, constructs an optimized and adjusted tunnel laser point cloud dataset, and acquires an optimized and adjusted tunnel laser point cloud dataset.

[0034] Preferably, the lining crack detection and interference reduction subsystem comprises:

[0035] The lining crack reduction detection subsystem detects tunnel lining cracks according to an optimized and adjusted tunnel laser point cloud data set, obtains tunnel lining crack detection information, and reduces interference factors in the tunnel.

[0036] The tunnel lining detection cloud transmission subsystem transmits the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0037] Preferably, the strong earthquake tunnel lining identification and warning subsystem comprises:

[0038] The high-definition image panoramic magnification subsystem sets up a high-definition camera visual recognition device, collects high-resolution images of the strong earthquake tunnel lining, performs panoramic magnification image visual intelligent recognition, and obtains complete information of the strong earthquake tunnel lining cracks.

[0039] The strong earthquake tunnel lining crack warning subsystem marks the tunnel lining cracks and warns the crack position and size according to the tunnel lining crack detection information and the strong earthquake tunnel lining crack identification information.

[0040] Compared with the prior art, the present application has at least the following beneficial effects:

[0041] The three-dimensional laser strong earthquake tunnel lining crack detection method and system can significantly improve the detection accuracy of tunnel lining cracks, the completeness and comprehensiveness of tunnel strong earthquake lining crack detection are significantly improved, the detection accuracy of the laser scanning device for detecting tunnels can be improved and the influence of interference factors can be reduced, the tunnel laser point cloud data after optimization and adjustment can be processed, the precise detection of tunnel lining cracks and the reduction of interference factors in the tunnel can be improved, the strong earthquake tunnel lining image can be processed and accurately recognized, the timeliness of the tunnel lining crack marking and crack position size warning is significantly improved, and the problem of low recognition accuracy and poor anti-interference ability of the three-dimensional laser scanner for detecting tunnel cracks is solved, a new idea of crack detection research based on the federal weighted learning algorithm is proposed, based on the tunnel laser point cloud data, firstly, the optimization federal weighted learning algorithm is used, and the asynchronous and residual test adaptive adjustment algorithm is used, so that the purpose of accurately detecting the tunnel cracks is achieved. The algorithm and the traditional algorithm are compared and analyzed in terms of reliability, accuracy and measurement accuracy of crack detection in Linfen Expressway tunnel, and the results show that the new method can effectively improve the reliability and accuracy of tunnel crack detection, and has good performance on the accuracy of crack width detection; when dust, exposed steel bars and other interference factors appear in the detection results, the new algorithm has obvious advantages in reliability over the traditional algorithm, and can achieve more than 95% recognition accuracy and less than 10% recognition error rate, and the robustness of the algorithm application effect is significantly improved; the present application has important technical significance and significant effect.

[0042] The three-dimensional laser strong earthquake tunnel lining crack detection method and system, other advantages, objects and features of the present application will be partially embodied by the following description, and partially understood by those skilled in the art through research and practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0043] The accompanying drawings are intended to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation of the present application. In the drawings:

[0044] Fig. 1Figure of one embodiment of the three-dimensional laser strong earthquake tunnel lining crack detection method according to the present application.

[0045] Fig. 2 Figure of one embodiment of the three-dimensional laser strong earthquake tunnel lining crack detection system according to the present application.

[0046] Fig. 3 Figure of another embodiment of the three-dimensional laser strong earthquake tunnel lining crack detection system according to the present application. DETAILED DESCRIPTION

[0047] The present application will be further described in detail below with reference to the accompanying drawings and embodiments, so that those skilled in the art can implement the present application according to the description; as shown, the present application provides a three-dimensional laser strong earthquake tunnel lining crack detection method, comprising: Figs. 1-3

[0048] S10, detecting the tunnel by a multi-directional focusing three-dimensional laser scanning device to obtain a plurality of groups of tunnel laser point cloud data;

[0049] S20, using an optimized federated weighted learning model to process the plurality of groups of tunnel laser point cloud data respectively, combining asynchronous and residual test to adaptively adjust the model, optimizing and adjusting the processed tunnel laser point cloud data to obtain an optimized and adjusted tunnel laser point cloud data set;

[0050] S30, accurately detecting the tunnel lining crack according to the optimized and adjusted tunnel laser point cloud data set, and reducing the interference factors in the tunnel;

[0051] S40, collecting high-resolution images of the strong earthquake tunnel lining, performing panoramic enlarged image visual intelligent recognition, combining strong earthquake tunnel lining crack identification information, and performing tunnel lining crack marking and crack position and size warning.

[0052] ​The principle and effects of the above technical solution are: the application provides a three-dimensional laser strong earthquake tunnel lining crack detection method, which comprises the following steps: detecting a tunnel through a multi-directional focusing three-dimensional laser scanning device to obtain multiple groups of tunnel laser point cloud data; an optimized federated weighted learning model is used to process the multiple groups of tunnel laser point cloud data respectively, and an asynchronous and residual test adaptive adjustment model is combined to optimize and adjust the processed tunnel laser point cloud data, so as to obtain the optimized and adjusted tunnel laser point cloud data set; according to the optimized and adjusted tunnel laser point cloud data set, the tunnel lining crack is accurately detected, and the interference factors in the tunnel are reduced; high-resolution images of the strong earthquake tunnel lining are collected, panoramic enlarged image visual intelligent recognition is performed, combined with strong earthquake tunnel lining crack identification information, tunnel lining crack marking and crack position and size warning are performed; the detection accuracy of the tunnel lining crack can be significantly improved; the completeness and comprehensiveness of the strong earthquake tunnel lining crack detection are significantly improved; the detection accuracy of the laser scanning device for detecting the tunnel can be improved, and the influence of interference factors can be reduced; the tunnel laser point cloud data after optimization and adjustment can be processed respectively; the accurate detection of the tunnel lining crack and the reduction of the interference factors in the tunnel can be improved; the strong earthquake tunnel lining images can be processed and accurately recognized; the timeliness of the tunnel lining crack marking and crack position and size warning is significantly improved; in order to solve the problems of low recognition accuracy and poor anti-interference ability of the three-dimensional laser scanner for detecting the tunnel crack, a new crack detection method based on the combination of the federated weighted learning algorithm, the crack extension surface internal trend and the crack depth detection factors is proposed; based on the tunnel laser point cloud data, the optimized federated weighted learning algorithm is used, and the asynchronous and residual test adaptive adjustment algorithm is adopted; the asynchronous and residual test adaptive adjustment algorithm comprises asynchronous differential and residual distribution sampling and residual distribution adaptive fine adjustment operation, so that the accuracy of accurately detecting the tunnel crack and the measurement accuracy are significantly improved; the test is carried out in a plurality of terrain multi-section high-speed tunnels, and the algorithm in the present application is compared with the traditional algorithm in terms of reliability, accuracy and measurement accuracy of crack detection. The results show that the new method can effectively improve the reliability and accuracy of tunnel crack detection, and has good performance on the accuracy of crack width detection; when dust, exposed steel bars and other interference factors appear in the detection results, the new algorithm has obvious advantages in reliability over the traditional algorithm, and has higher identification accuracy than the reference and lower identification error rate than the reference; the new algorithm has obvious advantages in reliability over the traditional algorithm, and has more than 95% identification accuracy and less than 10% identification error rate; the robustness of the algorithm application effect is significantly improved; the application has important technical significance and significant effects.

[0053] In one embodiment, S10 comprises:

[0054] S101, adopting angle-adjusting non-collinear multi-directional laser scanning detection, constructing a multi-directional focusing three-dimensional laser scanning device;

[0055] S102, detecting the tunnel by the multi-directional focusing three-dimensional laser scanning device to obtain a plurality of sets of tunnel laser point cloud data;

[0056] The multi-directional focusing three-dimensional laser scanning device comprises a laser autonomous positioning module, a laser scanning automatic moving module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module, and a laser focal point adjusting module. The multi-directional focusing three-dimensional laser scanning device can reduce the influence of interference factors in the tunnel on laser scanning.

[0057] The principle and effect of the above technical solution are as follows: the angle-adjusting non-collinear multi-directional laser scanning detection is used to construct the multi-directional focusing three-dimensional laser scanning device; the tunnel is detected by the multi-directional focusing three-dimensional laser scanning device to obtain a plurality of sets of tunnel laser point cloud data; the angle-adjusting non-collinear multi-directional laser scanning detection is used to construct the multi-directional focusing three-dimensional laser scanning device; the multi-directional focusing three-dimensional laser scanning device comprises a laser autonomous positioning module 1011, a laser scanning automatic moving module 1012, a three-dimensional laser scanning module 1013, a laser scanning control module 1014, a laser focusing monitoring module 1015, and a laser focal point adjusting module 1016; the laser autonomous positioning module is used for self-positioning of the multi-directional focusing three-dimensional laser scanning device and relative position positioning of the multi-directional focusing three-dimensional laser scanning device and other multi-directional focusing three-dimensional laser scanning devices for joint detection; the multi-directional focusing three-dimensional laser scanning devices scan and detect the relative distance and the relative angle therebetween to form a multi-directional focusing three-dimensional laser scanning device array network; the laser scanning automatic moving module is used for omnidirectional movement and keeping the three-dimensional laser scanning module perpendicular to the ground in the tunnel; the three-dimensional laser scanning module is used for laser scanning detection; the laser focusing monitoring module is used for monitoring the focusing positions of a plurality of multi-directional focusing three-dimensional laser scanning devices; the laser scanning control module is used for controlling the three-dimensional laser scanning modules of the multi-directional focusing three-dimensional laser scanning devices to differentiate laser wavelengths when scanning and detecting the same point position, and identifying the wavelengths of reflected lasers; the laser scanning control module controls the laser focal point adjusting module to adjust the focusing points of a plurality of multi-directional focusing three-dimensional laser scanning devices; and the multi-directional focusing three-dimensional laser scanning can significantly reduce the influence of various interference factors in the tunnel on laser scanning.

[0058] In one embodiment, S20 comprises:

[0059] S201, using an optimized federated weighted learning model to process a plurality of sets of tunnel laser point cloud data respectively to obtain tunnel laser point cloud data after optimization processing;

[0060] S202, combining the optimized federated weighted learning model with an asynchronous and residual test self-adaptive adjustment model to optimize and adjust the tunnel laser point cloud data after processing, constructing a set of tunnel laser point cloud data after optimization and adjustment processing, and obtaining a set of tunnel laser point cloud data after optimization and adjustment processing.

[0061] Principles and effects of the above technical solutions are: the optimized federated weighted learning model is used to process multiple groups of tunnel laser point cloud data respectively to obtain the tunnel laser point cloud data after optimization processing; the optimized federated weighted learning model is combined with an asynchronous and residual test adaptive adjustment model to optimize and adjust the tunnel laser point cloud data after processing, and a set of tunnel laser point cloud data after optimization and adjustment processing is constructed to obtain the set of tunnel laser point cloud data after optimization and adjustment processing.

[0062] In one embodiment, S30 includes:

[0063] S301, according to the set of tunnel laser point cloud data after optimization and adjustment processing, accurately detecting a tunnel lining crack to obtain tunnel lining crack detection information; reducing interference factors in the tunnel; the interference factors in the tunnel include dust interference factors or exposed steel interference factors, structure joint interference factors;

[0064] S302, sending the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0065] Principles and effects of the above technical solutions are: according to the set of tunnel laser point cloud data after optimization and adjustment processing, accurately detecting a tunnel lining crack to obtain tunnel lining crack detection information; reducing interference factors in the tunnel; the interference factors in the tunnel include dust interference factors or exposed steel interference factors, structure joint interference factors; sending the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0066] In one embodiment, S40 includes:

[0067] S401, setting a high-definition camera visual recognition device to collect high-resolution images of the strong earthquake tunnel lining, perform panoramic enlarged image visual intelligent recognition, and obtain complete information of the strong earthquake tunnel lining crack;

[0068] S402, according to the tunnel lining crack detection information, combining the strong earthquake tunnel lining crack recognition information, performing tunnel lining crack marking and crack position and size warning.

[0069] The principle and effect of the above technical scheme are: the high-definition camera visual recognition device is arranged to collect high-resolution images of the strong earthquake tunnel lining, panoramic enlarged image visual intelligent recognition is performed, and complete information of the strong earthquake tunnel lining crack is obtained; according to the tunnel lining crack detection information and in combination with the strong earthquake tunnel lining crack identification information, the tunnel lining crack marking and crack position and size warning are performed; the high-definition camera visual recognition device is arranged to collect high-resolution images of the strong earthquake tunnel lining, panoramic enlarged image visual intelligent recognition is performed, and complete information of the strong earthquake tunnel lining crack is obtained, including: the high-definition camera visual recognition device is arranged to collect high-resolution images of the strong earthquake tunnel lining; the high-resolution images of the tunnel lining are spliced and enlarged according to a set ratio to obtain a panoramic enlarged image of the tunnel lining; according to the panoramic enlarged image of the tunnel lining, panoramic enlarged image visual intelligent recognition is performed to obtain complete information of the strong earthquake tunnel lining crack; the complete information of the strong earthquake tunnel lining crack includes: crack starting position, crack extension, crack branch, crack branch end and crack extension end.

[0070] The application provides a three-dimensional laser strong earthquake tunnel lining crack detection system, which comprises:

[0071] The multi-directional focusing three-dimensional scanning subsystem detects the tunnel through a multi-directional focusing three-dimensional laser scanning device and obtains multiple groups of tunnel laser point cloud data.

[0072] The optimized federal asynchronous residual error adjustment subsystem uses an optimized federal weighted learning model to respectively process the multiple groups of tunnel laser point cloud data, combines an asynchronous and residual error test self-adaptive adjustment model, optimally adjusts the processed tunnel laser point cloud data, and obtains an optimized and adjusted tunnel laser point cloud data set.

[0073] The lining crack detection and interference reduction subsystem accurately detects the tunnel lining crack according to the optimized and adjusted tunnel laser point cloud data set and reduces interference factors in the tunnel.

[0074] The strong earthquake tunnel lining identification and warning subsystem collects high-resolution images of the strong earthquake tunnel lining, performs panoramic enlarged image visual intelligent recognition, combines the strong earthquake tunnel lining crack identification information, performs tunnel lining crack marking and crack position and size warning.

[0075] The principle and effect of the above technical solution are: the application provides a three-dimensional laser strong earthquake tunnel lining crack detection system, which comprises: a multi-directional focusing three-dimensional scanning subsystem, a multi-directional focusing three-dimensional laser scanning device is used to detect a tunnel, and a plurality of groups of tunnel laser point cloud data are obtained; an optimized federal asynchronous residual error adjustment subsystem, an optimized federal weighted learning model is used to process a plurality of groups of tunnel laser point cloud data respectively, an asynchronous and residual error test adaptive adjustment model is combined, tunnel laser point cloud data after processing is optimized and adjusted, and tunnel laser point cloud data set after optimization and adjustment processing is obtained; a lining crack detection and disturbance reduction subsystem, according to the tunnel laser point cloud data set after optimization and adjustment processing, tunnel lining cracks are accurately detected, and interference factors in the tunnel are reduced; a strong earthquake tunnel lining identification and warning subsystem, high-resolution images of the strong earthquake tunnel lining are collected, panoramic enlarged image visual intelligent identification is carried out, combined with strong earthquake tunnel lining crack identification information, tunnel lining crack marking and crack position and size warning are carried out; the tunnel lining crack can be accurately detected and the interference factors in the tunnel can be reduced; the strong earthquake tunnel lining image can be processed and accurately identified; the tunnel lining crack marking, crack position and size warning and timeliness are significantly improved; in order to solve the problems of low identification accuracy and poor anti-interference ability of the three-dimensional laser scanner in detecting tunnel cracks, a new crack detection method based on the federal weighted learning algorithm is proposed, which combines crack extension surface internal trend and crack depth detection factors; based on the tunnel laser point cloud data, firstly, the optimized federal weighted learning algorithm is used, and the asynchronous and residual error test adaptive adjustment algorithm is used; the asynchronous and residual error test adaptive adjustment algorithm comprises asynchronous differential and residual distribution sampling and residual distribution adaptive fine adjustment operation, so that the accuracy of accurately detecting the tunnel crack and the measurement accuracy are significantly improved; in the test of a plurality of terrain multi-section high-speed tunnels, the reliability, accuracy and measurement accuracy of crack detection are compared and analyzed with the traditional algorithm, and the results show that the new method can effectively improve the reliability and accuracy of tunnel crack detection, and has good performance on the accuracy of crack width detection; when dust, exposed steel bars and other interference factors appear in the detection results, the new algorithm has obvious advantages in reliability over the traditional algorithm, and the identification accuracy is more than 95% and above, and the identification error rate is significantly lower than the reference benchmark, and the robustness of the algorithm application effect is significantly improved; the application has important technical significance and significant effect.

[0076] In one embodiment, the multi-directional focusing three-dimensional scanning subsystem comprises:

[0077] The angle-adjusting multi-directional laser detection subsystem adopts angle-adjusting non-collinear multi-directional laser scanning detection to construct the multi-directional focusing three-dimensional laser scanning device.

[0078] The multi-directional focusing laser point cloud subsystem detects the tunnel through the multi-directional focusing three-dimensional laser scanning device, and obtains a plurality of groups of tunnel laser point cloud data.

[0079] The multi-directional focusing three-dimensional laser scanning device comprises a laser self-positioning module, a laser scanning automatic moving module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module and a laser focus adjusting module. The multi-directional focusing three-dimensional laser scanning reduces the influence of interference factors in the tunnel on laser scanning.

[0080] The principle and effect of the above technical solution are as follows: the multi-directional focusing three-dimensional scanning subsystem comprises an angle-adjusting multi-directional laser detection subsystem, a multi-directional focusing laser point cloud subsystem, and a multi-directional focusing three-dimensional laser scanning device. The angle-adjusting multi-directional laser detection subsystem adopts angle-adjusting non-collinear multi-directional laser scanning detection to construct the multi-directional focusing three-dimensional laser scanning device. The multi-directional focusing laser point cloud subsystem detects the tunnel through the multi-directional focusing three-dimensional laser scanning device and acquires multiple sets of tunnel laser point cloud data. The multi-directional focusing three-dimensional laser scanning device comprises a laser self-positioning module 1011, a laser scanning automatic moving module 1012, a three-dimensional laser scanning module 1013, a laser scanning control module 1014, a laser focusing monitoring module 1015 and a laser focus adjusting module 1016. The laser self-positioning module respectively performs self-positioning of the multi-directional focusing three-dimensional laser scanning device and relative position positioning with other multi-directional focusing three-dimensional laser scanning devices for joint detection. The multi-directional focusing three-dimensional laser scanning devices scan and detect the relative distance and the relative angle between each other to form a multi-directional focusing three-dimensional laser scanning device array network. The laser scanning automatic moving module performs universal movement and keeps the three-dimensional laser scanning module perpendicular to the tunnel floor. The three-dimensional laser scanning module performs laser scanning detection. The laser focusing monitoring module monitors the focusing positions of multiple multi-directional focusing three-dimensional laser scanning devices. The laser scanning control module controls the three-dimensional laser scanning modules of the multi-directional focusing three-dimensional laser scanning devices to differentiate laser wavelengths when scanning and detecting the same point, and identifies the wavelengths of the reflected lasers. The laser scanning control module controls the laser focus adjusting module to adjust the focusing points of multiple multi-directional focusing three-dimensional laser scanning devices. Through multi-directional focusing three-dimensional laser scanning, the influence of various interference factors in the tunnel on laser scanning is significantly reduced.

[0081] In one embodiment, the optimized federal asynchronous residual adjustment subsystem comprises:

[0082] The optimized federal weighted learning subsystem uses an optimized federal weighted learning model to process multiple sets of tunnel laser point cloud data respectively and acquires tunnel laser point cloud data after optimized processing.

[0083] The asynchronous residual error test adaptive subsystem combines the optimized federal weighted learning model with an asynchronous and residual error test adaptive adjustment model, optimally adjusts the processed tunnel laser point cloud data, constructs an optimized and adjusted tunnel laser point cloud data set, and obtains an optimized and adjusted tunnel laser point cloud data set.

[0084] The principle and effect of the above technical solution are as follows: the optimized federal asynchronous residual error adjustment system includes: an optimized federal weighted learning subsystem that uses an optimized federal weighted learning model to process multiple groups of tunnel laser point cloud data to obtain processed tunnel laser point cloud data; and an asynchronous residual error test adaptive subsystem that combines the optimized federal weighted learning model with an asynchronous and residual error test adaptive adjustment model to optimally adjust the processed tunnel laser point cloud data, construct an optimized and adjusted tunnel laser point cloud data set, and obtain an optimized and adjusted tunnel laser point cloud data set.

[0085] In one embodiment, the lining crack detection and interference reduction subsystem includes:

[0086] The lining crack detection and interference reduction subsystem accurately detects tunnel lining cracks based on the optimized and adjusted tunnel laser point cloud data set to obtain tunnel lining crack detection information, and reduces interference factors in the tunnel, including dust interference factors or exposed steel interference factors and structural joint interference factors.

[0087] The tunnel lining detection cloud transmission subsystem sends the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0088] The principle and effect of the above technical solution are as follows: the lining crack detection and interference reduction subsystem includes: a lining crack detection and interference reduction subsystem that accurately detects tunnel lining cracks based on the optimized and adjusted tunnel laser point cloud data set to obtain tunnel lining crack detection information, and reduces interference factors in the tunnel, including dust interference factors or exposed steel interference factors and structural joint interference factors; and a tunnel lining detection cloud transmission subsystem that sends the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform.

[0089] In one embodiment, the strong earthquake tunnel lining identification and warning subsystem includes:

[0090] The high-definition image panoramic magnification subsystem sets up a high-definition camera visual recognition device, collects high-resolution images of the strong earthquake tunnel lining, performs panoramic magnification image visual intelligent recognition, and obtains complete information of the strong earthquake tunnel lining cracks.

[0091] The strong earthquake tunnel lining crack warning subsystem can mark and warn the crack position and size of the tunnel lining crack according to the tunnel lining crack detection information and the strong earthquake tunnel lining crack identification information.

[0092] The principle and effect of the above technical solution are as follows: the strong earthquake tunnel lining identification and warning subsystem comprises: a high-definition image panoramic magnification subsystem, a high-definition camera visual identification device is arranged to collect high-resolution images of the strong earthquake tunnel lining, perform visual intelligent identification on the panoramic magnified images, and obtain complete information of the strong earthquake tunnel lining crack; the strong earthquake tunnel lining crack warning subsystem can mark and warn the crack position and size of the tunnel lining crack according to the tunnel lining crack detection information and the strong earthquake tunnel lining crack identification information; the detection precision of the tunnel lining crack can be significantly improved; the detection completeness and comprehensiveness of the strong earthquake tunnel lining crack are significantly improved; the high-definition camera visual identification device is arranged to collect high-resolution images of the strong earthquake tunnel lining, perform visual intelligent identification on the panoramic magnified images, and obtain complete information of the strong earthquake tunnel lining crack, including: the high-definition camera visual identification device is arranged to collect high-resolution images of the strong earthquake tunnel lining; the high-resolution images of the tunnel lining are spliced and enlarged according to a set ratio to obtain panoramic magnified images of the tunnel lining; the panoramic magnified images of the tunnel lining are used for visual intelligent identification to obtain complete information of the strong earthquake tunnel lining crack; the complete information of the strong earthquake tunnel lining crack includes: crack starting position, crack extension, crack branch, crack branch end, and crack extension end.

[0093] Although the embodiments of the present application have been disclosed as above, they are not limited to the application listed in the specification and the embodiments, and can be fully applied to various fields suitable for the present application, and additional modifications can be easily realized by those skilled in the art, and therefore the present application is not limited to specific details and the figures shown and described herein, without departing from the general concept defined by the claims and the equivalent scope.

Claims

1. A three-dimensional laser strong earthquake tunnel lining crack detection method, characterized in that, Comprise: S10, detect the tunnel by multi-directional focusing three-dimensional laser scanning device, obtain multiple groups of tunnel laser point cloud data; S20, use the optimized federal weighted learning model to process multiple groups of tunnel laser point cloud data respectively, combine the asynchronous and residual error test adaptive adjustment model, optimize and adjust the processed tunnel laser point cloud data, and obtain the optimized and adjusted tunnel laser point cloud data set; S30, according to the optimized and adjusted tunnel laser point cloud data set, accurately detect the tunnel lining crack, and reduce the interference factors in the tunnel; S40, collect high-resolution images of strong earthquake tunnel lining, perform panoramic enlarged image visual intelligent identification, combine strong earthquake tunnel lining crack identification information, and perform tunnel lining crack marking and crack position and size warning; S10 comprises: S101, adopt angle-adjusting non-collinear multi-directional laser scanning detection, and construct a multi-directional focusing three-dimensional laser scanning device; S102, detect the tunnel by the multi-directional focusing three-dimensional laser scanning device, and obtain multiple groups of tunnel laser point cloud data; The multi-directional focusing three-dimensional laser scanning device comprises a laser autonomous positioning module, a laser scanning automatic moving module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module and a laser focal point adjusting module; through multi-directional focusing three-dimensional laser scanning, the influence of interference factors in the tunnel on laser scanning is reduced; S20 comprises: S201, use the optimized federal weighted learning model to process multiple groups of tunnel laser point cloud data respectively, and obtain the processed tunnel laser point cloud data; S202, combine the optimized federal weighted learning model with the asynchronous and residual error test adaptive adjustment model, optimize and adjust the processed tunnel laser point cloud data, construct the optimized and adjusted tunnel laser point cloud data set, and obtain the optimized and adjusted tunnel laser point cloud data set; S30 comprises: S301, according to the optimized and adjusted tunnel laser point cloud data set, accurately detect the tunnel lining crack, obtain tunnel lining crack detection information, reduce the interference factors in the tunnel, and the interference factors in the tunnel include dust interference factors or exposed steel interference factors and structure joint interference factors; S302, send the tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform; S40 comprises: S401, set up a high-definition camera visual identification device, collect high-resolution images of strong earthquake tunnel lining, perform panoramic enlarged image visual intelligent identification, and obtain complete information of strong earthquake tunnel lining crack; S402, according to the tunnel lining crack detection information, combine the strong earthquake tunnel lining crack identification information, perform tunnel lining crack marking and crack position and size warning.

2. A three-dimensional laser strong motion tunnel lining crack detection system, characterized in that, Comprise: A multi-directional focusing three-dimensional scanning subsystem detects the tunnel by a multi-directional focusing three-dimensional laser scanning device, and obtains multiple groups of tunnel laser point cloud data; An optimized federal asynchronous residual error adjustment subsystem uses an optimized federal weighted learning model to process multiple groups of tunnel laser point cloud data respectively, combines an asynchronous and residual error test adaptive adjustment model, optimizes and adjusts the processed tunnel laser point cloud data, and obtains an optimized and adjusted tunnel laser point cloud data set; The lining crack detection and interference reduction sub-system accurately detects tunnel lining cracks and reduces interference factors in the tunnel according to the optimized and adjusted tunnel laser point cloud data set. The strong earthquake tunnel lining identification and warning sub-system collects high-resolution images of strong earthquake tunnel linings, performs panoramic zoom image visual intelligent identification, combines strong earthquake tunnel lining crack identification information, and performs tunnel lining crack marking and crack position and size warning. The multi-directional focusing three-dimensional scanning sub-system includes: The angle-adjustable multi-directional laser detection subsystem uses angle-adjustable non-collinear multi-directional laser scanning detection to construct a multi-directional focusing three-dimensional laser scanning device. The multi-directional focusing laser point cloud subsystem detects tunnels through the multi-directional focusing three-dimensional laser scanning device and obtains multiple sets of tunnel laser point cloud data. The multi-directional focusing three-dimensional laser scanning device includes a laser autonomous positioning module, a laser scanning automatic movement module, a three-dimensional laser scanning module, a laser scanning control module, a laser focusing monitoring module, and a laser focal point adjustment module. Through multi-directional focusing three-dimensional laser scanning, the influence of interference factors in the tunnel on laser scanning is reduced. The optimized federal asynchronous residual error adjustment sub-system includes: The optimized federal weighted learning subsystem uses an optimized federal weighted learning model to process multiple sets of tunnel laser point cloud data and obtain optimized and adjusted tunnel laser point cloud data. The asynchronous residual error test adaptive subsystem combines the optimized federal weighted learning model with an asynchronous and residual error test adaptive adjustment model to optimize and adjust the tunnel laser point cloud data, construct an optimized and adjusted tunnel laser point cloud data set, and obtain the optimized and adjusted tunnel laser point cloud data set. The lining crack detection and interference reduction sub-system includes: The lining crack interference reduction detection subsystem accurately detects tunnel lining cracks according to the optimized and adjusted tunnel laser point cloud data set, obtains tunnel lining crack detection information, and reduces interference factors in the tunnel, including dust interference factors or exposed steel interference factors and structural joint interference factors. The tunnel lining detection cloud transmission subsystem sends tunnel lining crack detection information to a tunnel detection control center and a tunnel lining crack detection cloud platform. The strong earthquake tunnel lining identification and warning sub-system includes: The high-resolution image panoramic zoom subsystem sets up a high-definition camera visual identification device, collects high-resolution images of strong earthquake tunnel linings, performs panoramic zoom image visual intelligent identification, and obtains complete information about strong earthquake tunnel lining cracks. The strong earthquake tunnel lining crack warning subsystem marks tunnel lining cracks and warns about crack positions and sizes according to tunnel lining crack detection information and strong earthquake tunnel lining crack identification information.

Images