Intelligent subway tunnel lining multiple disease detection equipment and detection method

Abstract

The invention discloses an intelligent subway tunnel lining multiple disease detection device and a detection method thereof, belonging to the technical field of tunnel lining maintenance detection, comprising a vehicle body, wherein the vehicle body is provided with a laser vibrometer, a photographic unit, a laser scanner and a computer medium; the invention adopts visual detection and laser vibrometer to collect tunnel lining surface and internal disease information, uses an established disease detection model and fundamental frequency damage degree to extract disease related information from the collected data information, and combines laser scanning to reconstruct a tunnel three-dimensional structure model, generates disease risk related reports and visual displays, the equipment can realize high-precision and intelligent detection of tunnel lining, and can be connected to an engineering management platform, so as to realize real-time detection of tunnel lining safety.

Description

Technical Field

[0001] The invention relates to the technical field of tunnel lining maintenance detection, and in particular to intelligent subway tunnel lining multiple disease detection equipment and a detection method thereof. Background Art

[0002] As a permanent supporting structure for tunnel engineering, tunnel lining is an important guarantee for the normal operation and maintenance of traffic tunnels. With the increase of years of use, various defects such as cracks, cavities, water leakage, peeling, deformation and misalignment have begun to appear one after another due to various internal and external factors. If they are not inspected, repaired and reinforced in time, it will inevitably endanger the driving safety of traffic vehicles. At present, the commonly used inspection method for subway tunnels is the traditional manual inspection method. This method is affected by the working status and experience level of the inspectors, and the quality of the inspection data cannot be guaranteed. At the same time, it consumes a lot of manpower and material resources and is inefficient. It is increasingly difficult to adapt to the actual needs of tunnel inspection projects.

[0003] There are various types of tunnel lining defects, but the relevant technical methods are mostly limited to single defect detection. For example, the detection methods based on ultrasound and geological radar are suitable for the detection of cavities inside the lining, the detection methods based on infrared are suitable for the detection of surface water leakage, and the detection of crack defects based on machine vision. The above methods have more or less limitations. For example, when the geological radar detects internal defects, it needs to be close to the lining to ensure the detection accuracy, so the work efficiency is limited. The existing detection methods usually conduct regular inspections at fixed points in key areas, which makes it difficult to achieve intelligent batch detection of tunnels.

[0004] In view of the above problems, the present invention document proposes an intelligent subway tunnel lining multiple disease detection device and a detection method thereof. Summary of the invention

[0005] The present invention provides an intelligent subway tunnel lining multiple disease detection device and a detection method thereof, which solves the above-mentioned problems.

[0006] The present invention provides the following technical solutions:

[0007] An intelligent subway tunnel lining multiple disease detection device comprises a car body, on which a laser vibrometer, a photographic unit, a laser scanner and a computer medium are arranged;

[0008] The laser vibrometer is used to obtain vibration data of the subway tunnel lining structure;

[0009] The photographing unit is used to obtain subway tunnel lining surface image data;

[0010] The laser scanner is used to emit a laser point cloud to obtain the laser point position coordinates to construct a three-dimensional model of the subway tunnel lining;

[0011] The computer medium is used to coordinate and control the photographing unit and the laser vibrometer.

[0012] In a possible design, a control unit is disposed on the vehicle body, a pan-tilt head is fixedly mounted on the vehicle body, and a top end of the pan-tilt head is connected to a laser vibrometer.

[0013] In a possible design, the photography unit includes a plurality of CCD linear array cameras, two groups of strip-shaped LED high-brightness light sources and a camera workbench.

[0014] In a possible design, the camera workbench is designed as an arched structure, and the plurality of CCD line array cameras are fixedly mounted on the camera workbench, and the plurality of CCD line array cameras are distributed at equal distances.

[0015] In a possible design, the two groups of the strip-shaped LED high-brightness light sources are fixedly mounted on the outer wall of the camera workbench, and the two groups of the strip-shaped LED high-brightness light sources are respectively arranged on both sides of the CCD linear array camera, and the positions of several of the strip-shaped LED high-brightness light sources in the two groups are respectively staggered.

[0016] An intelligent subway tunnel lining multiple disease detection method, the detection method comprising the following steps:

[0017] S1. Detect surface defects of subway tunnel linings, obtain and store surface defect information;

[0018] S2. Conduct internal disease detection on subway tunnel lining, obtain internal disease information and store it;

[0019] S3. Statistical analysis of internal and external diseases.

[0020] In a possible design, step S1 specifically includes the following steps:

[0021] S1.1. Obtain subway tunnel lining surface image data, inspect and screen it, and select surface disease images;

[0022] S1.2. Based on the selected surface damage images, a lining surface damage dataset is created;

[0023] S1.3. Based on the lining surface disease dataset, train a deep learning model to establish a disease detection model;

[0024] S1.4. Input the acquired subway tunnel lining surface image data into the defect detection model, and splice the surface defect image corresponding to the tunnel mileage information to extract and obtain defect-related feature information, which specifically includes area, length, width and position;

[0025] S1.5. Store the surface disease image and disease-related characteristic information in a computer medium.

[0026] In a possible design, step S2 specifically includes the following steps:

[0027] S2.1. Establish the relationship between the vibration data of subway tunnel lining structure and structural stiffness;

[0028] S2.2. Establish the deterioration degree describing the structural damage based on the structural stiffness and damage mechanics to describe the internal defects of the lining;

[0029] S2.3. Based on the conversion of the vibration data of the subway tunnel lining structure into the deterioration degree describing the internal defects of the lining, the internal defect information is obtained, and the internal defect information is stored in a computer medium.

[0030] In a possible design, step S3 specifically includes the following steps:

[0031] S3.1. Based on the tunnel model constructed by laser scanner and combined with the spatial positioning of tunnel mileage, the surface disease image and disease-related feature information are generated on the three-dimensional tunnel model;

[0032] S3.2, based on the tunnel model constructed by laser scanner, combined with the tunnel mileage spatial positioning, the internal disease information is generated on the three-dimensional tunnel model;

[0033] S3.3. Statistically analyze the surface disease images, disease-related characteristic information and internal disease information in the computer medium to generate reports.

[0034] In a possible design, in step S1.2, the specific steps of preparing the lining surface disease dataset based on the selected surface disease image are as follows:

[0035] The surface disease images were manually annotated, label images were made using Labelme, and the training set, validation set, and test set were divided in a ratio of 8:1:1.

[0036] It is to be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the invention.

[0037] The present invention uses visual inspection and laser vibration measurement to collect surface and internal disease information of tunnel linings, uses the established disease detection model and fundamental frequency damage degree to extract disease-related information from the collected data information, and combines laser scanning to reconstruct the three-dimensional structural model of the tunnel to generate disease risk-related reports and visual displays. This equipment can achieve high-precision and intelligent detection of tunnel linings, and can also be connected to the engineering management platform to achieve real-time detection of tunnel lining safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A system diagram of an intelligent subway tunnel lining multiple disease detection device provided by an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of the structure of an intelligent subway tunnel lining multiple disease detection device provided by an embodiment of the present invention;

[0040] Figure 3 A system diagram of a disease detection model for an intelligent subway tunnel lining multiple disease detection method provided by an embodiment of the present invention.

[0041] Reference numerals:

[0042] 11. Vehicle body; 12. Control unit; 21. CCD linear array camera; 22. Camera workbench; 23. Strip LED high-brightness light source; 31. Laser vibrometer; 32. PTZ; 4. Laser scanner; 5. Computer medium. DETAILED DESCRIPTION

[0043] The embodiments of the present invention are described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific implementation disclosed below.

[0045] Example 1

[0046] Reference Figure 1 and Figure 2 The intelligent subway tunnel lining multiple disease detection device of the present embodiment includes a vehicle body 11 on which a laser vibrometer 31, a photographic unit, a laser scanner 4 and a computer medium 5 are arranged.

[0047] The laser vibrometer 31 , the photographic unit, the laser scanner 4 and the computer medium 5 are all detachably assembled with the vehicle body 11 , such as being fixedly connected by bolts. The vehicle body 11 is equipped with four rollers and four handles.

[0048] Specifically, the vehicle body 11 control unit 12 can set different travel modes and different travel speeds according to detection requirements. The vehicle body 11 control unit 12 is built into the vehicle body 11 and is equipped with a signal terminal, which can be remotely controlled by a remote control device.

[0049] The laser vibrometer 31 is used to obtain vibration data of the subway tunnel lining structure.

[0050] Specifically, the laser vibrometer 31 is a Doppler laser vibrometer 31 , and the laser vibrometer 31 is controlled by the computer medium 5 via the ploytec software.

[0051] The photography unit is used to obtain subway tunnel lining surface image data.

[0052] The laser scanner 4 is used to emit a laser point cloud to obtain the position coordinates of the laser points so as to construct a three-dimensional model of the subway tunnel lining.

[0053] The computer medium 5 is used to coordinate and control the photographing unit and control the laser vibrometer 31 .

[0054] A control unit 12 is disposed on the vehicle body 11 , a pan / tilt platform 32 is fixedly mounted on the vehicle body 11 , and a top end of the pan / tilt platform 32 is connected to a laser vibrometer 31 .

[0055] Specifically, the photographing unit includes a plurality of CCD linear array cameras 21 , two groups of strip-shaped LED high-brightness light sources 23 and a camera workbench 22 .

[0056] Specifically, the camera workbench 22 is designed as an arched structure, and a plurality of CCD line array cameras 21 are fixedly installed on the camera workbench 22 , and the plurality of CCD line array cameras 21 are distributed at equal distances.

[0057] Specifically, two groups of strip-shaped LED highlight light sources 23 are fixedly mounted on the outer wall of the camera workbench 22, and the two groups of strip-shaped LED highlight light sources 23 are respectively arranged on both sides of the CCD linear array camera 21, and the positions of several strip-shaped LED highlight light sources 23 in the two groups are respectively staggered.

[0058] Working principle: the vehicle body 11 is placed on the subway track, the computer medium 5 is connected to the corresponding CCD camera, the strip LED high-brightness light source 23 is turned on, the camera focal length is adjusted, and the relevant parameters are adjusted to ensure the clarity of the image. The vehicle body 11 moves along the rail at a constant speed or a fixed distance according to the pre-set travel mode, and the image of the tunnel lining surface is collected. The collected data is saved in the computer medium 5. The pan / tilt 32 adjusts the laser vibrometer 31 to align with the point to be measured and fixes the posture. The laser vibrometer 31 emits laser, focuses the laser according to the signal feedback from the device end, adjusts the parameters, collects the lining vibration signal fed back by the laser, and stores the vibration data.

[0059] Example 2

[0060] The present embodiment provides an intelligent subway tunnel lining multiple disease detection method, the detection method comprising the following steps:

[0061] S1. Detect surface defects of subway tunnel linings, obtain surface defect information and store it:

[0062] S1.1. Obtain subway tunnel lining surface image data, inspect and screen it, and select surface disease images.

[0063] Specifically, the acquired subway tunnel lining surface image data includes lining surface defects such as cracks, peeling, and water leakage. The subway tunnel lining surface image data is inspected and screened based on a classification detection network, such as the commonly used ResNet series and YOLO series classification detection networks.

[0064] S1.2. Based on the selected surface damage images, a lining surface damage dataset is created.

[0065] S1.3. Based on the lining surface disease dataset, a deep learning model is trained to establish a disease detection model.

[0066] Specifically, Figure 3 As shown in the figure, the disease detection model is constructed with convolution blocks, pooling blocks, normalization blocks, activation blocks and other basic blocks. Specifically, the disease detection model consists of 9 combination blocks, which are divided into an encoder part and a decoder part. Each combination block consists of two convolution blocks, two normalization blocks and two activation blocks. The convolution block is a two-dimensional convolution, and the activation function is a ReLU function, which is specifically described as fx = max(0,x). Pooling is performed after each combination block of the encoder part, and upsampling is performed after each combination block of the decoder part to extract disease features.

[0067] S1.4. Input the acquired subway tunnel lining surface image data into the defect detection model, and splice the surface defect image corresponding to the tunnel mileage information, and extract it to obtain the defect-related feature information, which specifically includes area, length, width and position.

[0068] Specifically, the disease detection model performs feature fusion through jump connections. The feature fusion block is composed of self-attention and convolution fusion. Up and down sampling is required between feature layers to adjust the image size. Feature fusion is used to integrate the high and low-bit feature information of the network to improve the detection accuracy of the model. Then, the surface disease image corresponding to the tunnel mileage information is spliced ​​and extracted to obtain disease-related feature information, such as area, length, width, and position.

[0069] S1.5, storing the surface disease image and disease-related characteristic information in the computer medium 5;

[0070] S2. Conduct internal defect detection on subway tunnel lining, obtain internal defect information and store it:

[0071] S2.1. Establish the relationship between the vibration data of subway tunnel lining structure and structural stiffness;

[0072] S2.2. Establish the deterioration degree describing the structural damage based on the structural stiffness and damage mechanics to describe the internal defects of the lining;

[0073] S2.3, converting the vibration data of the subway tunnel lining structure into the deterioration degree describing the internal disease of the lining, obtaining the internal disease information, and storing the internal disease information in the computer medium 5;

[0074] S3. Statistical analysis of internal and external diseases:

[0075] S3.1, based on the tunnel model constructed by the laser scanner 4, combined with the tunnel mileage spatial positioning, the surface disease image and disease-related feature information are generated on the three-dimensional tunnel model;

[0076] S3.2, based on the tunnel model constructed by the laser scanner 4, combined with the tunnel mileage spatial positioning, the internal disease information is generated onto the three-dimensional tunnel model;

[0077] S3.3. Statistically analyze the surface disease images, disease-related characteristic information and internal disease information in the computer medium 5 to generate a report.

[0078] Specifically, the computer medium is used as a terminal to connect to the server via the Internet, and the detection data can be uploaded to the server.

[0079] In step S1.2, the specific steps of making a lining surface disease dataset based on the selected surface disease image are as follows:

[0080] The surface disease images were manually annotated, label images were made using Labelme, and the training set, validation set, and test set were divided in a ratio of 8:1:1.

[0081] In S2.1, the relationship between the vibration data of the subway tunnel lining structure and the structural stiffness is established, and its expression is:

[0082]

[0083] In the formula, is the natural vibration frequency of the structure, is the displacement mode of the structural vibration.

[0084] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0085] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0086] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. An intelligent subway tunnel lining multiple disease detection device, comprising a car body (11), Features: The vehicle body (11) is provided with a laser vibrometer (31), a photographic unit, a laser scanner (4) and a computer medium (5); The laser vibrometer (31) is used to obtain vibration data of the subway tunnel lining structure; The photographing unit is used to obtain subway tunnel lining surface image data; The laser scanner (4) is used to emit a laser point cloud to obtain the position coordinates of the laser points, so as to construct a three-dimensional model of the subway tunnel lining; The computer medium (5) is used to coordinate and control the photographing unit and control the laser vibrometer (31); The vehicle body (11) is provided with a control unit (12), and the control unit (12) sets different travel modes and different travel speeds according to detection requirements. The control unit (12) is built into the vehicle body (11) and is equipped with a signal terminal, and is remotely controlled by a remote control device; The detection method applied to the intelligent subway tunnel lining multiple disease detection device comprises the following steps: S1. Detect surface defects of subway tunnel lining, obtain and store surface defect information; specifically, the following steps are included: S1.

1. Obtain subway tunnel lining surface image data, inspect and screen it, and select surface disease images; S1.

2. Based on the selected surface damage images, a lining surface damage dataset is created; S1.

3. Based on the lining surface disease data set, a deep learning model is trained to establish a disease detection model; wherein the disease detection model is composed of 9 combination blocks, which are divided into an encoder part and a decoder part. Each combination block is composed of two convolution blocks, two standardization blocks, and two activation blocks. The convolution block is a two-dimensional convolution, and the activation function is a ReLU function. Pooling is performed after each combination block of the encoder part, and upsampling is performed after each combination block of the decoder part to extract disease features; S1.

4. Input the acquired subway tunnel lining surface image data into the disease detection model, and splice the surface disease image corresponding to the tunnel mileage information, and extract to obtain the disease-related feature information, the disease-related feature information specifically includes area, length, width and position; wherein, the disease detection model performs feature fusion through jump connection, the feature fusion block is composed of self-attention and convolution fusion, up and down sampling is performed between feature layers to adjust the image size, feature fusion is used to integrate the high and low bit feature information of the network, improve the model detection accuracy, and then splice the surface disease image corresponding to the tunnel mileage information, and extract to obtain the disease-related feature information; S1.5, storing the surface disease image and disease-related characteristic information in a computer medium (5); S2, performing internal disease detection on the subway tunnel lining, obtaining and storing internal disease information; specifically comprising the following steps: S2.

1. Establish the relationship between the vibration data of subway tunnel lining structure and structural stiffness; S2.

2. Establish the deterioration degree describing the structural damage based on the structural stiffness and damage mechanics to describe the internal defects of the lining; S2.3, converting the vibration data of the subway tunnel lining structure into the deterioration degree describing the internal defects of the lining, obtaining the internal defect information, and storing the internal defect information in the computer medium (5); S3. Statistical analysis of internal and external diseases; specifically including the following steps: S3.1, based on the tunnel model constructed by the laser scanner (4), combined with the tunnel mileage spatial positioning, the surface disease image and disease-related feature information are generated onto the three-dimensional tunnel model; S3.2, based on the tunnel model constructed by the laser scanner (4), combined with the tunnel mileage spatial positioning, the internal disease information is generated onto the three-dimensional tunnel model; S3.

3. Statistically analyze the surface disease images, disease-related characteristic information and internal disease information in the computer medium (5) to generate a report.

2. According to claim 1, an intelligent subway tunnel lining multiple disease detection device, It is characterized in that A pan head (32) is fixedly mounted on the vehicle body (11), and the top end of the pan head (32) is connected to the laser vibrometer (31).

3. According to claim 1, an intelligent subway tunnel lining multiple disease detection device, It is characterized in that The photographing unit comprises a plurality of CCD linear array cameras (21), two groups of strip-shaped LED high-brightness light sources (23) and a camera workbench (22).

4. According to claim 3, an intelligent subway tunnel lining multiple disease detection device, It is characterized in that The camera workbench (22) is designed as an arched structure, and the plurality of CCD line array cameras (21) are fixedly mounted on the camera workbench (22), and the plurality of CCD line array cameras (21) are distributed at equal distances.

5. According to claim 3, an intelligent subway tunnel lining multiple disease detection device, It is characterized in that The two groups of the strip-shaped LED high-brightness light sources (23) are both fixedly mounted on the outer wall of the camera workbench (22), and the two groups of the strip-shaped LED high-brightness light sources (23) are respectively arranged on both sides of the CCD linear array camera (21), and the positions of a plurality of the strip-shaped LED high-brightness light sources (23) in the two groups are respectively staggered.

6. According to claim 1, an intelligent subway tunnel lining multiple disease detection device, It is characterized in that In step S1.2, the specific steps of preparing the lining surface disease data set based on the selected surface disease image are as follows: The surface disease images were manually annotated, label images were made using Labelme, and the training set, validation set, and test set were divided in a ratio of 8:1:1.

Images