Tunnel visual facility identification system and storage medium thereof

Abstract

The invention discloses a tunnel visual facility identification system and a storage medium thereof. The tunnel visual facility identification system comprises an image acquisition module, an equipment identification module, a data storage module and a visual display module, the image acquisition module is provided with an improved photoelectric camera and is used for acquiring a high-definition image of the inner wall of the tunnel and transmitting data to the equipment identification module by means of a video transmission wire harness; and the equipment identification module is used for operating a target detection algorithm to identify tunnel visual facilities based on the images acquired by the image acquisition module, generating specific images, names, states and positions of the facilities and carrying out tunnel asset census work, and is connected with the data storage module and the visual display module through connecting wire harnesses. According to the system, the recognition precision is improved, the equipment recognition module optimizes parameters through a specific algorithm, and high-quality hardware configuration of the image acquisition module is combined, so that the recognition precision is greatly improved, missed detection and misjudgment are reduced, and reliable data are provided for tunnel management.

Description

Technical Field

[0001] The present invention relates to the technical field of visual recognition, and particularly to a tunnel visual facility recognition system and its storage medium. Background Art

[0002] Tunnel visual facility recognition refers to the use of specific technologies and equipment to accurately identify various facilities in the tunnel, detect their status, and present the results in a visual form to provide a basis for tunnel management. In terms of hardware, images of tunnel facilities are obtained through image acquisition devices such as cameras. At the software level, algorithms are used to analyze the collected images to identify the facility categories and detect whether there are any faults.

[0003] Existing tunnel facility recognition technologies have deficiencies in algorithms and hardware configurations, making it difficult to accurately identify various facilities and detect faults; in conventional highway, railway, and subway tunnel scenarios, the accuracy of identifying and fault detecting facilities such as lighting fixtures and fire-fighting equipment is low, resulting in a large number of missed detections and misjudgments. The existing technologies lack an efficient coordination mechanism among various modules and cannot achieve automated and intelligent patrol inspections. Taking the patrol inspection of highway tunnels as an example, it mainly relies on manual labor, and a single patrol inspection often takes more than 4 hours, with extremely low work efficiency. Existing tunnel facility recognition devices are not perfect in hardware design and technology application and are difficult to adapt to complex tunnel environments. In high-speed tunnels, stable and clear images cannot be obtained; in environments with weak signals, a lot of dust, or strong electromagnetic interference, the devices are prone to failures or data deviations and cannot operate stably and obtain accurate data, severely limiting their application scope. Therefore, a tunnel visual facility recognition system and its storage medium are proposed. Summary of the Invention

[0004] To overcome the above-mentioned defects of the prior art, the present invention provides a tunnel visual facility recognition system and its storage medium, which solves the problems in the prior art of being difficult to accurately identify various facilities and detect faults, unable to achieve automated and intelligent patrol inspections, and difficult to adapt to complex tunnel environments.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A tunnel visual facility recognition system includes an image acquisition module, a device recognition module, a data storage module, and a visual display module;

[0007] The image acquisition module: Configured with an improved optoelectronic camera, it is used to obtain high-definition images of the tunnel inner wall and transmit data to the device recognition module through a video transmission harness;

[0008] Device Identification Module: Based on the images obtained by the Image Acquisition Module, it runs a target detection algorithm to identify tunnel visualization facilities, generates specific images, names, statuses, and locations of the facilities, conducts a general survey of tunnel assets, and is connected to the Data Storage Module and the Visualization Display Module through a connection harness;

[0009] Data Storage Module: Used to store the image data of the Image Acquisition Module and the identification result data of the Device Identification Module, and is connected to the Device Identification Module through a connection harness;

[0010] Visualization Display Module: The display screen is connected to the Device Identification Module through a connection harness, and displays the original images taken, the tunnel facility identification results, and presents the device operation status and various detailed information of the facilities.

[0011] Preferably, the system is integrated into a tunnel visualization facility identification device, which includes an image acquisition component, a facility identification component, a data storage component, and a visualization display component.

[0012] Preferably, the image acquisition component includes an optoelectronic camera and an anti-shake support;

[0013] Optoelectronic Camera: Used to capture high-definition images of the tunnel inner wall;

[0014] Anti-shake Support: Utilizes anti-shake technology to support the optoelectronic camera to obtain high-definition and stable images under complex conditions in high-speed tunnels.

[0015] Preferably, the facility identification component includes a control panel, a GPU module, a satellite positioning module, an acceleration sensor, an interface, a heat dissipation module, and a power supply module;

[0016] Control Panel: Coordinates overall logical control and task scheduling, and coordinates the working timings of image acquisition, processing, storage, and display;

[0017] GPU Module: Runs a target detection algorithm to analyze images, meeting the real-time requirements of high-speed scenarios;

[0018] Satellite Positioning Module: Integrates GPS and inertial navigation modules, provides the geographical location information of the facilities, compensates for the problem of weak GPS signals in tunnels, and binds the facility identification results with longitude and latitude coordinates;

[0019] Acceleration Sensor: Dynamically senses the motion state of the device, and assists in optimizing the image acquisition and processing process;

[0020] Interface: Enables high-speed data interaction between different modules;

[0021] Heat Dissipation Module: Ensures the continuous and stable operation of the hardware;

[0022] Power Supply Module: Provides reliable power supply for the entire component.

[0023] Preferably, the data storage component includes a solid-state drive and a 4G communication module;

[0024] Solid-state drive: Stores the captured image data of the tunnel inner wall and the identification result data of tunnel facilities;

[0025] 4G communication module: Enables data transmission, facilitating remote monitoring and instant analysis.

[0026] Preferably, the visualization display component includes a capacitive display screen for showing the identification situation of tunnel facilities, facilitating on-site personnel to observe in real time;

[0027] The optoelectronic camera of the image acquisition module is equipped with a high quantum efficiency CMOS sensor, a narrowband filter and a linear polarizer lens are added, and a global shutter sensor is adopted to improve the light-sensing and shooting capabilities of the camera in complex lighting environments.

[0028] Preferably, the object detection algorithm optimizes the model parameters by calculating the loss function. The total loss L consists of the localization loss L loc and the confidence loss L conf , and the formula is where N is the number of prior boxes matched, x is the matching result between the prior box and the ground truth box, c is the predicted confidence, l is the predicted bounding box position, g is the position of the ground truth box, and α is a hyperparameter that adjusts the relative importance of the localization loss and the confidence loss.

[0029] Preferably, the anti-shake support estimates the camera movement based on feature point displacement. Let the coordinates of the feature points in two adjacent frames be (x1, y1) and (x2, y2) respectively. The feature point displacement vector (x1, y1) is the coordinate of a certain feature point in the first frame among two adjacent frames, and (x2, y2) is the coordinate of the same feature point in the second frame among two adjacent frames. By calculating the coordinate differences of the same feature point in the x and y directions in the two frames, this vector is obtained, which is used to estimate the movement state of the camera, and then drive the anti-shake support to perform compensation to ensure image stability.

[0030] Preferably, the GPS and inertial navigation module uses the extended Kalman filter algorithm to provide the geographical location information of the facilities. The state prediction equation is The observation update equation is where is the optimal estimate at the previous moment and the input u k The predicted state at the current moment, f(·) is the state transition function, which describes the evolution law of the system state from the k-1 moment to the k moment, is the optimal estimate of the system state at the k-1 moment, u kis the input control quantity of the system at time k, is the fused observation value z k the optimal estimate of the system state at time k after fusion, K k is the Kalman gain, used to balance the predicted value and the observation value z k the difference of, z k is the observation value at time k, h(·) is the observation function, which maps the system predicted state to the observation space for comparison with the actual observation value z k Compare.

[0031] Preferably, the storage medium of the tunnel visualization facility identification system stores computer instructions, which are used to implement the tunnel facility identification function of the tunnel visualization facility identification system as described in any one of claims 1-9 when executed by a processor, including controlling the image acquisition module to collect image data of facilities in the tunnel, identifying the types of facilities based on the image data through the device identification module, and displaying the identification results on the visualization display module.

[0032] Technical effects and advantages of a tunnel visualization facility identification system and its storage medium according to the present invention:

[0033] 1. For the present invention, the identification accuracy is improved. The device identification module optimizes parameters through a specific algorithm and combines with the high-quality hardware configuration of the image acquisition module to greatly improve the identification accuracy; in conventional highway, railway, and subway tunnel scenarios, the identification and fault detection accuracy of facilities such as lighting fixtures and fire-fighting equipment exceed 93%, reducing missed detections and misjudgments, and providing reliable data for tunnel management.

[0034] 2. For the present invention, the work efficiency is improved. Each module of the system closely cooperates to achieve automated and intelligent patrol inspection.

[0035] 3. For the present invention, the operation and maintenance cost is reduced. Technologies such as satellite positioning and 4G communication support remote operation and maintenance of extra-long tunnels. The rail vehicle automatically patrols with a small positioning error and can give early warnings of faults. The tunnel operation and maintenance cost is reduced by 30%, avoiding large-scale repairs and operation interruptions, and improving economic benefits.

[0036] 4. For the present invention, it adapts to complex environments. The optoelectronic camera is equipped with an anti-shake support, and the satellite positioning module integrates multiple technologies, enabling the system to adapt to complex tunnel environments. It can still operate stably and obtain accurate data in environments with high speed, weak signals, a lot of dust, and strong electromagnetic interference.

[0037] 5. For the present invention, it helps with intelligent management. The visualization display and data storage modules help with the intelligent management of tunnels. Operation and maintenance personnel can monitor the status of facilities in real time, and management personnel can analyze historical data to optimize management strategies, promoting the development of tunnel management towards intelligence and digitization. Description of the Drawings

[0038] Figure 1 It is the system block diagram of a tunnel visualization facility recognition system proposed by the present invention;

[0039] Figure 2 It is the schematic flow diagram of a tunnel visualization facility recognition system proposed by the present invention. Specific implementation manners

[0040] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0041] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, the elements defined by the statement "including..." do not exclude the existence of additional identical elements in the process, method, article or device including the said elements.

[0042] Embodiment 1

[0043] This embodiment provides a tunnel visualization facility recognition system and its storage medium for the implementation of routine highway tunnel inspections, including the following specific contents:

[0044] Experimental purpose:

[0045] To solve the problems of low efficiency, high missed inspection and misjudgment rate in traditional manual highway tunnel inspections, and to achieve fast and accurate tunnel facility inspections with the help of the tunnel visualization facility recognition system.

[0046] Experimental steps:

[0047] (1) Equipment installation and debugging. The staff of the highway management department install the tunnel visualization facility recognition equipment on a specially made inspection vehicle, check whether the connections of each module are stable, and debug the shooting parameters of the optoelectronic camera, the target detection algorithm parameters, etc., to ensure the normal operation of the equipment. Simulate environmental parameters such as different light intensities and tunnel wall materials, and test the stability of the equipment under various conditions;

[0048] (2) Real-time monitoring: Start the inspection vehicle and drive it into the tunnel at a constant speed of 60 km / h. The optoelectronic camera of the image acquisition component continuously takes images of the tunnel inner wall at 30 frames per second with the support of the anti-shake support, and transmits the image data to the device recognition module through the video transmission harness. The device recognition module uses the object detection algorithm to calculate the loss function Optimize the model parameters to identify tunnel facilities. The data storage component synchronously stores the images and recognition results, and the 4G communication module transmits the data to the monitoring center. The visualization display component displays the recognition results. In the simulation experiment, different numbers and distributions of facility samples are set to verify the recognition ability of the system.

[0049] When the optoelectronic camera takes images of the tunnel inner wall at 30 frames per second, the anti-shake support calculates the feature point displacement vector based on the feature point coordinates (x1, y1), (x2, y2) of two adjacent frames If the camera shakes due to the bump of the inspection vehicle, the displacement of adjacent two-frame feature points increases, and the system drives the anti-shake support to make a reverse compensation according to Δd to offset the vibration;

[0050] For example: If the feature point offsets 5 pixels in the x direction (x2 - x1 = 5) and 3 pixels in the y direction (y2 - y1 = 3), the anti-shake support moves the camera in the reverse direction (-5, -3) pixels through mechanical adjustment or electronic image stabilization technology to ensure the stability of the image and provide a clear input for the object detection algorithm.

[0051] Experimental results: The system successfully identified 50 lighting fixtures in the tunnel, and 8 of them had insufficient brightness; it identified 20 fire-fighting devices and recorded their status and locations. Compared with the traditional manual inspection, the inspection time was shortened from 4 hours to 48 minutes, the efficiency was increased by 80%, and the recognition accuracy reached 98%, greatly improving the efficiency and accuracy of highway tunnel inspection.

[0052] Example 2

[0053] This embodiment provides a tunnel visualization facility recognition system and its storage medium for the implementation of monitoring during the construction stage of railway tunnels, including the following specific contents:

[0054] Experimental purpose:

[0055] In the complex environment of railway tunnel construction, realize the real-time monitoring of temporary ventilation equipment and emergency lighting facilities, ensure the construction safety and progress, and avoid construction delays caused by facility failures.

[0056] Experimental steps:

[0057] (1) Equipment deployment and parameter setting: The railway tunnel construction party deploys the tunnel visualization facility identification equipment in the tunnel. According to the characteristics of the construction area, the coordinate parameters of the satellite positioning module are set, and the filtering and photosensitive parameters of the optoelectronic camera are debugged to ensure that the equipment adapts to the construction environment;

[0058] (2) Real-time monitoring: After the equipment runs, the optoelectronic camera of the image acquisition module uses a high quantum efficiency CMOS sensor, a narrowband filter, and a linearly polarized lens to collect images inside the tunnel. The equipment identification module analyzes the images to identify the status of temporary ventilation equipment and emergency lighting facilities. The satellite positioning module marks the location of the facilities, and the data storage component records the monitoring data. The system feeds back the facility status to the construction personnel in real time. In the simulation experiment, artificial fault scenarios of ventilation equipment and lighting facilities are set to test the fault detection ability of the system.

[0059] Experimental results: In the simulated construction environment, the accuracy rate of the system for detecting ventilation equipment faults reaches 96%, and the accuracy rate for detecting lighting facility faults is 95%. During the construction period, the system timely detected 3 ventilation equipment faults and 5 lighting facility damages. The construction personnel quickly responded for repair, avoiding delays in the construction progress caused by facility faults and ensuring construction safety. The system analyzes the historical data of the facility status, providing data support for the construction party to optimize the construction process. The simulation experiment shows that the construction efficiency is increased by 18%, and in actual construction, it is increased by 15%.

[0060] Example 3

[0061] This example provides a tunnel visualization facility identification system and its storage medium for the implementation of long tunnel remote operation and maintenance, including the following specific contents:

[0062] Experimental purpose:

[0063] To solve the problems of difficult management of long tunnels, high cost of traditional operation and maintenance, and slow response speed, realize remote operation and maintenance, and improve the efficiency and safety of tunnel management.

[0064] Experimental steps:

[0065] (1) Rail car and equipment installation: The long tunnel management unit installs the tunnel visualization facility identification equipment on the automatic inspection rail car, lays the track and debugs the equipment to ensure the stable operation of the rail car in the tunnel and the normal function of each module. Simulate different signal strength areas in the tunnel to test the performance of the satellite positioning module and the 4G communication module;

[0066] (2) Remote operation and maintenance work: The rail vehicle conducts 24-hour uninterrupted inspections along the preset track. The image acquisition module collects images, and the equipment recognition module analyzes the images to identify facilities. When the rail vehicle enters an area with weak signals, the satellite positioning module compensates for positioning with the help of the inertial navigation module. The data storage component stores data, and the 4G communication module transmits key information to the remote operation and maintenance center. The operation and maintenance personnel monitor the tunnel facilities through the visual interface, and the system issues fault warnings based on data analysis. In the simulation experiment, multiple facility fault scenarios are set to test the warning ability of the system.

[0067] When the rail vehicle enters an area with weak signals, the satellite positioning module uses the extended Kalman filter algorithm to achieve precise positioning and state prediction through the following equations. Based on the optimal estimate at time k-1 (such as the position and speed of the rail vehicle at the previous moment) and the input u k (the acceleration provided by the inertial navigation module), predict the state at time k When the rail vehicle is moving at a constant speed, predict the current position by combining the position and speed at the previous moment;

[0068] Observation update, The GPS shows that the position deviation of the rail vehicle is large. Through K k Fuse the inertial navigation data to correct the predicted value, make the positioning error less than 5 meters, ensure the accurate binding of the facility position and the longitude and latitude coordinates in remote operation and maintenance, and achieve fault warning.

[0069] Experimental results: The simulation experiment shows that in the area where the signal strength is lower than 10 dBm, the positioning error of the satellite positioning module is less than 5 meters. Through remote operation and maintenance, the simulated tunnel operation and maintenance cost is reduced by 32%, and the actual reduction is 30%. The fault response time is shortened from several hours in the simulation to 20 minutes, and the actual shortening is within half an hour. The facility fault discovery rate is increased by 25% in the simulation and 20% in reality, effectively ensuring the safe operation of the extra-long tunnel and improving the intelligent level of tunnel management.

[0070] Example 4

[0071] This example provides a tunnel visualization facility recognition system and its storage medium for the implementation of urban subway tunnel inspections, including the following specific contents:

[0072] Experimental purpose:

[0073] Solve the problem of difficult inspections caused by limited space and strong electromagnetic interference in the subway tunnel. With the help of the tunnel visualization facility recognition system, improve the inspection efficiency of the subway tunnel, eliminate potential safety hazards, and ensure the normal operation of the subway.

[0074] Experimental steps:

[0075] (1) Equipment Adaptation and Installation: The urban subway operation company installs the tunnel visualization facility identification equipment on the subway maintenance vehicle. For the electromagnetic interference environment in the subway tunnel, the equipment is optimized and debugged for anti-interference to ensure its stable operation. Simulate electromagnetic interference scenarios of different degrees to test the anti-interference performance of the equipment;

[0076] (2) Inspection Operation: During the non-operation period of the subway, the maintenance vehicle slowly travels along the track and the system starts. The image acquisition module collects images, and the equipment identification module identifies the facilities. The visualization display component shows the identification results. The maintenance personnel locate and repair the faulty facilities based on the results. The data storage component stores the inspection data to provide support for subsequent analysis. In the simulation experiment, set different types and quantities of subway tunnel facility faults to test the identification and location capabilities of the system.

[0077] Experimental Results: In the simulation experiment, when the electromagnetic interference intensity reaches 50 V / m, the recognition accuracy rate of the system for evacuation indicator signs and power supply cable joint faults still reaches over 93%. The system discovers 2 damaged evacuation indicator signs and 3 overheated power supply cable joints during one inspection. Through the application of the system, the simulated improvement of the subway tunnel inspection efficiency is 65%, and the actual improvement is 60%. The elimination rate of potential safety hazards reaches 98% in simulation and 95% in reality, providing data support for the subway operation company to optimize the equipment maintenance plan and ensuring the safe and stable operation of the subway.

[0078] In multiple key scenarios of tunnel operation and maintenance and construction management, the tunnel visualization facility identification system demonstrates excellent performance and significant application value, and the system is effectively applied in different types of tunnel scenarios. In the routine inspection scenario of highway tunnels, it solves the problems of low efficiency and poor accuracy of traditional manual inspections; during the construction stage of railway tunnels, it realizes real-time monitoring of temporary facilities to ensure construction safety and progress; for extra-long tunnels, with remote operation and maintenance, it breaks through the geographical limitations of management and reduces operation and maintenance costs; for urban subway tunnels, it overcomes the problems of limited space and strong electromagnetic interference and improves the inspection quality.

[0079] All four groups of embodiments utilize the core technologies of the tunnel visualization facility identification system. The image acquisition module uses an improved optoelectronic camera equipped with an anti-shake support, a high quantum efficiency CMOS sensor, etc. to obtain high-quality images; the device identification module uses object detection algorithms and related loss functions to optimize the model for accurate facility identification; the satellite positioning module integrates GPS and inertial navigation to ensure accurate positioning; the data storage and transmission module ensures secure data storage and timely transmission, and the visualization display module provides intuitive information for operators; through simulation experiments and practical applications, the system has achieved remarkable results in various scenarios. The accuracy of identification and fault detection is high. The accuracy of identifying highway tunnel lighting fixtures and fire-fighting equipment reaches 98.5% and 97.8% respectively, and the accuracy of fault detection of railway tunnel ventilation equipment and lighting facilities is 96% and 95% respectively. The accuracy of identifying evacuation indication signs and power supply cable joints in subway tunnels under strong electromagnetic interference exceeds 93%. At the same time, the work efficiency is greatly improved. The inspection efficiency of highway tunnels is increased by 80%, the construction efficiency of railway tunnels is increased by 15%, the operation and maintenance cost of extra-long tunnels is reduced by 30%, and the inspection efficiency of subway tunnels is increased by 60%, effectively ensuring the safe and stable operation and efficient management of various tunnels.

[0080] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0081] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0082] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0083] In addition, the functional modules in each embodiment of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0084] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the said claims.

[0085] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A tunnel visualization facility identification system, characterized in that, It includes an image acquisition module, a device identification module, a data storage module, and a visualization display module; Image acquisition module: Configure an improved optoelectronic camera to obtain high-definition images of the tunnel inner wall, and transmit data to the device identification module through a video transmission wire harness; Device identification module: Based on the images obtained by the image acquisition module, run an object detection algorithm to identify tunnel visualization facilities, generate specific images, names, statuses, and locations of the facilities, conduct a tunnel asset census, and at the same time connect to the data storage module and the visualization display module through a connection wire harness; Data storage module: Used to store the image data of the image acquisition module and the identification result data of the device identification module, and connect to the device identification module through a connection wire harness; Visualization display module: The display screen is connected to the device identification module through a connection wire harness to display the original captured images, tunnel facility identification results, present the device operation status, and the detailed information of the facilities.

2. The tunnel visualization facility identification system according to claim 1, wherein The image acquisition component includes an optoelectronic camera and an anti-shake support; Optoelectronic camera: Used to capture high-definition images of the tunnel inner wall; Anti-shake support: Apply anti-shake technology to support the optoelectronic camera to obtain high-definition and stable images under the complex working conditions of high-speed tunnels.

3. The tunnel visualization facility identification system according to claim 1, characterized in that, The facility identification component includes a control panel, a GPU module, a satellite positioning module, an acceleration sensor, an interface, a heat dissipation module, and a power supply module; Control panel: Coordinate the overall logic control and task scheduling, and coordinate the working timings of image acquisition, processing, storage, and display; GPU module: Run an object detection algorithm to analyze the images and meet the real-time requirements of high-speed scenarios; Satellite positioning module: Integrate GPS and inertial navigation modules to provide the geographical location information of the facilities, compensate for the problem of weak GPS signals in the tunnel, and bind the facility identification results with longitude and latitude coordinates; Acceleration sensor: Dynamically sense the motion state of the device and assist in optimizing the image acquisition and processing process; Interface: Realize high-speed data interaction between different modules; Heat dissipation module: Ensure the continuous and stable operation of the hardware; Power supply module: Provide reliable power supply for the entire component.

4. The tunnel visualization facility identification system according to claim 1, wherein The data storage component includes a solid-state drive and a 4G communication module; Solid-state drive: Store the image data of the tunnel inner wall captured and the tunnel facility identification result data; 4G communication module: Realize data transmission for convenient remote monitoring and instant analysis.

5. The tunnel visualization facility identification system according to claim 1, wherein, The visualization display component includes a capacitive display screen, which is used to display the identification situation of tunnel facilities for convenient on-site personnel to observe in real time; The optoelectronic camera of the image acquisition module is equipped with a high quantum efficiency CMOS sensor, equipped with a narrowband filter and a linear polarizer lens, and uses a global shutter sensor to improve the light-sensing and shooting capabilities of the camera in complex lighting environments.

6. The tunnel visualization facility identification system according to claim 1, characterized in that, The target detection algorithm optimizes the model parameters by calculating the loss function. The total loss L consists of the localization loss L loc and the confidence loss L conf . The formula is where N is the number of prior boxes matched, x is the matching result between the prior box and the ground truth box, c is the predicted confidence, l is the predicted bounding box position, g is the position of the ground truth box, and α is a hyperparameter that adjusts the relative importance of the localization loss and the confidence loss.

7. The tunnel visualization facility recognition system according to claim 3, characterized in that, The anti-shake support estimates the camera motion based on the displacement of feature points. Suppose the coordinates of feature points in two adjacent frames of images are (x1, y1) and (x2, y2) respectively, and the displacement vector of the feature points (x1, y1) is the coordinates of a certain feature point in the first frame among two adjacent frames of images, and (x 2 , y 2 ) are the coordinates of the same feature point in the second frame among two adjacent frames of images. By calculating the coordinate differences of the same feature point in the x and y directions in the two frames of images, this vector is obtained, which is used to estimate the motion state of the camera, and then drive the anti-shake support to perform compensation to ensure image stability.

8. The tunnel visualization facility recognition system according to claim 4, wherein, The GPS and inertial navigation module uses the extended Kalman filter algorithm to provide the geographical location information of the facility; the state prediction equation is The observation update equation is Where is the optimal estimate based on the previous moment and the input u k The predicted state at the current moment, f(·) is the state transition function, which describes the evolution law of the system state from the (k - 1)th moment to the kth moment. is the optimal estimate of the system state at the (k - 1)th moment, u k is the input control quantity of the system at the kth moment, is the optimal estimate of the system state at the kth moment after fusing the observation value z k , K k is the Kalman gain, which is used to balance the difference between the predicted value and the observation value z k , z k is the observation value at the kth moment, h(·) is the observation function, which maps the system predicted state to the observation space for comparison with the actual observation value z k .

9. A storage medium of a tunnel visualization facility identification system according to any one of claims 1-8, characterized in that, The storage medium of the tunnel visualization facility identification system stores computer instructions. When the computer instructions are executed by a processor, they are used to implement the tunnel facility identification function of the tunnel visualization facility identification system as described in any one of claims 1-8, including controlling the image acquisition module to acquire the image data of the facilities in the tunnel, identifying the facility types based on the image data through the device identification module, and displaying the identification results on the visualization display module.

Images