Tunnel leakage water automatic detection device integrated with visual identification

Abstract

The utility model discloses a kind of tunnel leakage water automatic detection devices of integrated visual identification, including running track, inspection car and electrical bin, running track is hoisted along the top wall of tunnel;Electrical bin is two, respectively installed in the both ends of running track, inspection car moves along running track.The utility model uses visual identification technical scheme, then carries out water leakage detection by the way of inspection, on one hand, visual identification scheme can have greater area monitoring, secondly, inspection can be completed by only one equipment, compared with fixed equipment, it is more cost-saving, failure rate is lower, based on visual identification technology, combined with edge computing technology, by the combination of modular hardware and intelligent embedded algorithm, the image data collected is transmitted to the edge computing module in inspection car, to carry out real-time offline processing, greatly reduce the storage capacity and transmission capacity of data, reduce the requirement to network software and hardware, realize the accurate monitoring and efficient operation of tunnel leakage water.

Description

Technical Field

[0001] This utility model relates to the field of tunnel leakage detection technology, and in particular to an automatic tunnel leakage detection device integrating visual recognition. Background Technology

[0002] Since tunnels are excavated inside mountains, monitoring for leaks and seepage is crucial. Traditional monitoring methods use sensors such as water immersion sensors and flow meters, but these can only be deployed at points prone to leakage, such as tunnel joints, and cannot provide comprehensive monitoring. With technological advancements, visual recognition technology has become more advantageous, but fixed cameras have limited range, and longer tunnels require more equipment, resulting in higher costs. Utility Model Content

[0003] The technical problem to be solved by this utility model is to provide an automatic detection device for tunnel water leakage that integrates visual recognition, so as to solve the technical problems in the background art mentioned above.

[0004] The technical solution of this utility model is as follows:

[0005] An integrated visual recognition automatic tunnel leakage detection device includes a running track, an inspection vehicle, and an electrical compartment. The running track is suspended along the tunnel's roof. Two electrical compartments are installed at opposite ends of the running track, and the inspection vehicle moves along the track. The inspection vehicle includes a vehicle body, a multispectral camera, an edge computing module, and a communication module A. The multispectral camera is a wide-angle model, single-axis vertically swinging, and is installed back-to-back at the bottom of the vehicle body, facing the tunnel's two side walls. The edge computing module and communication module A are both located inside the vehicle body. The edge computing module is connected to the multispectral camera for real-time image processing. The communication module is connected to the edge computing module for data transmission. The electrical compartment has an opening at the front, and a charger and communication module B are installed on the side walls of the electrical compartment. The charger is compatible with the charging device equipped on the vehicle body, and the communication module B is used to interact with and transmit data to the communication module A.

[0006] Furthermore, the running track has an I-shaped cross-section and a lifting bolt hole at the top; two sets of vertical rollers are installed on the top wall of the vehicle body corresponding to both sides of the running track, and horizontal drive wheels are installed between the vertical rollers in the same set. The horizontal drive wheels on both sides clamp the vertical wall of the running track, and their shafts are connected to the servo motor installed in the vehicle body.

[0007] Furthermore, the vertical hanging wheel is installed on the top of the L-shaped wheel seat, and the bottom of the L-shaped wheel seat is detachably fixed to the vehicle body by bolts. The horizontal drive wheel is detachably fixed to its shaft by nuts, pins or transmission keys.

[0008] Furthermore, the charger is a wireless charger, the charging device is an inductive charging device, and the electrical compartment uses wireless charging to charge the battery inside the vehicle.

[0009] Furthermore, the inspection vehicle is also equipped with an integrated SSD solid-state drive, which connects to the edge computing module to store data and supports offline caching and data resume transmission.

[0010] Furthermore, communication module A uses a Bluetooth transmission module or a LoRa module, while communication module B supports the transmission protocol of communication module A and can be used in a hybrid network with the 5G module.

[0011] Furthermore, the edge computing module uses either NVIDIA Jetson AGX Xavier or Eastcom HOURSIS2025 traffic server.

[0012] Furthermore, distance sensors are installed at the front and rear of the vehicle body, respectively, and are connected to the MCU of the inspection vehicle to provide distance feedback when the inspection vehicle enters the electrical compartment.

[0013] The advantages of this utility model are:

[0014] This invention employs a visual recognition technology solution, followed by inspection for leak detection. Firstly, the visual recognition solution allows for monitoring of a larger area; secondly, inspection only requires a single device, resulting in lower costs and failure rates compared to fixed equipment. Based on visual recognition technology and combined with edge computing technology, the collected image data is transmitted to the edge computing module within the inspection vehicle via modular hardware and intelligent embedded algorithms for real-time offline processing. This significantly reduces data storage and transmission volumes, lowers the demands on network hardware and software, and enables accurate monitoring and efficient maintenance of tunnel leaks. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the structure of this utility model;

[0016] Figure 2 This is a three-dimensional illustration of the present utility model. Figure 1 ;

[0017] Figure 3 This is a three-dimensional illustration of the present utility model. Figure 2 .

[0018] In the diagram: 1-Electrical compartment, 2-Communication module B, 3-Charger, 31-Charging device, 4-Inspection vehicle, 41-Vehicle body, 42-Vertical lifting wheel, 421-L-type wheel seat, 43-Horizontal drive wheel, 44-Servo motor, 45-Distance sensor, 5-Multispectral camera, 6-Running track, 61-Lifting bolt hole. Detailed Implementation

[0019] The specific embodiments of this utility model will be further described below with reference to the accompanying drawings. It should be noted that these descriptions are for the purpose of aiding understanding of this utility model, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of this utility model described below can be combined with each other as long as they do not conflict with each other.

[0020] like Figure 1-3 As shown:

[0021] An integrated visual recognition automatic tunnel leakage detection device includes a running track 6, an inspection vehicle 4, and an electrical compartment 1. The running track 6 is suspended along the top wall of the tunnel. There are two electrical compartments 1, which are respectively installed at both ends of the running track 6. The inspection vehicle 4 moves along the running track 6. The inspection vehicle 4 includes a vehicle body 41, a multispectral camera 5, an edge computing module, and a communication module A. The multispectral camera 5 is a wide-angle model, with a single-axis up-and-down swing, and is installed back-to-back at the bottom of the vehicle body 41, respectively facing the two side walls of the tunnel. The edge computing module and the communication module A are both located inside the vehicle body 41. The edge computing module is connected to the multispectral camera 5 for real-time processing of captured images. The communication module is connected to the edge computing module for data transmission. The electrical compartment 1 has an opening at the front end, and a charger 3 and a communication module B2 are respectively installed on the side wall of the electrical compartment 1. The charger 3 is matched with the charging device 31 equipped on the vehicle body 41, and the communication module B2 is used to interact with the communication module A and transmit data.

[0022] Considering the curvature of the tunnel, the running track 6 is designed with an I-shaped cross section and has a lifting bolt hole 61 at the top to facilitate the operation of the inspection vehicle 4. Two sets of vertical rollers 42 are installed on the top wall of the vehicle body 41 on both sides of the running track 6. Horizontal drive wheels 43 are installed between the vertical rollers 42 in the same set. The horizontal drive wheels 43 on both sides clamp the vertical wall of the running track 6. Their shafts are connected to the servo motors 44 installed in the vehicle body 41. This structure, which is suspended by rollers and driven by horizontal drive wheels 43, is more suitable for operation in curved conditions.

[0023] For ease of installation and maintenance, the vertical wheel 42 is mounted on the top of the L-shaped wheel seat 421. The bottom of the L-shaped wheel seat 421 is detachably fixed to the vehicle body 41 by bolts. The horizontal drive wheel 43 is detachably fixed to its shaft by nuts, pins or transmission keys.

[0024] Considering the usage environment, the high humidity inside the tunnel makes it inconvenient to use methods such as sliding power supply or plug-in charging. Therefore, a wireless charger is used, and the charging device is an inductive charging device. The electrical compartment uses wireless charging to charge the battery inside the vehicle body 41, which can improve the safety of the equipment.

[0025] Because some tunnels are long and winding, making internal signal transmission difficult, the inspection vehicle 4 is also equipped with an integrated SSD solid-state drive (such as a Samsung 870EVO) to connect to the edge computing module for data storage. It supports offline caching and data resume transmission. Each time the inspection vehicle 4 approaches the electrical compartment 1 or is charging in the electrical compartment, it transmits data to the communication module B2. The communication module A uses a Bluetooth or LoRa module. The communication module B2 supports the transmission protocol of the communication module A and can be networked with the 5G module. In this way, the communication module A only performs short-range communication, and then the communication module B2 transmits the data out.

[0026] Each inspection vehicle 4 performs a single-sided inspection as one cycle. Except during inspection time, it stays in the electrical compartment 1 in standby or charging mode. Distance sensors 45 can be installed at the front and rear ends of the vehicle body 41, respectively, and connected to the MCU of the inspection vehicle 4 to provide distance feedback when the inspection vehicle 4 enters the electrical compartment 1.

[0027] This utility model adopts a visual recognition technology solution and then conducts water leakage detection through inspection. On the one hand, the visual recognition solution can monitor a larger area, and on the other hand, the inspection only requires one device, which is more cost-effective than fixed equipment.

[0028] During the inspection, multispectral camera 5 is responsible for collecting image data from both sides of the tunnel. It is a wide-angle model with a single-axis vertical swing (maximum vertical shooting angle up to 180°), sufficient to cover half of the tunnel wall. Multispectral camera 5 uses a combination of visible light and infrared thermal imaging (such as the FLIRA series thermal imager) to cover different lighting conditions. At night, it detects temperature differences in seepage areas using the infrared band, thus eliminating the need for supplementary lighting and not affecting vehicles inside the tunnel. The collected image data is transmitted to the edge computing module inside the inspection vehicle 4 for real-time offline processing. The edge computing unit uses either NVIDIA Jetson AGXXavier or the Eastcom HOURSIS 2025 traffic server, employing Cascade Mask. The R-CNN framework, combined with the lightweight MobileNetV3 backbone network, achieves pixel-level segmentation of leakage areas (IoU threshold 0.6-0.7). Based on annotated leakage datasets (including water stains, crystals, etc.), it adapts to different tunnel structures through transfer learning and online update mechanisms, thereby selecting images suspected of leakage. Only these images need to be transmitted, which can greatly reduce the amount of data storage and transmission. Data processing is performed at the monitoring end, reducing the requirements for network hardware and software.

[0029] In this utility model, for example, the control and drive of the inspection vehicle are all existing technologies. All other parts not mentioned are existing technologies. Those skilled in the art can select and use them according to specific circumstances within the existing technologies. The shape of electrical modules not shown (such as communication module A and edge computing module) is not an improvement point, and they are all existing technologies selected. Based on their understanding of their functions, those skilled in the art can complete the wiring connection themselves.

[0030] The embodiments of this utility model have been described in detail above with reference to the accompanying drawings, but this utility model is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of this utility model, and these variations still fall within the protection scope of this utility model.

Claims

1. An automatic detection device for tunnel water leakage integrating visual recognition, characterized in that: The system includes a running track, an inspection vehicle, and an electrical compartment. The running track is suspended along the tunnel's roof. There are two electrical compartments, one at each end of the running track, and the inspection vehicle moves along the running track. The inspection vehicle includes a vehicle body, a multispectral camera, an edge computing module, and a communication module A. The multispectral camera is a wide-angle model, single-axis vertically swinging, and is mounted back-to-back at the bottom of the vehicle body, facing the two side walls of the tunnel. The edge computing module and communication module A are both located inside the vehicle body. The edge computing module is connected to the multispectral camera for real-time image processing. The communication module is connected to the edge computing module for data transmission. The electrical compartment has an opening at the front, and a charger and communication module B are installed on the side walls of the electrical compartment. The charger is compatible with the charging device equipped on the vehicle body, and the communication module B is used to interact with and transmit data to the communication module A.

2. The automatic tunnel leakage detection device integrating visual recognition according to claim 1, characterized in that: The running track has an I-shaped cross section and a lifting bolt hole at the top. Two sets of vertical rollers are installed on the top wall of the vehicle body corresponding to both sides of the running track. Horizontal drive wheels are installed between the vertical rollers in the same set. The horizontal drive wheels on both sides clamp the vertical wall of the running track, and their shafts are connected to a servo motor installed in the vehicle body.

3. The automatic tunnel leakage detection device with integrated visual recognition according to claim 2, characterized in that: The vertical hanging wheel is installed on the top of the L-shaped wheel seat, and the bottom of the L-shaped wheel seat is detachably fixed to the vehicle body by bolts. The horizontal drive wheel is detachably fixed to its shaft by nuts, pins or transmission keys.

4. The automatic tunnel leakage detection device integrating visual recognition according to claim 1, characterized in that: The charger is a wireless charger, and the charging device is an inductive charging device. The electrical compartment uses wireless charging to charge the battery inside the vehicle.

5. The automatic tunnel leakage detection device integrating visual recognition according to claim 1, characterized in that: The inspection vehicle is also equipped with an integrated SSD solid-state drive, which is connected to the edge computing module to store data and supports offline caching and data resume transmission.

6. The automatic tunnel leakage detection device with integrated visual recognition according to claim 5, characterized in that: The communication module A uses a Bluetooth transmission module or a LoRa module, and the communication module B supports the transmission protocol of the communication module A to form a hybrid network with the 5G module.

7. The automatic tunnel leakage detection device with integrated visual recognition according to claim 1, characterized in that: The edge computing module uses either NVIDIA Jetson AGX Xavier or Eastcom HOURSIS2025 traffic server.

8. The automatic tunnel leakage detection device with integrated visual recognition according to any one of claims 1-7, characterized in that: Distance sensors are installed at the front and rear of the vehicle body, respectively, and are connected to the MCU of the inspection vehicle to provide distance feedback when the inspection vehicle enters the electrical compartment.

Images