Tunnel security detection method, device and readable storage medium

Abstract

The application discloses a tunnel safety detection method and device and a readable storage medium, and belongs to the technical field of unmanned planes. The application is used for controlling an unmanned plane cluster, the unmanned plane cluster comprising a cluster head unmanned plane arranged at a tunnel entrance and at least one cluster member unmanned plane arranged in a tunnel, the cluster head unmanned plane and the cluster member unmanned plane being connected through a chain network, after the cluster member unmanned plane collects detection data, marking the detection data with a node mark and a time mark; acquiring routing information, and sending the detection data to a processing end based on the routing information; when the processing end receives the detection data, determining a tunnel position according to the node mark and the time mark; and determining a safety detection result of the tunnel position according to the detection data and a tunnel detection model. The technical effect that the unmanned plane automatically completes the safety detection of the tunnel without manual participation and improves the tunnel safety detection efficiency is achieved.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) technology, and in particular to tunnel safety inspection methods, equipment, and readable storage media. Background Technology

[0002] For tunnel environments without GNSS (Global Navigation Satellite System) signals, the main inspection methods are currently manual inspection and robot inspection on fixed tracks.

[0003] However, in the context of long and deep tunnels, manual inspections are inefficient, and the reliability of inspection results is often reduced due to subjective human factors. Robotic inspections, on the other hand, are limited by the installation track and cannot effectively monitor blind spots.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this application is to provide a tunnel safety inspection method, apparatus, equipment, and readable storage medium, aiming to solve the technical problem of low efficiency in existing tunnel inspection schemes.

[0006] To achieve the above objectives, this application provides a tunnel safety detection method for controlling a drone swarm. The drone swarm includes a leader drone deployed at the tunnel entrance and at least one member drone deployed inside the tunnel. The leader drone and the member drones are connected via a chain network. The tunnel safety detection method includes the following steps:

[0007] The tunnel safety inspection method includes the following steps:

[0008] After the cluster member UAV collects the detection data, it adds node markers and time markers to the detection data;

[0009] Obtain routing information, and send the detection data to the processing terminal based on the routing information;

[0010] When the processing terminal receives the detection data, it determines the tunnel location based on the node marker and the time marker.

[0011] Based on the detection data and the tunnel detection model, the safety detection results for the tunnel location are determined.

[0012] Optionally, the step of adding node and time markers to the detection data after the cluster member UAV collects the detection data includes:

[0013] When the processing terminal receives a tunnel detection task, it obtains the tunnel parameters corresponding to the tunnel detection task.

[0014] Based on the tunnel detection task and the tunnel parameters, determine the control scheme for the UAV swarm;

[0015] When the cluster leader UAV receives the control scheme and the GNSS signal, it transmits the control scheme and the GNSS signal to the cluster member UAVs through the chain network.

[0016] Optionally, the tunnel parameters include tunnel width, tunnel length, and tunnel route. The step of determining the control scheme for the UAV swarm based on the tunnel detection task and the tunnel parameters includes:

[0017] The number of cluster-head drones is determined based on the tunnel length;

[0018] Based on the tunnel width, tunnel route, and number of cluster head drones, determine the route planning and task allocation for the cluster member drones;

[0019] The control parameters of the cluster-member UAV are determined based on the task level of the tunnel detection task.

[0020] Based on the number of the cluster-leading UAVs, the route planning, the task allocation, and the control parameters, a control scheme for the UAV cluster is determined.

[0021] Optionally, the detection data includes point cloud data inside the tunnel, and the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes:

[0022] The point cloud data is visualized to generate cross-sectional data of the tunnel location;

[0023] Obtain the historical cross-sectional dataset of the tunnel location;

[0024] Based on the cross-sectional data, the historical cross-sectional dataset, and the tunnel detection model, the safety detection result for whether the tunnel exhibits spatial deformation is output.

[0025] Optionally, the tunnel detection model includes a tunnel crack defect identification model, and the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes the following before:

[0026] Acquire a set of images of tunnel cracks;

[0027] Multi-scale crack features are extracted and labeled from the tunnel crack image set, including crack edge shape, gray-level abrupt changes at crack defects, and crack development texture.

[0028] Based on the multi-scale features, the tunnel crack defect identification model is established and trained.

[0029] Optionally, the tunnel detection model includes a tunnel leakage area identification model, and the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes:

[0030] Acquire a set of infrared images of the leaking area in the tunnel;

[0031] Extract and label multi-scale leakage features from the infrared image set, the multi-scale leakage features including temperature at the leakage point, leakage path, and leakage diffusion point;

[0032] Based on the multi-scale leakage characteristics, a model for identifying leakage areas in the tunnel is established and trained.

[0033] Optionally, the tunnel detection model includes a tunnel void identification model, and the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes:

[0034] Acquire a set of images showing tunnel hollowness;

[0035] Extract and label the hollow features of the tunnel hollow image set;

[0036] Based on the hollow features, the tunnel hollow recognition model is established and trained.

[0037] Optionally, the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes:

[0038] When the cluster member UAV receives the safety detection result, it determines the emergency device to be invoked based on the safety detection result. The emergency device includes a waterproof coating spraying device and a warning information marking device.

[0039] The emergency device is invoked to handle the abnormal locations marked in the safety detection results.

[0040] In addition, to achieve the above objectives, this application also provides a tunnel safety detection device, the device comprising: a memory, a processor, and a tunnel safety detection program stored in the memory and executable on the processor, the tunnel safety detection program being configured to implement the steps of the tunnel safety detection method described above.

[0041] In addition, to achieve the above objectives, this application also provides a readable storage medium storing a tunnel security detection program, which, when executed by a processor, implements the steps of the tunnel security detection method described above.

[0042] In this application, to address the inefficiency of existing tunnel inspection schemes, a chain-like network of drones is constructed. Each drone possesses terminal node and routing capabilities, enabling both point-to-point communication and data packet relay. The drone swarm includes a leader drone deployed at the tunnel entrance and at least one member drone deployed within the tunnel. After the member drones collect inspection data, they tagged the data with node and time stamps; obtained routing information; and sent the inspection data to a processing terminal based on the routing information. When the processing terminal receives the inspection data, it determines the tunnel location based on the node and time stamps; and, based on the inspection data and the tunnel inspection model, determines the safety inspection result for the tunnel location. This achieves the technical effect of fully automated tunnel safety inspection by drones without human intervention, significantly improving the efficiency of tunnel safety inspection. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the UAV network inspection system for the tunnel safety inspection method of this application;

[0044] Figure 2 This is a flowchart of the first embodiment of the tunnel safety inspection method of this application;

[0045] Figure 3 This is a flowchart of the second embodiment of the tunnel safety inspection method of this application;

[0046] Figure 4 This is a flowchart of the third embodiment of the tunnel safety inspection method of this application;

[0047] Figure 5 This is a schematic diagram of the structure of a tunnel safety detection device in the hardware operating environment involved in the embodiments of this application.

[0048] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0049] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0050] For tunnel environments without GNSS (Global Navigation Satellite System) signals, the main inspection methods are currently manual inspection and robot inspection on fixed tracks.

[0051] However, in the context of long and deep tunnels, manual inspections are inefficient, and the reliability of inspection results is often reduced due to subjective human factors. Robotic inspections, on the other hand, are limited by the installation track and cannot effectively monitor blind spots.

[0052] To address the aforementioned issues, this application, based on the characteristics of Flying Ad-Hoc networks (FANETs), designs a UAV network inspection system for tunnel environments, referring to... Figure 1 , Figure 1 This is a schematic diagram of the UAV network inspection system for tunnel safety inspection according to this application. The system includes a satellite, UAVs at the tunnel entrance, inspection UAVs inside the tunnel, and ground terminal equipment. The UAVs outside the tunnel and the inspection UAVs inside the tunnel are connected in series to form a chain-like network. The UAVs at the tunnel entrance act as cluster heads, and the inspection UAVs inside the tunnel act as cluster members. The cluster head UAVs transmit control schemes and GNSS signals to the cluster member UAVs deep inside the tunnel via multi-hop routing, achieving full signal coverage inside the tunnel. After receiving the control schemes and GNSS signals, the cluster member UAVs execute the control schemes to perform tunnel safety inspections. After the cluster member UAVs collect inspection data, they mark the inspection data with node tags and time stamps; obtain routing information, and send the inspection data to the processing terminal based on the routing information; when the processing terminal receives the inspection data, it determines the tunnel location based on the node tags and time stamps; and determines the safety inspection result of the tunnel location based on the inspection data and the tunnel inspection model. This achieves the technical effect of fully automatic UAV safety inspection of tunnels without human intervention, improving the efficiency of tunnel safety inspection.

[0053] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0054] This application provides a tunnel safety detection method, referring to... Figure 2 , Figure 2 This is a flowchart of the first embodiment of a tunnel safety detection method according to this application.

[0055] In this embodiment, the tunnel safety detection method is used to control a drone swarm. The drone swarm includes a cluster leader drone deployed at the tunnel entrance and at least one cluster member drone deployed inside the tunnel. The cluster leader drone and the cluster member drone are connected by a chain network. The tunnel safety detection method includes the following steps:

[0056] Step S10: After the cluster member UAV collects the detection data, the detection data is marked with node markers and time markers;

[0057] In this embodiment, by equipping the cluster-member UAV with payloads such as a 5G wireless communication transceiver, a visible light camera, an infrared camera, and a lidar, the cluster-member UAV can collect detection data during the inspection process. This monitoring data includes visible light image data, infrared image data, point cloud data, etc. After collecting the detection data, the cluster-member UAV packages the detection data and adds node markers and collection time markers to obtain a detection data package.

[0058] Step S20: Obtain routing information, and send the detection data to the processing terminal based on the routing information;

[0059] Optionally, the cluster member drone sends the aforementioned detection data packet to neighboring cluster member drones based on the obtained routing information, and transmits it hop-by-hop to the processing end.

[0060] Optionally, a mobile edge computing (MEC) server can be installed on the cluster-head drone, with the cluster-head drone acting as the processing end to process the detection data.

[0061] Optionally, the cluster-head UAV connects to the ground terminal device through the aforementioned mobile edge computing server, transmitting the aforementioned detection data to the ground terminal device, which then acts as the processing end, providing remote wireless access and low-latency edge computing services.

[0062] Furthermore, satellites can be used to provide efficient and fast cloud computing services for ground terminal equipment.

[0063] Step S30: When the processing terminal receives the detection data, it determines the tunnel location based on the node marker and the time marker;

[0064] Step S40: Determine the safety detection result of the tunnel location based on the detection data and the tunnel detection model.

[0065] Optionally, when the processing end receives the detection data packet, it first determines the drone that collected the detection data based on the node marker of the detection data packet, and determines the collection time of the detection data based on the timestamp. Then, based on the location of the drone at the determined collection time, it determines the location of the tunnel being detected. Finally, based on the content of the detection data packet and the tunnel detection model, it determines the safety detection result of the tunnel location.

[0066] Optionally, the tunnel inspection model includes a tunnel crack defect identification model, a tunnel leakage area identification model, a tunnel hollow area identification model, and a tunnel surface spalling identification model, etc. Different types of inspection data subsets are extracted based on the inspection data, and these subsets are input into the corresponding tunnel inspection models.

[0067] In one implementation, the detection data includes a subset of infrared image data, which is input into the tunnel leakage area detection model.

[0068] As another implementation, the detection data also includes a subset of visible light image data, which is input into the tunnel crack defect identification model, the tunnel hollow identification model, and the tunnel surface spalling identification model.

[0069] In actual detection, the required detection model can be pre-trained using deep learning algorithms according to actual needs, including but not limited to the recognition models enumerated in this embodiment.

[0070] Optionally, the tunnel detection model includes a tunnel crack defect identification model, and the steps prior to step S40 include:

[0071] Step S31: Obtain a set of tunnel crack images;

[0072] Step S32: Extract and label the multi-scale crack features of the tunnel crack image set, including crack edge shape, gray-scale abrupt changes at crack defects, and crack development texture;

[0073] Step S33: Based on the multi-scale features, establish and train the tunnel crack defect identification model.

[0074] Optionally, a set of tunnel crack images (visible light images) is acquired, and multi-scale crack features of the tunnel crack image set are extracted and labeled. The multi-scale crack features include crack edge shape, gray-scale abrupt changes at crack defects, and crack development texture. Based on the multi-scale features, a dataset of various cracks and other defects under different tunnel environments is constructed and labeled. A binary image segmentation model is constructed using deep learning frameworks such as U-Net to learn the pixel features of crack defects in the images, and the tunnel crack defect recognition model is established and trained.

[0075] Optionally, the tunnel detection model includes a tunnel leakage area identification model, and the steps prior to step S40 include:

[0076] Step S34: Obtain an infrared image set of the leaking area in the tunnel;

[0077] Step S35: Extract and label the multi-scale leakage features of the infrared image set, the multi-scale leakage features including the temperature at the leakage point, the leakage path, and the leakage diffusion point;

[0078] Step S36: Based on the multi-scale leakage features, establish and train the tunnel leakage area identification model.

[0079] Optionally, an infrared image set (infrared images) of the tunnel leakage area is acquired, and multi-scale leakage features of the infrared image set are extracted and labeled. The multi-scale leakage features include the temperature at the leakage point, the leakage path, and the leakage diffusion point. Based on the multi-scale features, various leakage area datasets under different tunnel environments are constructed and labeled. A binary image segmentation model is constructed using deep learning frameworks such as U-Net to learn the pixel features of the leakage area in the image, and the tunnel leakage area recognition model is established and trained.

[0080] Optionally, the tunnel detection model includes a tunnel void recognition model, and the steps prior to step S40 include:

[0081] Step S37: Obtain a set of tunnel hollow images;

[0082] Step S38: Extract and label the hollow features of the tunnel hollow image set;

[0083] Step S39: Based on the hollow features, establish and train the tunnel hollow recognition model.

[0084] Optionally, a set of tunnel hollow images (visible light images) is acquired, and multi-scale hollow features of the tunnel hollow image set are extracted and labeled. The multi-scale hollow features include the shape of the hollow edge, gray-scale abrupt changes at the hollow protrusion, and texture changes in the hollow region. Based on the multi-scale features, various hollow region datasets under different tunnel environments are constructed and labeled. A binary image segmentation model is constructed using deep learning frameworks such as U-Net to learn the pixel features of the hollow region in the image, and the tunnel hollow recognition model is established and trained.

[0085] Furthermore, the detection data includes point cloud data inside the tunnel, and step S40 includes:

[0086] Step S41: Visualize the point cloud data to generate cross-sectional data of the tunnel location;

[0087] Step S42: Obtain the historical cross-sectional dataset of the tunnel location;

[0088] Step S43: Based on the cross-sectional data, the historical cross-sectional dataset, and the tunnel detection model, output the safety detection result of whether the tunnel has spatial deformation.

[0089] Optionally, the cluster-based UAV is equipped with a lidar to perform planar three-dimensional laser scanning of the tunnel interior. Through site stitching, calibration, and adjustment, high-precision and high-density point cloud data is obtained. Data visualization processing technology is used to generate tunnel cross-sectional data, and tunnel deformation analysis is performed by comparing data from multiple periods, thereby enabling real-time, accurate, and comprehensive detection of tunnel spatial deformation.

[0090] Alternatively, due to limitations imposed by environmental and other factors, the point cloud acquired by lidar in a single acquisition process can only cover a portion of the tunnel surface. In order to obtain complete tunnel point cloud information, it is necessary to scan the target object multiple times and perform a rigid body transformation on the coordinate system of the obtained three-dimensional point cloud data to convert the local point cloud data of the tunnel surface to the same coordinate system.

[0091] Optionally, the point cloud registration process described above can be divided into two steps: coarse registration and fine registration. Coarse registration, or initial registration of the point clouds, refers to aligning two point clouds at different positions as closely as possible using an initial value for a rotation and translation matrix. Mainstream methods for coarse registration include RANSAC and 4PCS. After coarse registration, the overlapping portions of the two point clouds can be roughly aligned, but the accuracy is still insufficient to detect slight deformations in the tunnel, requiring further fine registration. Fine registration involves calculating approximate rotation and translation matrices between the two point clouds based on the initial registration. Mainstream methods for fine registration include ICP, NDT, and deep learning.

[0092] In this embodiment, drones outside the tunnel and inspection drones inside the tunnel are connected in series to form a chain-like network. The drone at the tunnel entrance acts as the cluster head, and the inspection drones inside the tunnel act as cluster members. The cluster head drone transmits the control scheme and GNSS signals to the cluster member drones deep inside the tunnel via multi-hop routing, achieving full signal coverage within the tunnel. After receiving the control scheme and GNSS signals, the cluster member drones execute the control scheme to perform tunnel safety inspections. After the cluster member drones collect the inspection data, they mark the data with node tags and time stamps; obtain routing information, and send the inspection data to the processing terminal based on the routing information; when the processing terminal receives the inspection data, it determines the tunnel location based on the node tags and time stamps; and based on the inspection data and the tunnel inspection model, it determines the safety inspection result for the tunnel location. This achieves the technical effect of fully automated tunnel safety inspection by drones without human intervention, improving the efficiency of tunnel safety inspection.

[0093] Furthermore, referring to Figure 3 , Figure 3 This is a flowchart illustrating the second embodiment of the tunnel safety detection method of this application. The second embodiment of this application includes the following steps before step S10:

[0094] Step S00: When the processing terminal receives a tunnel detection task, it obtains the tunnel parameters corresponding to the tunnel detection task;

[0095] Step S01: Determine the control scheme for the UAV swarm based on the tunnel detection task and the tunnel parameters;

[0096] Step S02: When the cluster head UAV receives the control scheme and the GNSS signal, it transmits the control scheme and the GNSS signal to the cluster member UAVs through the chain network.

[0097] In this embodiment, the processing end automatically generates a control scheme for the UAV based on the received tunnel detection task, and transmits the control scheme and GNSS signal to each member UAV through a chain network, so that the member UAV can have GNSS signal for take-off, navigation and positioning, and perform detection tasks according to the control scheme.

[0098] Optionally, the tunnel inspection task includes the tunnel to be inspected and the task level, which includes depth inspection tasks and routine inspection tasks. Compared with routine inspection tasks, depth inspection tasks involve controlling the drone swarm to stay inside the tunnel for a longer period of time, resulting in richer and more comprehensive inspection data and higher inspection accuracy.

[0099] Optionally, based on the tunnel to be detected, corresponding tunnel parameters are obtained, including tunnel width, tunnel length, and tunnel route, etc.

[0100] Optionally, the number of cluster-head drones is determined based on the tunnel length. As one implementation, it is determined whether the tunnel length is greater than a preset length. If so, two cluster-head drones (i.e., two drone clusters) are set up, entering from the entrances at both ends of the tunnel and simultaneously performing tasks. If not, one cluster-head drone (i.e., one drone cluster) is set up, entering from the selected tunnel entrance to perform the detection task.

[0101] Optionally, the distribution of the cluster-member UAVs is determined based on the tunnel length. As one implementation, it is determined whether the tunnel width is greater than a preset width. If so, the same tunnel segment detection task is assigned to two cluster-member UAVs. Depending on actual needs, both sides can be inspected, or each UAV can inspect half of the tunnel segment. If not, one cluster-member UAV is assigned to each tunnel segment to perform the detection task.

[0102] As another implementation, the tunnel parameters also include the number of lanes, and the distribution of the cluster drones is determined based on the number of lanes, that is, the number of cluster drones allocated to the same tunnel segment.

[0103] Optionally, the hardware information of the cluster-member UAV is obtained, including at least one of battery level, service life, maintenance history, payload equipment, and flight time. Based on the hardware information, the tunnel section to be inspected by the cluster-member UAV is determined.

[0104] In one implementation approach, the cluster-member UAV with the most remaining battery power, shortest service life, fewest maintenance history, highest-level load equipment, or highest overall score is selected to perform the inspection task in the deepest tunnel section. Conversely, the cluster-member UAV with the least remaining battery power, longest service life, most maintenance history, lowest-level load equipment, or lowest overall score is selected to perform the inspection task in the tunnel section closest to the tunnel entrance.

[0105] As another implementation, when determining the tunnel section to be inspected by the cluster-member UAV based on hardware information from two or more items, the weight of the corresponding item for the hardware information is obtained, and a comprehensive score is determined based on the hardware information and the weight. The comprehensive score of the hardware information of the cluster-member UAV is proportional to the depth of the tunnel section inspected by the cluster-member UAV.

[0106] Optionally, the route planning and task allocation of the cluster member drones are determined based on the tunnel width, tunnel route, number of cluster head drones, and hardware information of the cluster member drones.

[0107] Optionally, control parameters for the UAV are determined based on the task level. These control parameters include the UAV's flight speed when performing inspection tasks and its hovering time when collecting detection data.

[0108] Optionally, the control scheme for the drone swarm is determined based on the number of the cluster leader drones, the route planning, the task allocation, and the control parameters.

[0109] In this embodiment, the processing unit can automatically acquire relevant data information, such as tunnel parameters and hardware information of the cluster drones, based on the received tunnel detection task. Based on this data, it can automatically formulate a control scheme for the drone swarm. The entire process requires no manual intervention, effectively improving the efficiency of tunnel detection.

[0110] Furthermore, referring to Figure 4 , Figure 4 This is a flowchart illustrating the third embodiment of the tunnel safety detection method of this application. The third embodiment of this application includes the following steps after step S40:

[0111] Step S41: When the cluster member UAV receives the safety detection result, determine the emergency device to be called based on the safety detection result. The emergency device includes a waterproof coating spraying device and a warning information marking device.

[0112] Step S42: Invoke the emergency device to perform emergency handling on the abnormal locations marked in the safety detection results.

[0113] Optionally, the cluster-member UAV determines, based on the safety detection results, that there is an anomaly in the tunnel section where it performs the detection task, and determines the emergency device to be invoked based on the anomaly.

[0114] As one implementation method, if the anomaly is water leakage, a waterproof coating spraying device is activated to spray the leaking area marked in the safety inspection results as an emergency measure.

[0115] As another implementation, the anomaly is the detachment of the tunnel surface. In order to prevent the surface from continuing to detach and potentially hitting passing vehicles or pedestrians, a warning information marking device is activated to warn of the detachment location marked in the safety inspection results as an emergency measure.

[0116] Optionally, the aforementioned warning information marking device can be a fluorescent paint spraying device, or a warning flag that can be adsorbed onto the tunnel surface, etc. Any marking device that can serve a warning purpose is acceptable, and this embodiment does not impose any restrictions on it.

[0117] Furthermore, based on the safety detection results, an anomaly level is determined. If the anomaly level is a preset high-risk level, the cluster drone will issue an alarm to evacuate vehicles and pedestrians.

[0118] Optionally, the cluster-member UAV also has terminal functionality. Besides transmitting detection data packets, its own nodes perform real-time analysis of these packets, enabling it to quickly determine the safety detection sub-results for the detected tunnel segment. Based on these safety detection sub-results, the cluster-member UAV executes the aforementioned emergency procedures.

[0119] In this embodiment, by equipping the drone with an emergency device, abnormalities in the safety inspection results can be handled without waiting for manual repairs, effectively improving the efficiency of tunnel safety inspections. Alternatively, if the safety inspection results show serious anomalies that necessitate manual repairs, the timely emergency handling by the drone can effectively reduce the risks associated with anomalies occurring before manual repairs.

[0120] In addition, this application also proposes a tunnel safety detection device.

[0121] Reference Figure 5 , Figure 5 This is a schematic diagram of the tunnel safety detection equipment structure in the hardware operating environment involved in the embodiments of this application.

[0122] like Figure 5 As shown, the tunnel safety detection device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0123] Those skilled in the art will understand that Figure 5The structure shown does not constitute a limitation on tunnel safety inspection equipment and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0124] like Figure 5 As shown, the memory 1005, which serves as a readable storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and a tunnel security detection program.

[0125] exist Figure 5 In the tunnel safety detection equipment shown, the network interface 1004 is mainly used for data communication with other devices, including satellites, cluster-head UAVs, and other ground terminal equipment; the user interface 1003 is mainly used for data interaction with users; the processor 1001 and memory 1005 in the tunnel safety detection equipment of this application can be set in the tunnel safety detection equipment. The tunnel safety detection equipment calls the tunnel safety detection program stored in the memory 1005 through the processor 1001 and executes the tunnel safety detection method provided in the embodiment of this application. The specific embodiments are basically the same as the above-mentioned tunnel safety detection method embodiments, and will not be described in detail here.

[0126] Furthermore, embodiments of this application also propose a readable storage medium.

[0127] This application has a readable storage medium storing a tunnel security detection program, which, when executed by a processor, implements the steps of the tunnel security detection method described above.

[0128] The specific embodiment in which the tunnel security detection program stored in the readable storage medium of this application is executed by the processor is basically the same as the embodiments of the tunnel security detection method described above, and will not be repeated here.

[0129] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitations, an element defined by the phrase "comprising a tunnel security detection" does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0130] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0132] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A tunnel safety inspection method, characterized in that, The tunnel safety detection method is used to control a drone swarm, which includes a leader drone deployed at the tunnel entrance and at least one member drone deployed inside the tunnel. The leader drone and the member drones are connected by a chain network. The tunnel safety detection method includes the following steps: When the processing end receives a tunnel detection task, it obtains the tunnel parameters corresponding to the tunnel detection task, including the tunnel width, tunnel length and tunnel route; The number of cluster-head drones is determined based on the tunnel length; Obtain the hardware information of the cluster member drone, which includes at least one of the following: battery level, service life, maintenance history, payload equipment, and flight time; Based on the tunnel width, tunnel route, number of cluster head drones, and hardware information of the cluster member drones, the route planning and task allocation of the cluster member drones are determined. The control parameters of the cluster-member UAV are determined based on the task level of the tunnel detection task. Based on the number of the cluster-leading UAVs, the route planning, the task allocation, and the control parameters, a control scheme for the UAV cluster is determined. When the cluster leader UAV receives the control scheme and the GNSS signal, it transmits the control scheme and the GNSS signal to the cluster member UAVs through the chain network; After the cluster member UAV collects the detection data, it adds node markers and time markers to the detection data; Obtain routing information, and send the detection data to the processing terminal based on the routing information; When the processing terminal receives the detection data, it determines the tunnel location based on the node marker and the time marker. Based on the detection data and the tunnel detection model, the safety detection results for the tunnel location are determined; The detection data includes visible light image data, infrared image data, and point cloud data. The tunnel detection model includes a tunnel crack defect identification model, a tunnel leakage area identification model, a tunnel hollow area identification model, and a tunnel surface spalling identification model. The step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes: The subset of infrared image data is input into the tunnel leakage area detection model; The subset of visible light image data is input into the tunnel crack defect identification model, the tunnel hollowing identification model, and the tunnel surface spalling identification model; Input the point cloud data into the tunnel detection model; Each tunnel detection model outputs the corresponding safety detection results.

2. The tunnel safety inspection method as described in claim 1, characterized in that, The detection data includes point cloud data inside the tunnel. The step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes: The point cloud data is visualized to generate cross-sectional data of the tunnel location; Obtain the historical cross-sectional dataset of the tunnel location; Based on the cross-sectional data, the historical cross-sectional dataset, and the tunnel detection model, the safety detection result for whether the tunnel exhibits spatial deformation is output.

3. The tunnel safety inspection method as described in claim 1, characterized in that, The tunnel detection model includes a tunnel crack defect identification model. Prior to the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model, the following steps are included: Acquire a set of images of tunnel cracks; Multi-scale crack features are extracted and labeled from the tunnel crack image set, including crack edge shape, gray-level abrupt changes at crack defects, and crack development texture. Based on the multi-scale features, the tunnel crack defect identification model is established and trained.

4. The tunnel safety inspection method as described in claim 1, characterized in that, The tunnel detection model includes a tunnel leakage area identification model. Prior to the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model, the following steps are included: Acquire a set of infrared images of the leaking area in the tunnel; Extract and label multi-scale leakage features from the infrared image set, the multi-scale leakage features including temperature at the leakage point, leakage path, and leakage diffusion point; Based on the multi-scale leakage characteristics, a model for identifying leakage areas in the tunnel is established and trained.

5. The tunnel safety inspection method as described in claim 1, characterized in that, The tunnel detection model includes a tunnel void recognition model. Prior to the step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model, the following steps are included: Acquire a set of images showing tunnel hollowness; Extract and label the hollow features of the tunnel hollow image set; Based on the hollow features, the tunnel hollow recognition model is established and trained.

6. The tunnel safety inspection method as described in claim 1, characterized in that, The step of determining the safety detection result of the tunnel location based on the detection data and the tunnel detection model includes: When the cluster member UAV receives the safety detection result, it determines the emergency device to be invoked based on the safety detection result. The emergency device includes a waterproof coating spraying device and a warning information marking device. The emergency device is invoked to handle the abnormal locations marked in the safety detection results.

7. A tunnel safety inspection device, characterized in that, The device includes: a memory, a processor, and a tunnel security detection program stored in the memory and executable on the processor, the tunnel security detection program being configured to implement the steps of the tunnel security detection method as described in any one of claims 1 to 6.

8. A readable storage medium, characterized in that, The readable storage medium stores a tunnel safety detection program, which, when executed by a processor, implements the steps of the tunnel safety detection method as described in any one of claims 1 to 6.

Images