Tunnel monitoring unmanned aerial vehicle, unmanned aerial vehicle base station and wall surface monitoring method

Abstract

The invention relates to the field of unmanned aerial vehicles and tunnel automatic monitoring, and particularly provides a tunnel monitoring unmanned aerial vehicle, an unmanned aerial vehicle base station and a wall surface monitoring method. The unmanned aerial vehicle comprises an unmanned aerial vehicle main body, a rotating component, a rotor wing, an image acquisition device, a magnetic attraction guiding charging structure and a three-dimensional laser radar, and the base station comprises a box body, a bin door, a magnetic attraction seat, an annular two-dimensional code and a separation mechanism. The method comprises the following steps: after the unmanned aerial vehicle is separated from the base station, fusing multi-sensor positioning, flying according to a preset route to collect wall data, docking with the base station through magnetic attraction after completing a task, and analyzing data by the remote control center to generate a health report. The problems that the unmanned aerial vehicle in the tunnel is difficult to position and unreliable in charging and needs manual operation are solved, automatic monitoring is achieved, efficiency and reliability are improved, and tunnel safety is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of unmanned aerial vehicles and automatic tunnel monitoring, in particular to a tunnel monitoring unmanned aerial vehicle, an unmanned aerial vehicle base station and a wall surface monitoring method. BACKGROUND

[0002] At present, under long-term operation, cracks, leakage and other diseases may occur on the tunnel wall surface, and in order to ensure the brightness of the tunnel, reflective materials or coatings are usually used, and dirt and damage may occur after long-term operation. In the prior art, tunnel wall surface state monitoring mainly relies on manual or track-type robots, which has the defects of low efficiency and limited coverage. Although using unmanned aerial vehicles for monitoring can partially solve the above problems, there are still many challenges in the closed tunnel environment.

[0003] Firstly, the traditional unmanned aerial vehicle relies on GPS satellite signals for positioning, and the signal in the tunnel is weak, so the position of the unmanned aerial vehicle cannot be effectively obtained. Secondly, the unmanned aerial vehicle needs to be manually set to start and end points, and cannot complete the monitoring task remotely and automatically, so it cannot be operated unmanned. Thirdly, in a long-distance tunnel, the battery of the unmanned aerial vehicle may not be able to complete the monitoring task completely, and needs to be frequently returned for charging, and the reliability of the traditional plug-in interface in the dusty environment of the tunnel is difficult to guarantee. SUMMARY

[0004] The present application relates to the field of unmanned aerial vehicles and automatic tunnel monitoring, in particular to a tunnel monitoring unmanned aerial vehicle, an unmanned aerial vehicle base station and a wall surface monitoring method.

[0005] In a first aspect, the present application provides a tunnel monitoring unmanned aerial vehicle, which comprises an unmanned aerial vehicle body, a rotating member, a rotor, an image acquisition device, a magnetic attraction guiding and charging structure and a three-dimensional laser radar. The unmanned aerial vehicle body is provided with a cantilever around the periphery. The cantilever is fixedly connected to the side wall of the unmanned aerial vehicle body, and the cantilevers are uniformly distributed along the circumference of the unmanned aerial vehicle body. The rotating member is arranged on the side wall of the unmanned aerial vehicle body. The rotor is provided with four rotors, which are installed at the end of the cantilever. The image acquisition device is arranged on the rotating member, and the image acquisition device is rotatably connected to the rotating member. The magnetic attraction guiding and charging structure is arranged at the top of the unmanned aerial vehicle body. The three-dimensional laser radar is arranged at the bottom of the unmanned aerial vehicle body.

[0006] In a second aspect, the application provides a tunnel monitoring unmanned aerial vehicle base station, comprising a box body, a warehouse door, a magnetic seat, an annular two-dimensional code, and a separation mechanism, wherein the box body is fixedly arranged at the top of the tunnel and outside the tunnel construction limit; the warehouse door is arranged at the bottom of the box body; the magnetic seat is arranged inside the box body, the magnetic seat is provided with a groove, the groove is matched with the table body structure at the top of the unmanned aerial vehicle, the magnetic seat is provided with a second annular permanent magnet array, a second guide magnet, and a spring thimble, the second annular permanent magnet array is arranged at the bottom of the groove, the spring thimble is arranged in the middle of the second annular permanent magnet array, and the second guide magnet is uniformly distributed along the side surface of the groove; the annular two-dimensional code is arranged on one side of the magnetic seat; and the separation mechanism comprises at least four electromagnetic push rods, and the electromagnetic push rods are arranged around the groove.

[0007] In a third aspect, the application provides a tunnel wall monitoring method, which comprises the following steps:

[0008] obtaining an operation instruction, wherein the operation instruction is used to instruct the unmanned aerial vehicle to perform wall monitoring operation;

[0009] obtaining initialized positioning information according to the operation instruction, and determining a take-off pose according to the initialized positioning information;

[0010] obtaining sensor data, wherein the sensor data comprises point cloud data, IMU data, and UWB data;

[0011] adjusting the take-off pose and position of the unmanned aerial vehicle in real time according to the sensor data, and flying along a preset flight route;

[0012] when the unmanned aerial vehicle is flying, an image acquisition device always faces the tunnel wall to obtain image information;

[0013] when the unmanned aerial vehicle completes the monitoring operation, the image acquisition device rotates upward, recognizes the annular two-dimensional code of the base station, adjusts the attitude to fly upward, and is adsorbed on the magnetic seat through the magnetic attraction and guidance charging structure;

[0014] uploading the image information to a remote control center to generate a tunnel wall health report.

[0015] The application has the following beneficial effects:

[0016] This invention utilizes a tunnel monitoring drone equipped with a high-definition camera, 3D LiDAR, and other equipment. Combined with the drone's base station's magnetic guidance charging and data transmission interface, it employs a multi-sensor fusion positioning and magnetic-visual collaborative docking method to achieve tunnel wall monitoring. This not only solves the problems of difficult drone positioning, low charging reliability, and the need for manual operation within tunnels, achieving automatic monitoring, precise positioning, and efficient data acquisition and transmission to ensure tunnel safety, but also unexpectedly enables stable charging and data transmission in dusty environments through the magnetic guidance structure. Furthermore, the multi-sensor fusion effectively suppresses the errors of individual sensors, significantly improving monitoring accuracy and endurance, reducing redundant scanning, and providing a stable and reliable solution for long-distance tunnel monitoring.

[0017] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the tunnel monitoring drone structure described in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the magnetic attraction-guided charging structure described in an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of the tunnel monitoring drone base station structure described in an embodiment of the present invention.

[0022] Figure 4 This is a schematic diagram of the magnetic base structure described in an embodiment of the present invention.

[0023] Figure 5 This is a schematic diagram of the ring-shaped QR code structure described in an embodiment of the present invention.

[0024] Figure 6 The diagram shows the magnetic orientation of the magnetic charging structure and the magnetic base.

[0025] Marked in the figure: 101, unmanned aerial vehicle body; 102, rotating member; 103, rotor; 104, three-dimensional laser radar; 105, magnetic attraction guiding charging structure; 121, high-definition camera; 122, light supplement lamp; 151, first annular permanent magnet array; 152, charging and data transmission contact; 153, first guiding magnet; 201, box body; 202, door; 203, magnetic attraction seat; 204, electromagnetic push rod; 205, annular two-dimensional code; 231, second annular permanent magnet array; 232, second guiding magnet; 233, spring needle. DETAILED DESCRIPTION

[0026] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work, fall within the scope of protection of the present application.

[0027] It should be noted that: similar labels and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0028] Embodiment 1

[0029] As Figure 1As shown, the embodiment provides a tunnel monitoring unmanned aerial vehicle, which comprises an unmanned aerial vehicle body 101, a rotating member 102, a rotor 103, an image acquisition device, a magnetic attraction guiding and charging structure 105, and a three-dimensional laser radar 104. The unmanned aerial vehicle body 101 is provided with a cantilever around. The cantilever is fixedly connected with the side wall of the unmanned aerial vehicle body 101, and is uniformly distributed along the circumference of the unmanned aerial vehicle body 101. The rotating member 102 is arranged on the side wall of the unmanned aerial vehicle body 101. The rotor 103 is provided with four rotors 103, which are installed at the end of the cantilever. The image acquisition device is arranged on the rotating member 102, and the image acquisition device is rotatably connected with the rotating member 102. The magnetic attraction guiding and charging structure 105 is arranged at the top of the unmanned aerial vehicle body 101. The three-dimensional laser radar 104 is arranged at the bottom of the unmanned aerial vehicle body 101. The present application guarantees the stable flight of the unmanned aerial vehicle in the tunnel through the uniformly arranged cantilever and the rotor 103. The rotating member 102 is arranged to make the image acquisition device flexibly align with the wall surface. The magnetic attraction guiding and charging structure 105 solves the problems of automatic charging and docking of the unmanned aerial vehicle, solves the problems of low efficiency of traditional manual or track monitoring, difficult positioning of the unmanned aerial vehicle in the tunnel, unreliable charging and the like, and realizes automatic and accurate wall surface data acquisition.

[0030] In one specific embodiment of the present disclosure, a motor fixing seat is fixedly arranged at the end of the cantilever, an electric motor is arranged in the motor fixing seat, a power output shaft of the electric motor extends in the vertical direction, and the rotor 103 is horizontally arranged at the top end of the power output shaft. The tunnel is a closed space, there is a building limit restriction (the unmanned aerial vehicle needs to avoid invading the road surface to affect traffic), and the unmanned aerial vehicle needs to maintain a fixed distance from the wall surface to ensure image acquisition accuracy. The vertically extending power output shaft cooperates with the horizontally arranged rotor to directly drive the rotor to generate vertical lift through the electric motor, reduces power transmission loss, and at the same time facilitates accurate control of the lift size by adjusting the rotor speed, meets the needs of stable hovering and translation of the unmanned aerial vehicle in the preset flight path, provides a stable flight basis for wall surface monitoring, and maximizes the utilization of the radial space of the unmanned aerial vehicle through the layout of the horizontal rotor and the vertical power shaft, avoids collision between the rotor and the tunnel wall surface and the top structure, adapts to the flight demand of the narrow space of the tunnel, directly drives the rotor through the vertical shaft, reduces the energy loss and delay of the transmission mechanism, makes the lift adjustment faster and more linear, facilitates accurate control of the attitude of the unmanned aerial vehicle in the scenes of take-off, hovering and docking, and at the same time, due to the airflow disturbance caused by the vehicle driving in the tunnel, the rigid connection structure of the horizontal rotor and the vertical power shaft makes the rotor force more stable, is less likely to cause attitude deviation due to airflow fluctuation compared with the flexible transmission structure, indirectly improves the clarity and positioning accuracy of image acquisition, and in addition, the reduction of power transmission loss reduces the invalid energy consumption, indirectly prolongs the monitoring distance of a single flight in a long-distance tunnel monitoring, reduces the frequency of returning to the base station for charging, and improves the detection efficiency.

[0031] In one specific embodiment of the present disclosure, the image acquisition device comprises a high-definition camera 121 and a fill light 122, the fill light 122 is arranged along the circumference of the high-definition camera 121, and the illumination direction of the fill light 122 is the same as the shooting direction of the high-definition camera 121. In the tunnel, due to the problems of insufficient light and uneven light, it is necessary to ensure that the high-definition camera 121 can still clearly collect the wall details in a low-light environment. Therefore, the fill light 122 is arranged around the high-definition camera 121 and synchronously irradiated, which can reduce the shooting shadow and reflection, uniformly supplement the light, improve the image contrast and detail recognition, cooperate with the rotating member 102 to make the fill light 122 and the high-definition camera 121 always adapt to the wall direction, and provide reliable visual data support for automatic monitoring. In addition, when the tunnel wall is a reflective material, the circumferential distribution of the fill light 122 can effectively weaken the strong light reflection interference and avoid local overexposure caused by a single light source. The synchronicity of the fill light and the shooting can improve the consistency of the image brightness of different tunnel sections (such as the strong light area at the entrance and the dark light area in the deep part), facilitate the remote control center to compare and analyze the trend of crack growth and pollution diffusion, and indirectly improve the accuracy of the health report.

[0032] In one specific embodiment of the present disclosure, the unmanned aerial vehicle body 101 is internally provided with a central control device, a UWB tag and an INS inertial navigation system. The UWB tag realizes distance positioning by communicating with the base station attached positioning anchor points. The INS inertial navigation system provides continuous attitude and position information through an accelerometer and a gyroscope. The central control device, as the core, integrates the data of UWB and INS to correct the positioning error in real time, provides accurate parameters for the flight path control and attitude adjustment of the unmanned aerial vehicle, solves the positioning problem of the unmanned aerial vehicle in the tunnel without GPS signal, and ensures that the unmanned aerial vehicle can determine its position and attitude in real time and fly stably along the preset flight path in the closed and complex tunnel environment.

[0033] As shown in Figure 2 In one specific embodiment of the present disclosure, the magnetic attraction guiding and charging structure 105 is a table body structure, which comprises a first annular permanent magnet array 151, a charging and data transmission contact 152 and a first guiding magnet 153. The first annular permanent magnet array 151 is arranged at the top of the table body structure, the charging and data transmission contact 152 is arranged in the middle of the first annular permanent magnet array 151, and the first guiding magnet 153 is arranged on the side of the table body structure. The first guiding magnet 153 is uniformly distributed along the circumference of the table body structure, the magnetic pole directions of adjacent two first guiding magnets 153 are opposite, and the magnetic force arrangement directions of adjacent two magnets in the first annular permanent magnet array 151 are opposite.

[0034] Example 2

[0035] AsFigure 3 、 Figure 4 and Figure 5 As shown in the drawings, the embodiment provides a tunnel monitoring unmanned aerial vehicle base station, which comprises a box body 201, a warehouse door 202, a magnetic seat 203, a ring-shaped two-dimensional code 205 and a detachment mechanism. The box body 201 is fixedly arranged at the top of the tunnel and located outside the tunnel construction limit. The warehouse door 202 is arranged at the bottom of the box body 201. The magnetic seat 203 is arranged in the box body 201, and a groove is formed in the magnetic seat 203. The groove is matched with the table body structure at the top of the unmanned aerial vehicle. A second ring-shaped permanent magnet array 231, a second guide magnet 232 and a spring needle 233 are arranged in the magnetic seat 203. The second ring-shaped permanent magnet array 231 is arranged at the bottom of the groove. The spring needle 233 is arranged in the middle of the second ring-shaped permanent magnet array 231. The second guide magnet 232 is uniformly distributed along the side of the groove. The ring-shaped two-dimensional code 205 is arranged on one side of the magnetic seat 203. The detachment mechanism comprises at least four electromagnetic push rods 204, which are arranged around the groove. It should be noted that the warehouse door of the present application is a flexible roller shutter door. A fixed guide rail, a roller shutter door accommodating box and a pull rope accommodating box are arranged below the box body 201. The roller shutter door accommodating box and the pull rope accommodating box are arranged at the two ends of the fixed guide rail respectively. The roller shutter door is controlled by a motor to be pulled out of and returned to the roller shutter door accommodating box, so as to close and open the warehouse door. When the unmanned aerial vehicle performs a monitoring operation, the warehouse door 202 is first pulled out of the roller shutter door accommodating box and opened through the motor linkage at the two ends of the fixed guide rail. Then, the electromagnetic push rod 204 in the warehouse is extended, gradually increasing the distance between the unmanned aerial vehicle and the magnetic seat 203 to weaken the magnetic force. At the same time, the unmanned aerial vehicle synchronously increases the rotor lift to dynamically compensate for the magnetic force attenuation, avoiding the risk of attitude loss of control or falling caused by sudden drop of the magnetic force, and finally realizing stable detachment from the magnetic seat. After the monitoring operation is completed, the unmanned aerial vehicle first completes preliminary positioning by identifying the ring-shaped two-dimensional code 205 of the base station and accurately hovers directly below the magnetic seat 203. As the unmanned aerial vehicle gradually rises, the distance between the unmanned aerial vehicle and the magnetic seat is reduced, and the magnetic force is gradually increased. At this time, the electromagnetic push rod 204 is extended and contacts the unmanned aerial vehicle, forming a flexible limit to prevent the unmanned aerial vehicle from colliding with the magnetic seat 203 due to sudden increase of the magnetic force before the attitude adjustment is completed, reserving sufficient time for the unmanned aerial vehicle to perform attitude calibration and significantly improving the docking accuracy. In addition, a controller is arranged in the box body 201 for whole-process cooperative operation: real-time acquisition of the state of the unmanned aerial vehicle for linkage control of the warehouse door, synchronous storage of the video monitoring data returned by the unmanned aerial vehicle, and transmission of the data to a remote control center to realize fully automated closed-loop management of the monitoring process.

[0036] As Figure 6As shown, in one specific embodiment of the present disclosure, the magnetic pole direction of the first annular permanent magnet array 151 is opposite to that of the second annular permanent magnet array 231, the magnetic pole direction of the first guide magnet 153 is opposite to that of the second guide magnet 232, and the magnetic pole direction of the two adjacent second guide magnets 232 is opposite, which forms a synergistic magnetic force adjustment mechanism: when the UAV flies to the vicinity of the magnetic suction seat 203 and there is a deviation in the horizontal direction, the first guide magnet 153 and the second guide magnet 232 with opposite magnetic pole directions will generate a couple moment in the tangential direction, which can preliminarily correct the horizontal attitude of the UAV; at the same time, the first annular permanent magnet array 151 and the second annular permanent magnet array 231 can further fine-tune the angle deviation of the UAV during the process of providing vertical attraction, so as to realize accurate positioning of the pose. During the UAV ascending docking process, the circumferential magnetic force of the guide magnet synergistically acts to effectively suppress the horizontal shaking, and the vertical attraction of the annular permanent magnet array ensures the stable approach of the UAV, and the two together form a "three-dimensional fixation" effect. Even if there is a small amount of dust on the contact or the surface of the magnet, the magnetic field gradient formed by the synergistic action of the two-stage magnetic force can still accurately push the UAV to adjust the attitude, thereby reducing the influence of the dust in the tunnel on the docking of the UAV, and finally realizing the accurate alignment of the UAV and the magnetic suction seat. This design successfully solves the problems of low docking accuracy, unstable fixation and unreliable function interaction of the UAV and the base station in a tunnel with a lot of dust and a closed environment.

[0037] Embodiment 3

[0038] The present application provides a tunnel wall monitoring method, which comprises:

[0039] Step S1, obtaining an operation instruction, the operation instruction being used to instruct the UAV to perform a wall monitoring operation;

[0040] In this step, the remote control center issues an instruction to replace the manual on-site start of the equipment, breaks the dependence of traditional monitoring on manual presence, lays a foundation for unmanned tunnel wall monitoring, and improves the operation response efficiency.

[0041] Step S2, obtaining initialized positioning information according to the operation instruction, and determining a take-off pose according to the initialized positioning information;

[0042] In this step, the unmanned aerial vehicle performs pre-flight self-checking, preheating detection of the three-dimensional laser radar 104, verification that the motor speed and point cloud density meet the minimum requirements, static zero offset calibration of the INS inertial navigation system, rotation test of the rotating member 102, and positioning initialization after self-checking is completed. The UWB tag carried by the unmanned aerial vehicle communicates with multiple anchor points around it, measures the time difference of the signal arriving at each anchor point based on the anchor point coordinate database, calculates the three-dimensional coordinates of the unmanned aerial vehicle, performs attitude initialization of the INS inertial navigation system 114, measures the gravity vector components by the accelerometer, calculates the initial pitch angle and roll angle, and determines the yaw angle with the assistance of the magnetometer. Subsequently, a Kalman filter is started for tight coupling fusion, the Kalman gain is calculated and the state vector is corrected, until the position fluctuation converges to <0.05 m and the attitude change rate is <0.001 rad / s.

[0043] Step S3, acquiring sensor data, the sensor data including point cloud data, IMU data and UWB data;

[0044] In this step, by collecting multi-dimensional environmental data, the defects of single sensor are compensated to form multi-source data complementation, which provides basis for real-time positioning, attitude adjustment and obstacle avoidance of the unmanned aerial vehicle.

[0045] Step S4, real-time adjusting the take-off pose and position of the unmanned aerial vehicle according to the sensor data, and flying according to the preset route;

[0046] In this step, the extended Kalman filter fuses multi-sensor data to correct the position and attitude in real time (such as using LiDAR to suppress the short-term drift of INS, and using UWB to suppress the cumulative error of LiDAR), and combines the preset tunnel section and building clearance data to realize automatic obstacle avoidance and route constraint, solving the problem of safe flight of the unmanned aerial vehicle in the narrow environment of the tunnel and ensuring the integrity and accuracy of the monitoring coverage.

[0047] In this step, the step S4 includes steps S41, S42, S43, S44, S45 and S46, which specifically include:

[0048] Step S41, aligning the sensor data to obtain aligned sensor data;

[0049] Step S42, acquiring a preset state vector and an error vector;

[0050] In this step, the preset state vector is:

[0051]

[0052] Wherein, p = [p x ,p y ,p z ]T Let v be the three-dimensional coordinates of the UAV in the tunnel's global coordinate system (unit: meters), v = [v x ,v y ,v z ] T For the speed of the UAV in the tunnel global coordinate system

[0053] Degrees (unit: meters per second), q = [q w ,q x ,q y ,q z ] T Quaternions represent the attitude of a drone (unitless, must satisfy...) b g =[b gx ,b gy ,b gz ] T b is the constant zero bias of the IMU gyroscope (unit: radians / second). a =[b ax ,b ay ,b az ] T This is the constant zero bias of the IMU accelerometer (unit: m / s²).

[0054] The preset error vector is:

[0055]

[0056] Where δp=[δp x ,δp y ,δp z ] T Let δv be the position error (m), then δv = [δv x ,δv y ,δv z ] T For the velocity error (m / s), δθ=[δθ x ,δθ y ,δθ z ] T Let δb be the attitude error (rotation vector, rad). g =[δb gx ,δb gy ,δb gz ] T δb represents the gyroscope's zero bias error (rad / s). a =δb ax ,δb ay ,δb az ] T For zero bias error of acceleration (m / s) 2 ).

[0057] Step S43, substituting the aligned sensor data and the preset state vector into the inertial navigation equation to perform state prediction, to obtain an updated first error state and a first error covariance state;

[0058] In this step, the updated first error state includes:

[0059] δx t = F t δx t-1 + G t w t

[0060] wherein F t is a 15x15 error state transition matrix, G t is a 15x12 noise driving matrix, δx t-1 is an error state at t-1 moment, w t is an angular velocity measured by an IMU gyroscope.

[0061] The updated first error covariance state includes:

[0062]

[0063] wherein P t-1 is a covariance matrix at t-1 moment, F t is a Jacobian matrix, which is a partial derivative matrix (dimension 16x16) of a state transition function f(·) with respect to a state vector x at a current state estimation point, and it linearly approximates how the state evolves from t-1 moment to t moment, P t|t-1 is a predicted state covariance matrix at t moment, Q t is a process noise covariance.

[0064] It can be understood that the state prediction by the inertial navigation equation is a technical solution known to those skilled in the art, and thus will not be described here.

[0065] Step S44, updating the updated first error state and the first error covariance state by using a LiDAR observation model combined with point cloud data, to obtain a second error state and a second error covariance state;

[0066] In this step, the LiDAR laser radar receives point cloud once every 100 ms, and the position is corrected. The LiDAR observation model is specifically:

[0067]

[0068] In the above formula, the observation value z LiDAR ∈R 3The three-dimensional position [p x ,p y ,p z ] T , The theoretical absolute position obtained by matching the point cloud with the preset map, v LiDAR The observation noise, H LiDAR =[I3,0 3×12 ]∈R 3×15 The observation matrix, indicating that only the three-dimensional position p of the tunnel global coordinate system in the state vector is directly corrected, x is the preset state vector, and if the residual δ = |z LiDAR -p t | is less than 0.05 m, it indicates that the LiDAR observation is effective, triggering the state update, and the Kalman gain is calculated:

[0069]

[0070] Among them, K t is the Kalman gain, R LiDAR is the covariance matrix of the LiDAR observation noise, and P t|t-1 is the updated first error covariance state.

[0071] The error state update is calculated:

[0072] δx t = K t δz LiDAR

[0073] In the above formula, δz LiDAR is the LiDAR observation residual.

[0074] The covariance state update is:

[0075] P t =(I-K t H LiDAR )P t|t-1

[0076] Among them, I represents the unit matrix.

[0077] Step S45, the updated first error state and the first error covariance state are updated by using the UWB observation model combined with the UWB data, to obtain the third error state and the third error covariance state;

[0078] In this step, UWB ranging values are received every 50 ms in turn, and the position is corrected. The UWB observation model is specifically:

[0079]

[0080] where anchor is the UWB anchor number, z UWB is the distance from the UAV to each anchor, h UWB (x) is the UWB observation model function, v UWB is the observation noise, p anchor is the three-dimensional position of each anchor in the global coordinate system of the tunnel, p is the three-dimensional position of the UAV in the global coordinate system of the tunnel, ||p-p anchor || is the distance from the UAV to the anchor, UWB ranging values are received every 50 ms, after removing outliers, a maximum of 4 anchor data are used to update the state per frame;

[0081] Calculate the residual:

[0082]

[0083] where z UWB is the actual UWB ranging value, is the predicted distance;

[0084] Calculate the Kalman gain:

[0085]

[0086] where P t|t-1 is the updated first error covariance state, H UWB is the observation function h UWB is the partial derivative matrix of the prediction state x t|t-1 , R UWB is the covariance matrix of the UWB observation noise.

[0087] Calculate the error state update:

[0088] δx t = K t δz UWB

[0089] where δz UWB is the UWB observation residual

[0090] Calculate the covariance state update:

[0091] P t = (I-K t H UWB )P t|t-1

[0092] where I represents the identity matrix.

[0093] Step S46, inject the second error state and the second error covariance state and the third error state and the third error covariance state into a preset state vector to obtain a state vector of the current UAV, the state vector of the current UAV being used to update the attitude, speed and position of the UAV.

[0094] In this step, the state vector of the current UAV is obtained after injection, and then the error state and covariance are reset.

[0095] Step S5, when the UAV is flying, the image acquisition device always acquires image information towards the tunnel wall;

[0096] In this step, the image acquisition device is always directed towards the wall by the rotating member 102, and the circumferential light compensation lamp 122 compensates for the uneven lighting in the tunnel, reducing shadows and reflections during shooting; hovering to store images when flying to the preset collection point ensures that the wall conditions are recorded completely.

[0097] It can be understood that when the UAV detects an obstacle, the RTT path planning algorithm is started, the path is re-planned locally, the new path avoids the obstacle while maintaining a preset distance from the tunnel wall, and automatically returns to the original task route after bypassing the obstacle. If the task is in progress, when the power is less than 20%, trigger the breakpoint saving process, save the completed collection point index and the uncompleted collection point index, save the current three-dimensional position, attitude, speed, then use the preset "return path" to return to the base station, and transmit the completed monitoring data to the base station controller and to the remote control center. The UAV is charged to 80% and then takes off again, resets the task route, and continues the task from the next uncompleted collection point, solving the problem of task failure or repeated work caused by insufficient endurance of the UAV in long-distance tunnel monitoring, realizing precise breakpoint continuation when the power is insufficient, avoiding the inefficient problem of forced termination of the task or returning to charge and starting from scratch in the traditional scheme, significantly improving the continuity and execution efficiency of the monitoring task, supporting the closed-loop implementation of full-automatic tunnel monitoring. In addition, the design of real-time transmission of completed data during breakpoint saving can reduce the risk of data loss caused by UAV failure.

[0098] Step S6, when the UAV completes the monitoring task, the image acquisition device rotates upwards below the base station, identifies the base station ring two-dimensional code 205 and adjusts the attitude to fly upwards, and is adsorbed on the magnetic seat 203 through the magnetic attraction guiding charging structure 105;

[0099] In this step, when the annular two-dimensional code 205 appears in a specific position in the image acquisition device, the electromagnetic push rod 204 extends to contact the rising unmanned aerial vehicle, and then the electromagnetic push rod 204 slowly retracts. When the unmanned aerial vehicle deviates horizontally, the magnetic field gradient formed by the two-stage magnetic force can still accurately push the unmanned aerial vehicle to adjust the posture, so that the magnetic attraction guiding charging structure 105 is adsorbed on the magnetic attraction seat 203, the spring needle 233 is pressed to be conductive, and charging and data transmission start. The video and image data are transmitted to the base station controller, the controller transmits the image data to the remote control control center. The application solves the docking accuracy problem in the dusty environment of the tunnel through visual positioning (recognizing the annular two-dimensional code) combined with magnetic attraction guiding. The magnetic poles of the magnetic attraction structure are designed alternately to generate a self-correcting torque, which cooperates with the flexible limiting of the electromagnetic push rod to realize the posture correction of the inclined unmanned aerial vehicle, greatly improves the docking success rate, and realizes the unmanned monitoring operation.

[0100] Step S7, uploading the image information to the remote control center to generate a tunnel wall health report.

[0101] In this step, the remote control center has a target detection function, which can identify cracks and leaking water information in the wall, record the crack length, identify the wall pollution situation and gray scale, and compare with the last monitoring image to obtain the wall cleanliness decline trend and crack growth trend, generate a tunnel wall health report. It should be noted that the control center uses a target detection algorithm to identify cracks and pollutants, including: pre-training a crack detection data set; the control center receives video and image data transmitted remotely from the base station; video data is archived, and image data is imported into a crack target detection algorithm; identify the position and length of the crack in the image and archive it; identify the number and area of the pollution area in the image and archive it; generate a crack and pollution report; compare with the last pollution report to generate a tunnel wall health report containing crack growth and cleanliness decline.

[0102] In the description of the application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, or the orientation or positional relationship commonly used when the product is used, and are only for the convenience of describing the application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the application. In addition, the terms "first", "second", "third" and the like are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0103] In the description of the application, it also needs to be explained that, unless otherwise explicitly specified and limited, the terms "set", "install", "connect", "connect" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be directly connected, or indirectly connected through an intermediate medium, or it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the application can be understood according to the specific circumstances.

[0104] The above is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0105] The above is only the specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A tunnel monitoring drone, characterized in that, Include: The unmanned aerial vehicle body (101) is provided with a cantilever around the unmanned aerial vehicle body (101), the cantilever is fixedly connected with the side wall of the unmanned aerial vehicle body (101), and the cantilevers are uniformly distributed along the circumference of the unmanned aerial vehicle body (101); Rotary member (102), the rotary member (102) is arranged on the side wall of the unmanned aerial vehicle body (101); Rotor (103), the rotor (103) is provided with four, the rotor (103) is installed at the end of the cantilever; Image acquisition device, the image acquisition device is arranged on the rotary member (102), and the image acquisition device is rotatably connected with the rotary member (102); Magnetic attraction guiding and charging structure (105), the magnetic attraction guiding and charging structure (105) is arranged at the top of the unmanned aerial vehicle body (101); Three-dimensional laser radar (104), the three-dimensional laser radar (104) is arranged at the bottom of the unmanned aerial vehicle body (101).

2. The tunnel monitoring drone of claim 1, wherein: The end of the cantilever is fixedly provided with a motor fixing seat, a motor is arranged in the motor fixing seat, a power output shaft of the motor extends in the vertical direction, and the rotor (103) is horizontally arranged at the top end of the power output shaft.

3. The tunnel monitoring drone of claim 1, wherein: The image acquisition device includes a high-definition camera (121) and a fill light (122), the fill light (122) is arranged along the circumference of the high-definition camera (121), and the illumination direction of the fill light (122) is the same as the shooting direction of the high-definition camera (121).

4. The tunnel monitoring drone of claim 1, wherein: The unmanned aerial vehicle body (101) is internally provided with a central control device, a UWB tag and an INS inertial navigation system.

5. The tunnel monitoring drone of claim 1, wherein: The magnetic attraction guiding and charging structure (105) is a table body structure, the magnetic attraction guiding and charging structure (105) includes a first annular permanent magnet array (151), a charging and data transmission contact (152) and a first guiding magnet (153), the first annular permanent magnet array (151) is arranged at the top of the table body structure, the charging and data transmission contact (152) is arranged in the first annular permanent magnet array (151), the first guiding magnet (153) is arranged on the side surface of the table body structure, and the first guiding magnet (153) is uniformly distributed along the circumference of the table body structure.

6. The tunnel monitoring drone of claim 5, wherein: The magnetic pole directions of two adjacent first guiding magnets (153) are opposite, and the magnetic force arrangement directions of two adjacent magnets in the first annular permanent magnet array (151) are opposite. 7.A tunnel monitoring unmanned aerial vehicle base station, characterized in that, Include: Box (201), the box (201) is fixedly arranged at the top of the tunnel and located outside the tunnel building limit; Warehouse door (202), the warehouse door (202) is arranged at the bottom of the box (201); A magnetic seat (203) is arranged inside the box (201), the magnetic seat (203) is provided with a groove, the groove is matched with the table body structure on the top of the unmanned aerial vehicle, the second annular permanent magnet array (231), the second guide magnet (232) and the spring needle (233) are arranged in the magnetic seat (203), the second annular permanent magnet array (231) is arranged at the bottom of the groove, the spring needle (233) is arranged in the second annular permanent magnet array (231), and the second guide magnet (232) is uniformly distributed along the side of the groove; A ring-shaped two-dimensional code (205) is arranged on one side of the magnetic seat (203); The disengagement mechanism includes at least four electromagnetic push rods (204), and the electromagnetic push rods (204) are arranged around the groove.

8. The drone base station of claim 7, wherein: The magnetic pole direction of the first annular permanent magnet array (151) is opposite to that of the second annular permanent magnet array (231), the magnetic pole direction of the first guide magnet (153) is opposite to that of the second guide magnet (232), and the magnetic pole directions of two adjacent second guide magnets (232) are opposite.

9. A method of monitoring a tunnel wall surface, characterized by, Comprise: Obtaining a work instruction, the work instruction is used for instructing the unmanned aerial vehicle to carry out wall monitoring work; According to the initialization of the positioning information, the take-off pose is determined; Obtaining sensor data, the sensor data includes point cloud data, IMU data and UWB data; According to the sensor data, the take-off pose and position of the unmanned aerial vehicle are adjusted in real time, and the unmanned aerial vehicle flies along the preset route; When the unmanned aerial vehicle flies, the image acquisition device always faces the tunnel wall to obtain image information; When the unmanned aerial vehicle completes the monitoring work, it is located below the base station, the image acquisition device rotates upward, identifies the ring-shaped two-dimensional code (205) of the base station and adjusts the attitude to fly upward, and is adsorbed on the magnetic seat (203) through the magnetic attraction guide charging structure (105); The image information is uploaded to the remote control center to generate a tunnel wall health report.

10. The tunnel wall monitoring method according to claim 9, wherein According to the sensor data, the take-off pose and position of the unmanned aerial vehicle are adjusted in real time, and the unmanned aerial vehicle flies along the preset route; Aligning the sensor data to obtain aligned sensor data; Obtaining a preset state vector and an error vector; The aligned sensor data and the preset state vector are substituted into the inertial navigation equation to perform state prediction, and updated first error state and first error covariance state are obtained; The updated first error state and first error covariance state are updated by using a LiDAR observation model combined with point cloud data to obtain second error state and second error covariance state; The updated first error state and first error covariance state are updated by using a UWB observation model combined with UWB data to obtain third error state and third error covariance state; The second error state and the second error covariance state and the third error state and the third error covariance state are injected into a preset state vector to obtain a state vector of the current UAV, and the state vector of the current UAV is used to update a pose, a speed and a position of the UAV.

Images