CN118196611A - Dyke leakage dangerous case quick disposal device based on vehicle-mounted platform and underwater robot - Google Patents

Abstract

The invention discloses a rapid dam leakage dangerous case disposal device based on a vehicle-mounted platform and an underwater robot, and relates to the technical field of dam engineering leakage dangerous case rescue. Compared with the prior dam leakage plugging device, the device improves the conditions of poor mobility and poor material one-time supply of the existing equipment, and for the problem of poor plugging effect of larger leakage points, the device is combined with generation of an opposing network through a digital twin technology, the characteristic reconstruction of the leakage points is carried out based on the characteristic reconstruction of the learning points, the quality of the characteristic of the leakage points is improved through a characteristic reinforcing algorithm, meanwhile, the problem of model gradient disappearance is solved through an activating function, the corresponding plugging material feeding quantity prediction is finally carried out, the problem of one-time supply shortage of the plugging materials can be solved, the effective utilization rate of the plugging materials is greatly improved by utilizing the positioning operation of an underwater robot, and the device is flexible, flexible and strong in applicability, and can meet the rapid treatment of larger leaks and leakage point groups.

Description

Rapid disposal device for dam leakage based on vehicle-mounted platform and underwater robot

Technical Field

The invention relates to the technical field of emergency rescue of leakage danger of dam engineering, and in particular to a device for quickly handling leakage danger of dam engineering based on a vehicle-mounted platform and an underwater robot.

Background technique

Leakage is a common hazard in dam projects. If it is not handled promptly and effectively, it can easily develop into disasters such as landslides and breaches. Therefore, timely and efficient sealing of leakage points is an important measure to ensure the safety of dams and eliminate risks. Traditional methods of sealing leakage in dams include laying anti-seepage blankets, throwing clay, grouting, etc., but they are very costly. In recent years, new plugging technologies have gradually developed. The existing Chinese patent number CN111501675A, "Underwater leakage self-priming plugging method based on negative pressure suction", has been applied to dam leakage plugging. This method uses phase change materials to self-prime and plug the dam, but the equipment has problems such as poor maneuverability and insufficient one-time supply of materials. The plugging effect on larger leakage points is not good, which limits the application of this type of method.

In order to solve the above problems, the present invention proposes a device for quickly handling dangerous situations of dike leakage based on a vehicle-mounted platform and an underwater robot.

Summary of the invention

The purpose of the present invention is to propose a device for quickly handling dike leakage danger based on a vehicle-mounted platform and an underwater robot to solve the problems raised in the background technology:

The existing equipment has poor mobility, is prone to insufficient material supply at one time, and is not effective in sealing larger leaks.

In order to achieve the above object, the present invention adopts the following technical solutions:

A rapid disposal device for dike leakage based on a vehicle-mounted platform and an underwater robot, including: a vehicle-mounted platform, a robotic arm, a material storage cabin, a delivery pipe, an above-water beacon, an underwater robot and a central control module;

The vehicle-mounted platform is the carrier of the entire device and is used to carry each module of the device for movement;

The mechanical arm is installed on the vehicle-mounted platform and is used for rotation and extension. The end of the mechanical arm is connected to an underwater robot;

The material storage cabin is installed at the rear of the vehicle-mounted platform and is used to provide materials for plugging leakage points;

The delivery pipe is arranged along the mechanical arm, connecting the material storage tank and the underwater robot, and is used to deliver the plugging material;

The above-water beacon is connected to the underwater robot and is used to provide an above-water reference for positioning the underwater robot;

The underwater robot is used to judge the leakage risk of the dike and control the material delivery module to deliver the plugging materials at the leakage point;

The central control module is used to transmit, process and control device information;

The water beacon includes a buoy body, a positioning module, and a mooring module;

The buoy body floats on the water surface and serves as a carrier of the entire water beacon;

The positioning module is used to locate the position coordinates of the buoy body in real time;

The mooring module includes a mooring rope, a retracting module and a mooring sensor;

The mooring rope is used to connect the buoy body and the underwater robot; the retracting module is used to make the mooring rope scaled as the underwater robot moves, so that the mooring rope between the buoy body and the underwater robot remains in a tensioned state; the mooring rope sensor is used to obtain the length and angle of the mooring rope released;

The relative position of the underwater robot and the water beacon is calculated by the length and angle of the mooring rope, and the spatial position of the underwater robot is calculated by using the positioning data obtained in real time by the water beacon; the position of the underwater robot is located or manipulated;

The underwater robot includes a detection module, a propulsion module, a manipulator, a cleaning module, a video module, an inkjet module, and a material delivery module;

The detection module includes a groundwater level sensor, a seepage rate monitoring instrument and an underwater geological radar; the groundwater level sensor is used to monitor the groundwater level in real time, the seepage rate monitoring instrument is used to measure the water flow rate in the levee for real-time monitoring; the underwater geological radar is used to detect the infiltration point detected by the video module;

The propulsion module is used to provide power to the underwater robot and control the movement of the underwater robot;

The manipulator is controlled to perform gripping and rotating operations, and is used to position and operate the conveying pipe in an underwater environment;

The cleaning module is used to clean up debris by means of a scraper and a brush;

The video module is used to record underwater images;

The inkjet module is used to spray pigment to indicate the direction of leakage water flow, thereby finding the leakage point;

The material delivery module includes a heat-insulating flexible edge guard connected to the end of the conveying pipe for releasing the plugging material.

Preferably, the robotic arm is a three-axis type for realizing multi-angle rotation.

Preferably, the material storage compartment has a pressure regulating function, and the leakage point plugging material stored in the material storage compartment is a material that can be melted by heating.

Preferably, the delivery pipe has a heat preservation function.

Preferably, the central control module is used to control the rotation of the operating robot arm, the movement, cleaning, operation of the underwater robot, video recording and playback, coordinate feedback of the underwater robot, control the placement of plugging materials, and also receive detection feedback signals from the detection module of the underwater robot.

The method for using the device for quickly handling dike leakage danger based on the vehicle-mounted platform and the underwater robot comprises the following steps:

S1: Control the robotic arm to release the underwater robot and the water beacon, and use the detection module to preliminarily detect whether there are leakage points in the dam. After the detection is completed, the underwater robot is recovered;

S2: The central control module simulates and analyzes the characteristics of the leakage point based on the data collected by the underwater robot's detection module and calculates the amount of plugging material to be put in;

S3: Check whether there is enough material in the material storage compartment and whether the functions of each module are normal;

S4: Move the entire device to the leakage area through the vehicle-mounted platform;

S5: Control the robotic arm again to release the underwater robot and the water beacon;

S6: Control the underwater robot to the leakage point by manual operation or by inputting coordinate points;

S7: Accurately determine the leakage point through the video module and inkjet module;

S8: Control the underwater robot to fix the material delivery module at the underwater leakage point;

S9: The plugging material is pumped from the material storage tank to the material delivery module through the delivery pipe by pumping, and the plugging material flows from the leakage point to the leakage point through the material delivery module, thereby plugging the leakage point;

S10: The underwater robot performs leakage detection again through the detection module. If the detected dam has no leakage, the processing is completed. If the detected dam still has leakage points, S7 to S10 are repeated until the dam has no leakage.

Preferably, in S2, the central control module receives the detection feedback signal of the detection module of the underwater robot, and simulates the characteristics of the leakage point of the levee through the digital twin technology, reconstructs the characteristics of the leakage point based on the generative adversarial network, and calculates the amount of plugging material to be put in, and predicts the location and corresponding characteristics of the possible future seepage points of the levee based on the transfer learning method;

In the generative adversarial network, the LSTM network is used to establish the basic model in the generative adversarial network framework, the collected leakage point features are enhanced, and applied to both the generator and the discriminator:

The image features of the previous hidden layer are converted into two feature space functions, and the feature map hj _,i is obtained by softmax normalization:

Where _Wa and _Wb represent learning weights; _xi represents the i-th position data; _xj represents the j-th region data; N represents the product of the width and height of the convolution channel;

The obtained feature map _hj,i is multiplied pixel by pixel through a 1*1 convolution unit to obtain the feature-enhanced feature map _yi . The output _yi of the feature map is multiplied by the scale parameter λ and added back to the input feature map:

_zi ＝ _λyi + _xi

Among them, z _i is the final output of the feature map;

Except for the last layer, the generator of the model uses the rectified linear unit as the activation function, and the last layer of the generator uses the hyperbolic tangent activation function. Except for the last layer, the discriminator uses the leaky rectified linear unit as the activation function, and the last layer of the discriminator uses the sigmoid activation function. In this way, a generative adversarial network is constructed to reconstruct the leakage point features.

Preferably, in S7, when the underwater vision is clear, the leakage point is observed through the video module to accurately find the leakage point; when the underwater water is turbid or the vision is unclear, the inkjet module is used to perform inkjet operation and then the position of the leakage point is determined according to the ink flow direction, and the robot is manipulated to approach the position; when approaching the embankment, the video module is used for close observation to accurately find the leakage point.

Compared with the prior art, the present invention provides a rapid disposal device for dike leakage based on a vehicle-mounted platform and an underwater robot, which has the following beneficial effects:

The present invention combines digital twin technology with generative adversarial networks, and reconstructs leakage point features based on learned reconstructed point features. The quality of leakage point features is improved through a feature enhancement algorithm, and the problem of model gradient vanishing is solved through an activation function. Finally, the corresponding plugging material delivery amount is predicted, which can solve the problem of insufficient one-time supply of plugging materials. The underwater robot positioning and manipulation is used for precise plugging, which greatly improves the effective utilization rate of plugging materials. The device is flexible and maneuverable, highly applicable, and can meet the needs of rapid disposal of larger leaks and leakage point groups.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the structural connection of the device mentioned in Example 1 of the present invention;

FIG2 is a schematic diagram of the partial structural connection of the device mentioned in Example 1 of the present invention;

FIG3 is a flow chart of the method mentioned in Example 2 of the present invention.

Meaning of the symbols in the figure:

1. Vehicle-mounted platform; 2. Robotic arm; 3. Material storage compartment; 4. Delivery pipe; 5. Water beacon; 510. Buoy body; 520. Positioning module; 530. Mooring module; 531. Mooring rope; 532. Retracting and releasing module; 533. Mooring sensor; 6. Underwater robot; 610. Detection module; 620. Propulsion module; 630. Robotic arm; 640. Cleaning module; 650. Video module; 660. Inkjet module; 670. Material delivery module; 7. Central control module.

Detailed ways

The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

The present invention combines digital twin technology with generative adversarial networks, reconstructs leakage point features based on learned reconstruction point features, improves the quality of leakage point features through feature enhancement algorithms, and solves the problem of model gradient vanishing through activation functions. Finally, the corresponding plugging material delivery amount is predicted, which can solve the problem of insufficient one-time supply of plugging materials, and uses underwater robots for positioning and manipulation to perform precise plugging, which greatly improves the effective utilization rate of plugging materials. The device is flexible and mobile, highly applicable, and can meet the needs of rapid disposal of large leaks and leakage point groups. Specifically, it includes the following contents.

Embodiment 1:

Please refer to Figures 1-2, the rapid disposal device for dike leakage based on a vehicle-mounted platform and an underwater robot of the present invention comprises: a vehicle-mounted platform 1, a mechanical arm 2, a material storage cabin 3, a conveying pipe 4, an above-water beacon 5, an underwater robot 6 and a central control module 7;

The vehicle-mounted platform 1 is an off-road flatbed vehicle, which is composed of a cockpit, an operation cabin and an equipment platform from front to back. The control terminal of the central control module 7 in the operation cabin allows technicians to operate each module; the equipment platform is equipped with modules such as a robotic arm 2, a material storage cabin 3, a conveying pipe 4, and an underwater robot 6.

The arm length of the robot arm 2 is 10m, and it is a three-axis type that can achieve multi-angle rotation.

The material storage cabin 3 is a sealed stainless steel tank with heating and heat preservation functions. An exhaust valve, a pressurizing device, etc. are installed on the upper part. A gate valve is installed at the lower part to connect the delivery pipe 4.

The delivery pipe 4 is a hose with a heat-insulating function and has an inner diameter of 20 mm.

The water beacon 5 is a floating hollow plastic ball, in which a positioning module 520 and a mooring module 530 are installed, and the ball can float on the water surface.

The underwater robot 6 is selected to have a detection module 610, a crawler chassis, a propeller, a manipulator 630, a video module 650, and a lighting system. The material delivery module 670 is independently made, with a medium-density sponge frame on the outside and a hose end on the inside, and the two are sealed with potting glue.

The central control module 7 is respectively embedded in each component and connected to the station platform control cabin.

Embodiment 2:

Please refer to FIG3 , the present invention is based on the device for quickly handling leakage danger of dikes in Example 1, and the device for quickly handling leakage danger includes the following steps:

S1: Control the robotic arm 2 to release the underwater robot 6 and the water beacon 5, and use the detection module 610 to preliminarily detect whether there is a leakage point in the dam. After the detection is completed, the underwater robot 6 is recovered; the details are as follows:

The groundwater level sensor and seepage rate monitoring instrument of the detection module 610 detect the change of groundwater level and the rate of water flow, and preliminarily judge that there is leakage in a certain embankment section K36+100~K36+152. The leakage area can be determined by underwater geological radar.

S2: The central control module 7 simulates and analyzes the characteristics of the leakage point based on the data collected by the detection module 610 of the underwater robot 6 and calculates the amount of plugging material to be put in; the details are as follows:

The central control module 7 reconstructs the characteristics of the leakage points of the dam based on the acquired data based on the data twin technology, and can calculate the required amount of plugging material to facilitate the accurate loading of materials in the material storage compartment 3 and the subsequent one-time release.

S3: Check whether the material in the material storage compartment 3 is sufficient and whether the functions of each module are normal; the details are as follows:

Prepare sufficient plugging materials based on the calculated plugging material delivery amount to avoid insufficient materials during the plugging process. At the same time, check whether other modules function normally to ensure that the next step is performed after normal loading.

S4: Move the entire device to the leakage area via the vehicle-mounted platform 1; details are as follows:

Based on the initially determined location of the leakage area, the entire device can be moved to the leakage area via the vehicle-mounted platform 1 .

S5: controlling the robotic arm 2 again to release the underwater robot 6 and the water beacon 5;

S6: Control the underwater robot 6 to the leakage point by manual operation or by inputting coordinate points; the details are as follows:

The underwater robot 6 is controlled to the leakage point by manual operation or inputting the coordinate points according to the coordinates of the leakage point determined by the underwater geological radar.

S7: Accurately determine the location of the leakage point through the video module 650 and the inkjet module 660; the details are as follows:

By collecting underwater video through the video module 650, the exact location of the leakage point can be accurately found. If the underwater is turbid or the vision is unclear due to water quality problems or the movement of the underwater robot 6, the lighting system can be used to illuminate the water and find the leakage point. If the underwater robot 6 has an inkjet module 660, the inkjet module 660 can be used to perform inkjet operation to determine the direction of the leakage point according to the ink flow direction, and the underwater robot 6 can be controlled to approach the direction; when approaching the leakage point, the video module 650 can be used to observe closely to accurately find the leakage point.

S8: Control the underwater robot 6 to fix the material delivery module 670 at the underwater leakage point;

S9: The plugging material is pumped from the material storage tank 3 to the material delivery module 670 through the delivery pipe 4 by pumping, and the plugging material flows from the leakage point to the leakage point through the material delivery module 670, thereby plugging the leakage point;

S10: The underwater robot 6 performs leakage detection again through the detection module 610. If the detected dam has no leakage, the process is completed. If the detected dam still has leakage points, S7 to S10 are repeated until the dam has no leakage. The details are as follows:

The underwater robot 6 performs another detection through the groundwater level sensor and seepage rate monitoring instrument of the detection module 610. If the detected groundwater level and water flow rate are both within the normal value range, the surface leakage point is blocked. If the monitoring is still within the abnormal value range, it is judged that there is still leakage in the dam. The leakage point is detected again by underwater geological radar and the corresponding sealing material is filled.

The plugging material delivery amount calculated based on the leakage point reconstructed by the digital twin algorithm and the standard delivery amount within an error of ±5% are taken as accurate data, and the other data are taken as inaccurate data. The single LSTM network, the traditional LSTM network based on the generative adversarial network, i.e., 1-LSTM, and the LSTM network based on the generative adversarial network of the present invention, i.e., 2-LSTM, are used to calculate the plugging material delivery amount for leakage points at 10 locations. The accuracy rates can be referred to Table 1:

Table 1 Accuracy of plugging material placement for 10 leakage points using LSTM, 1-LSTM and 2-LSTM

As can be seen from Table 1, the generative adversarial network is beneficial to improving the feature quality of LSTM network reconstruction. Feature enhancement measures can improve the quality of reconstructed features and obtain leakage point feature information more comprehensively, which is beneficial to improving the training effect of the model.

The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims (8)

1. A device for rapidly handling dike leakage hazards based on a vehicle-mounted platform and an underwater robot, characterized in that it comprises: a vehicle-mounted platform (1), a mechanical arm (2), a material storage compartment (3), a conveying pipe (4), an above-water beacon (5), an underwater robot (6) and a central control module (7); The vehicle-mounted platform (1) is a carrier of the entire device and is used to carry various modules of the device for movement; The mechanical arm (2) is mounted on the vehicle-mounted platform (1) and is used for rotation and extension, and the end of the mechanical arm (2) is connected to an underwater robot (6); The material storage cabin (3) is installed at the rear of the vehicle-mounted platform (1) and is used to provide materials for plugging leakage points; The conveying pipe (4) is arranged along the mechanical arm (2), connecting the material storage cabin (3) and the underwater robot (6), and is used to convey the plugging material; The above-water beacon (5) is connected to the underwater robot (6) and is used to provide an above-water reference object for positioning the underwater robot (6); The underwater robot (6) is used to judge the leakage risk of the dike and control the material delivery module (670) to deliver the plugging material to the leakage point; The central control module (7) is used to transmit, process and control device information; The water beacon (5) comprises a buoy body (510), a positioning module (520), and a mooring module (530); The buoy body (510) floats on the water surface and serves as a carrier of the entire water beacon (5); The positioning module (520) is used to locate the position coordinates of the buoy body (510) in real time; The mooring module (530) comprises a mooring cable (531), a retracting and releasing module (532) and a mooring sensor (533); The mooring rope (531) is used to connect the buoy body (510) and the underwater robot (6); the retracting module (532) is used to make the mooring rope (531) scale as the underwater robot (6) moves, so that the mooring rope (531) between the buoy body (510) and the underwater robot (6) remains in a tensioned state; the mooring rope sensor (533) is used to obtain the length and angle of the mooring rope (531) being released; The relative position of the underwater robot (6) and the water beacon (5) is calculated by the length and the angle of release of the mooring rope (531), and the spatial position of the underwater robot (6) is calculated by using the positioning data acquired in real time by the water beacon (5); thereby positioning or manipulating the position of the underwater robot (6); The underwater robot (6) comprises a detection module (610), a propulsion module (620), a manipulator (630), a cleaning module (640), a video module (650), an inkjet module (660), and a material delivery module (670); The detection module (610) comprises a groundwater level sensor, a seepage rate monitoring instrument and an underwater geological radar; the groundwater level sensor is used to monitor the groundwater level in real time, the seepage rate monitoring instrument is used to measure the water flow rate in the levee for real-time monitoring; the underwater geological radar is used to detect the infiltration point detected by the video module (650); The propulsion module (620) is used to provide power for the underwater robot (6) and control the movement of the underwater robot (6); The manipulator (630) controls the clamping and rotating operations, and is used to position and operate the conveying pipe (4) in an underwater environment; The cleaning module (640) is used to clean debris by using a scraper and a brush; The video module (650) is used to record underwater images; The inkjet module (660) is used to spray pigment to indicate the direction of leakage water flow, thereby finding the leakage point; The material delivery module (670) comprises a heat-insulating flexible edge guard connected to the end of the delivery pipe (4) for releasing the plugging material. 2. According to the device for quickly handling dike leakage hazards based on a vehicle-mounted platform and an underwater robot as described in claim 1, it is characterized in that the mechanical arm (2) is a three-axis type for realizing multi-angle rotation. 3. According to the device for rapid disposal of dike leakage hazards based on a vehicle-mounted platform and an underwater robot as described in claim 1, it is characterized in that the material storage compartment (3) has a pressure regulation function, and the leakage point plugging material stored in the material storage compartment (3) is a material that can be melted by heating. 4. According to the device for quickly handling dike leakage hazards based on a vehicle-mounted platform and an underwater robot as described in claim 1, it is characterized in that the delivery pipe (4) has a heat preservation function. 5. According to claim 1, the rapid disposal device for dike leakage hazards based on a vehicle-mounted platform and an underwater robot is characterized in that the central control module (7) is used to control the rotation of the operating robot arm (2), the movement, cleaning, and operation of the underwater robot (6), video recording and playback, coordinate feedback of the underwater robot (6), control the placement of plugging materials, and also receive the detection feedback signal of the detection module (610) of the underwater robot (6). 6. A method for rapidly handling dike leakage danger based on a vehicle-mounted platform and an underwater robot used in the device according to any one of claims 1 to 5, characterized in that it comprises the following steps: S1: Controlling the mechanical arm (2) to release the underwater robot (6) and the water beacon (5), using the detection module (610) to preliminarily detect whether there is a leakage point in the dam, and recovering the underwater robot (6) after the detection is completed; S2: The central control module (7) simulates and analyzes the characteristics of the leakage point based on the data collected by the detection module (610) of the underwater robot (6) and calculates the amount of plugging material to be put in; S3: Check whether the material in the material storage compartment (3) is sufficient and whether the functions of each module are normal; S4: moving the entire device to the leakage area via the vehicle-mounted platform (1); S5: controlling the robotic arm (2) again to release the underwater robot (6) and the water beacon (5); S6: Controlling the underwater robot (6) to the leakage point by manual operation or by inputting coordinate points; S7: Accurately determine the location of the leakage point through the video module (650) and the inkjet module (660); S8: controlling the underwater robot (6) to fix the material delivery module (670) at the underwater leakage point; S9: Pumping the plugging material from the material storage compartment (3) through the delivery pipe (4) to the material delivery module (670), and the plugging material flows from the leakage point into the leakage point through the material delivery module (670), thereby plugging the leakage point; S10: The underwater robot (6) performs leakage detection again through the detection module (610). If the detected dam is leak-free, the process is completed. If the detected dam still has leakage points, S7 to S10 are repeated until the dam is leak-free. 7. The method for rapidly handling dike leakage hazards based on a vehicle-mounted platform and an underwater robot according to claim 6 is characterized in that, in S2, the central control module (7) receives the detection feedback signal of the detection module (610) of the underwater robot (6), and simulates the characteristics of the dike leakage point through the digital twin technology, reconstructs the leakage point characteristics based on the generative adversarial network, and calculates the amount of plugging material to be put in, and predicts the location and corresponding characteristics of the possible future seepage points of the dike based on the transfer learning method; In the generative adversarial network, the LSTM network is used to establish the basic model in the generative adversarial network framework, the collected leakage point features are enhanced, and applied to both the generator and the discriminator: The image features of the previous hidden layer are converted into two feature space functions, and the feature map hj _,i is obtained by softmax normalization: Where _Wa and _Wb represent learning weights; _xi represents the i-th position data; _xj represents the j-th region data; N represents the product of the width and height of the convolution channel; The obtained feature map _hj,i is multiplied pixel by pixel through a 1*1 convolution unit to obtain the feature-enhanced feature map _yi . The output _yi of the feature map is multiplied by the scale parameter λ and added back to the input feature map: _zi ＝ _λyi + _xi Among them, z _i is the final output of the feature map; Except for the last layer, the generator of the model uses the rectified linear unit as the activation function, and the last layer of the generator uses the hyperbolic tangent activation function. Except for the last layer, the discriminator uses the leaky rectified linear unit as the activation function, and the last layer of the discriminator uses the sigmoid activation function. In this way, a generative adversarial network is constructed to reconstruct the leakage point features. 8. The method for using the device for rapid disposal of dike leakage hazards based on a vehicle-mounted platform and an underwater robot according to claim 6 is characterized in that, in S7, when the underwater vision is clear, the leakage point is observed through the video module (650) to accurately find the leakage point; when the underwater is turbid or the vision is unclear, the inkjet module (660) performs inkjet operation and then determines the position of the leakage point according to the ink flow direction, and controls the robot to approach the position; when approaching the dike slope, the video module (650) is used for close observation to accurately find the leakage point.