Unmanned aerial vehicle paying-off device and using method

Abstract

The invention discloses an unmanned aerial vehicle pay-off device and a using method, and belongs to the technical field of power transmission line construction, the unmanned aerial vehicle pay-off device comprises an unmanned aerial vehicle, a central control unit, a pay-off device, a tension pay-off device, a guide rope and an identification device, the pay-off device is installed at the bottom of the unmanned aerial vehicle, and the pay-off device is connected with one end of the guide rope in a clamping mode; the other end of the guide rope is connected with a cable unwound by the tension pay-off device, the tension pay-off device is arranged on the ground, the pay-off speed of the tension pay-off device is controlled to be consistent with the flight speed of the unmanned aerial vehicle, so that the guide rope and the cable are in a straightened state, and the situation that pay-off is affected by too large sag of the cable is avoided; the unmanned aerial vehicle, the pay-off device and the tension pay-off device are all in signal connection with the central control unit, operation of workers is facilitated, accurate positioning of the unmanned aerial vehicle in complex terrains is ensured, and meanwhile, an identification device is arranged on a tower where a cable needs to be installed so that the unmanned aerial vehicle can accurately recognize the preset position.

Description

Technical Field

[0001] The present invention belongs to the technical field of power transmission line construction, and in particular relates to an unmanned aerial vehicle (UAV) line-laying device and a use method thereof. Background Art

[0002] With the rapid development of the power grid, electricity has become an indispensable part of people's daily life, and there are more and more new line projects. Transmission lines are important tools for power and information transmission. Before the conductors of new lines are laid out, guide ropes are required to connect each tower in the tension section, and then the guide ropes are used to drag the conductors to achieve the laying of the conductors.

[0003] With the development of drone technology, the use of drones to deploy guide ropes and cables can achieve the goal of crossing obstacles such as rivers, woods, and roads, making it easier for workers to build transmission lines faster. However, in the current drone line-laying technology, the only way is to use the drone to carry the end of the guide rope to the target tower and then use the servo to drop the guide rope, which will cause the entire guide rope to lose tension. The guide rope may even fall into rivers, woods, and roads, affecting normal traffic operations, and may even cause friction or jamming of the guide rope or cable in the woods, making it impossible to move forward or rise, seriously affecting the quality and efficiency of engineering construction. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the present invention provides a drone line-releasing device and a method for using the device.

[0005] The present invention adopts the following technical solution.

[0006] The first aspect of the present invention discloses a UAV wire-releasing device, comprising: a UAV, a central control unit, a wire-releasing device, a tension wire-releasing device, a guide rope and an identification device. The wire-releasing device is installed at the bottom of the UAV, the wire-releasing device clamps one end of the guide rope, and the other end of the guide rope is connected to a cable released by the tension wire-releasing device. The tension wire-releasing device is arranged on the ground, and the wire-releasing speed of the tension wire-releasing device is controlled to be consistent with the flight speed of the UAV, so as to make the guide rope and the cable in a straightened state, thereby avoiding the sag height of the cable affecting the wire-releasing, and the UAV, the wire-releasing device and the tension wire-releasing device are all connected to the central control unit by signal, and an identification device is arranged on the pole tower where the cable needs to be installed, so that the UAV can identify the preset position.

[0007] Preferably, the drone has a built-in management system, which includes: a positioning unit, and the central control unit includes: a map system. The positioning unit and the map system signals of the central control unit are connected, so that the staff can control the flight path and line-laying position of the drone through the central control unit, wherein the positioning unit includes: a high-precision positioning system, a visual positioning system and an inertial measurement system. The high-precision positioning system is positioned through GPS or GLONASS, the visual positioning system VPS and the high-precision positioning system are operated in combination, and the inertial measurement unit uses IMU to provide flight attitude and heading information.

[0008] Preferably, the management system also includes: an image unit, and the central control unit also includes: a monitoring system. The monitoring system and the image unit in the management system of the UAV are electrically connected to facilitate the staff to obtain the flight status and line-laying progress of the UAV in real time. The image unit includes: a front camera, a rear camera and a visual processing module. The front camera and the rear camera are arranged on the UAV to fully capture the image information of the preset pole tower. The visual processing module is responsible for processing and analyzing the images captured by the camera for automatic flight control of the UAV.

[0009] Preferably, the visual processing module includes: an input layer, a convolution layer, a residual block, a global average pooling layer and an output layer. The visual processing module uses Mobile_ResNet for image processing. The input layer is connected to both the front camera and the rear camera. The input layer receives the image taken by the camera and performs preprocessing, and then the preprocessed image is transmitted to the convolution layer. The convolution layer uses depthwise separable convolution and standard convolution to jointly extract image features of the pole tower and the identification device. The convolution layer contains multiple convolution kernels for capturing features of different scales and directions. At the same time, a residual block is introduced to construct a deep network. The residual block contains multiple convolution layers and jump connections, so that the input data can be directly passed to the output layer, avoiding the problem of gradient disappearance. The image with extracted features is transmitted to the global average pooling layer for global average pooling and extracting global features. Finally, the image is transmitted to the output layer, and the output layer uses a fully connected layer and an activation function for classification or regression tasks.

[0010] Preferably, the wire-releasing device has a built-in release structure and a tension sensor, the release structure allows the guide rope and the cable to be smoothly released, the tension sensor and the tension wire-releasing device are signal-connected, and then the tension sensor protects the guide rope and the cable from overload through the tension wire-releasing device.

[0011] Preferably, the tension sensor has a preset tension threshold, and the tension threshold is divided into three levels: a safe range, a warning range and a dangerous range. When the guide rope tension is within the safe range, the tension-wire-releasing device normally releases the cable; when the cable tension reaches the warning range, the tension-wire-releasing device automatically stops releasing the cable and sends a signal alarm to the central control unit to alert the operator; when the guide rope tension reaches the dangerous range, the tension-wire-releasing device automatically stops releasing the cable and sends an emergency alarm to the central control unit.

[0012] Preferably, a physical sensor is installed on the guide rope to monitor the stress change and vibration of the guide rope in real time.

[0013] Preferably, the management system of the drone is also provided with an obstacle avoidance system, which is a fusion global planning algorithm based on the Q-Learning algorithm and the particle swarm optimization algorithm. It creates a Q table to store the Q value of each state-action pair, integrates Q-Learning with PSO and performs iterative updates, shares experience and adjusts parameters, then extracts the global optimal solution, outputs the global optimal solution as the solution to the problem, and enables the drone to perform obstacle avoidance movements according to the state-action corresponding to the Q value.

[0014] The second aspect of the present invention discloses a method for using a drone line-releasing device, based on the drone line-releasing device, comprising the following steps:

[0015] Survey the construction area and obtain geographic information and data;

[0016] Conduct pre-flight inspection and debugging of drones;

[0017] Connect the drone to the wire-laying device, install the guide rope in the wire-laying device, connect the guide rope to the wire to be deployed, and set the marking device on the designated tower;

[0018] According to the survey results, the geographical information and line-laying requirements are input through the central control unit, thereby controlling the flight path and line-laying position of the drone;

[0019] The staff monitors the flight status and line-release progress of the drone in real time through the central control unit and makes fine adjustments as needed;

[0020] After the drone has completed the line release, it returns to the take-off point.

[0021] Preferably, the staff monitors the flight status and line-release progress of the drone in real time through the central control unit, and makes fine adjustments as needed, including:

[0022] The drone flies according to the path preset by the central control unit;

[0023] When the drone approaches the marking device, the camera captures the image information of the tower area, and the visual processing module processes and analyzes the image to identify the location of the tower;

[0024] Based on the recognition results, the drone automatically adjusts its flight attitude and path, allowing it to fly over the tower and complete the deployment of the cables.

[0025] The beneficial effect of the present invention is that, compared with the prior art, when line engineering construction personnel use drones to lay out wires, they install a wire-laying device on the drone, and then connect the end of the guide rope to the wire-laying device, and the other end of the guide rope is connected to the cable deployed by the tension wire-laying device. The drone carries the wire-laying device connected to the guide rope, and adjusts the wire-laying speed to be consistent with the flight speed, so that the guide rope and the cable are in a straightened state, and the sag is not too large to hang on the trees below, and even affect the traffic; after the end of the guide rope is fixed on the target tower by the wire-laying device, the guide rope is tightened by the tension wire-laying device, so that the guide rope is tensioned along the entire line, and the sag of the guide rope and the cable is increased, and they will not fall to the ground or the woods, effectively reducing the impact on traffic and forest obstacles, and greatly improving the reliability of the wire-laying work.

[0026] At the same time, identification devices are set up on the poles or fixed cross bars where cables need to be installed, so that drones can be accurately identified and guided to fly.

[0027] The drone, wire-paying device and wire-tensioning device are all connected to the central control unit via signals, making it convenient for staff to operate. At the same time, the visual positioning system and high-precision positioning system in the drone work together to ensure accurate positioning of the drone in complex terrain. The inertial measurement unit helps the drone maintain stability in complex environments.

[0028] At the same time, a tension sensor is set in the wire-releasing device to monitor the tension status of the guide rope and the cable in real time; the tension sensor sets the tension threshold by preset to protect the guide rope and the cable from overload.

[0029] The obstacle avoidance system of the UAV combines the strategy learning ability of Q-Learning and the global search ability of particle swarm optimization, which improves the performance and efficiency of the algorithm. It introduces the experience sharing strategy to enhance the learning ability between particles and balance the global and local search capabilities of the algorithm. It uses Q-Learning to update the parameters of the particle swarm algorithm online, which improves the adaptability and robustness of the algorithm, enabling the UAV to avoid obstacles more accurately and quickly.

[0030] The image unit of the drone is highly separable through deep convolution and residual connection. Mobile_ResNet can significantly reduce the amount of calculation and the number of parameters while ensuring high accuracy. It also has great scalability and can adjust the network depth and width according to specific application requirements to balance the computational overhead and model performance. It combines the stability of ResNet and the efficiency of MobileNet. Mobile_ResNet is more efficient, more stable and has significant performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a first structural diagram of a UAV line-releasing device of the present invention;

[0032] Figure 2 This is a second structural diagram of a UAV line-releasing device of the present invention;

[0033] In the figure: 1. UAV; 2. Wire-releasing device; 3. Tension wire-releasing device; 4. Guide rope; 5. Marking device; 6. Cable. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0035] like Figure 1-2 As shown, the present invention discloses a UAV line-releasing device and a method for using the device, the device comprises: a UAV 11, a central control unit, a marking device 5, a line-releasing device 2 and a tension line-releasing device 3, the line-releasing device 2 is installed at the bottom of the UAV 11, the line-releasing device 2 is clamped to one end of a guide rope 4, the other end of the guide rope 4 is connected to a cable 6 deployed by the tension line-releasing device 3, the tension line-releasing device 3 is arranged on the ground, and the line-releasing speed of the tension line-releasing device 3 is adjusted to be consistent with the flight speed of the UAV 11, so that the UAV 11 can drive one end of the guide rope 4 to fly at a uniform speed, and at the same time, the other end of the guide rope 4 is connected to the cable 6, and the cable 6 is released at a uniform speed through the tension line-releasing device 3, so that the guide rope 4 and the cable 6 are kept in a taut state, so as to prevent the guide rope 4 and the cable 6 from drooping too much and hanging on the trees below, and even affecting the traffic. The UAV 11, the wire-releasing device 2 and the tension wire-releasing device 3 are all connected to the central control unit by signal. The central control unit is operated by staff. At the same time, an identification device 5 is set on the tower or fixed crossbar where the cables need to be installed, so that the UAV 11 can be accurately identified and guided to fly.

[0036] The central control unit includes: a map system, which is used to plan the flight path and line-laying position of the drone 11.

[0037] The UAV 11 adopts a multi-rotor UAV with good vertical take-off and landing and hovering capabilities, and is suitable for precise operations in complex terrain. The UAV 11 has a built-in management system, which includes: a positioning unit, a perception unit and an image unit. The positioning unit includes: a high-precision positioning system, a visual positioning system and an inertial measurement system. The positioning unit is connected to the map system signal of the central control unit, and can perform mobile positioning of the UAV 11 through the map system of the central control unit. The high-precision positioning system performs precise positioning through global navigation satellite systems such as GPS or GLONASS to ensure that the UAV 11 can accurately reach the preset line-laying position. The visual positioning system (VPS) and the high-precision positioning system are combined to ensure the precise positioning of the UAV 11 in complex terrain. The inertial measurement unit uses a high-precision IMU to provide accurate attitude and heading information to help the UAV 11 maintain stability in complex environments.

[0038] The central control unit 7 also includes a monitoring system, which is connected to the sensing unit and the image unit in the drone 11 by signals, so that the operator can grasp the flight status and line-laying progress of the drone 11 in real time.

[0039] The perception unit senses the surrounding environment by fusing multiple sensors, such as fusing data from lidar, camera and ultrasonic sensor, to improve environmental perception capability and enhance the stability of the drone. At the same time, the perception unit can transmit drone 11 data to the monitoring system of the central control unit 7.

[0040] The image unit includes: a front camera, a rear camera and a visual processing module. The front camera and the rear camera are set on the drone 11 to fully capture the image information of targets such as poles and fixed crossbars. The visual processing module is responsible for processing and analyzing the images captured by the camera to identify the poles and identification devices 5 to achieve automatic flight control of the drone. At the same time, the image unit is connected to the monitoring system signal in the central control unit 7, so that the staff can obtain the flight status and line-laying progress of the drone 11.

[0041] The visual processing module adopts Mobile_ResNet, which is a neural network model that combines the characteristics of MobileNet depthwise separable convolutions (Depthwise Separable Convolutions) and ResNet residual network. Mobile_ResNet provides an efficient convolutional neural network (CNN) architecture for the drone 11.

[0042] The convolution method of MobileNet significantly reduces the amount of calculation and model parameters, and improves the calculation efficiency. At the same time, ResNet solves the gradient vanishing problem in deep neural network training by introducing residual blocks and skip connections. Mobile_ResNet combines the efficient calculation of MobileNet and the model performance of ResNet, ensuring high accuracy while significantly reducing the amount of calculation and the number of parameters, making Mobile_ResNet suitable for efficient image processing tasks in resource-constrained environments.

[0043] At the same time, the visual processing module includes: input layer, convolution layer, residual block, global average pooling layer and output layer. The input layer is connected to both the front camera and the rear camera. The input layer receives the images taken by the two cameras and performs necessary preprocessing, such as resizing and normalization. Then the preprocessed images are transmitted to the convolution layer. The convolution layer uses depth-separable convolution and standard convolution to jointly extract the image features of the tower and the identification device 5. The convolution layer contains multiple convolution kernels for capturing features of different scales and directions. At the same time, the residual block is introduced to construct a deep network. The residual block contains multiple convolution layers and jump connections, so that the input data can be directly transmitted to the output layer, thereby avoiding the influence of the gradient disappearance problem. The image with extracted features is transmitted to the global average pooling layer, which performs global average pooling on the image with extracted features to reduce the dimension of the feature map and extract global features. Finally, the image is transmitted to the output layer, and the output layer uses the fully connected layer Dense Layer and the activation function Softmax for classification or regression tasks.

[0044] The wire-releasing device 2 is used to be connected to one end of the guide rope 4 to be released, and has good load-bearing and buffering capabilities. The built-in release structure of the wire-releasing device 2 can smoothly release the guide rope 4 and the cable 6.

[0045] At the same time, a tension sensor is arranged in the pay-off device 2 for real-time monitoring of the tension status of the guide rope 4 and the cable 6; the tension sensor presets a tension threshold, and sets a reasonable tension threshold according to the material, specification of the cable and the specific requirements of the pay-off task, and the tension sensor is connected with the tension pay-off device 3 by signal, and then the tension sensor performs overload protection on the guide rope 4 and the cable 6 through the tension pay-off device 3, and the tension threshold is divided into three levels: a safe range, a warning range and a dangerous range; when the guide rope tension does not exceed the safe range, the tension pay-off device normally pays off the cable; when the cable tension reaches the warning range, the tension pay-off device automatically stops paying off the cable, and sends a signal alarm to the central control unit to remind the operator to pay attention; when the guide rope tension reaches the dangerous range, the tension pay-off device automatically stops paying off the cable, and sends an emergency alarm to the central control unit.

[0046] Preferably, the alarm should alert the operator through multiple means such as sound or light to ensure timely action.

[0047] Physical sensors, such as strain gauges or acceleration sensors, are installed on the guide rope 4 to monitor the stress changes and vibrations of the guide rope 4 in real time. When the physical sensor detects that the stress of the guide rope 4 suddenly decreases or the vibration is abnormal, it will be judged that the guide rope 4 is broken, and the central control unit will immediately stop the operation of the wire-releasing device 2 and issue an emergency alarm. The drone 11 quickly adjusts its flight attitude to ensure that it hovers or lands in a safe area to avoid further losses.

[0048] Preferably, the drone 11 management system also adopts an adaptive control algorithm, so that the drone can automatically adjust flight parameters according to environmental changes in real time to maintain stability. At the same time, the AI ​​algorithm is used to enable the drone to learn from experience and optimize its flight behavior to cope with complex flight environments; the drone 11 management system also sets up a multiple backup system, using multiple IMU sensors and GPS modules to ensure that when one component fails, there are still spare components that remain functional.

[0049] The drone 11 management system is also provided with an obstacle avoidance system, which is a fusion global planning algorithm based on the Q-Learning algorithm and the particle swarm optimization algorithm. The obstacle avoidance system is a new algorithm that combines the advantages of the Q-Learning algorithm in reinforcement learning and the particle swarm optimization algorithm (PSO).

[0050] Q-Learning is a model-free reinforcement learning algorithm that learns a strategy to maximize long-term cumulative rewards through the agent's exploration and utilization of the environment. Q-Learning finds the optimal strategy by learning a state-action value function (Q function). In Q-Learning, the agent learns by interacting with the environment, and its Q value is updated after each interaction to reflect the expected returns from taking different actions in different states.

[0051] Particle swarm optimization is an optimization algorithm based on swarm intelligence. It searches for the optimal solution through collaboration and competition among particles. Particle swarm optimization simulates the foraging behavior of biological groups such as bird flocks or fish schools. In PSO, each particle represents a potential solution and updates its position and velocity through an iterative process to find the optimal solution.

[0052] The fusion algorithm combines the strategy learning ability of Q-Learning with the global search ability of particle swarm optimization to improve the performance and efficiency of the algorithm.

[0053] In a specific embodiment, the state space and action space are first determined. The state space is all possible states that the drone 11 may be in, and the action space is all possible actions that the drone 11 can take. Then the Q table is initialized: a Q table is created to store the Q value corresponding to each state-action. The initialization of the Q table is usually zero or a random value. The particle swarm is initialized: the particle swarm is initialized, including the position, speed and individual optimal solution of each particle. The position of the particle can be regarded as a potential solution to the problem, and the speed determines the direction and speed of the particle's movement in the search space. Integrate Q-Learning with PSO: During the movement of drone 11, changes will occur and iterative updates will be made. In each iteration, drone 11 selects an action based on the current state and Q table, and executes the action to observe the next state and reward. Then, the Q-Learning update rule is used to update the Q table. At the same time, the particle swarm also updates its position and velocity based on the individual historical optimal solution and the global optimal solution. Combine Q-Learning with PSO, combine the update of the Q table with the position update of the particle swarm, and enable the Q table to update the Q value in the Q table according to environmental feedback. For example, use the Q value to guide the movement direction of the particle in the search space or use the Q table as an additional information source in the particle swarm optimization process. At the same time, experience sharing and parameter adjustment are performed, and an experience sharing strategy is designed so that particles can share the experience of the optimal particles through the Q table, accelerate the convergence of the Q table, and enhance the learning ability between particles. In addition, Q-Learning is used to update the inertia weight and acceleration parameters of the particle swarm algorithm online, so as to improve the convergence ability of the particle swarm algorithm and avoid falling into the local optimum. Then the global optimal solution is extracted: during the iteration process, the global optimal solution, that is, the position of the particle with the highest Q value, is continuously tracked and recorded. When the termination condition is met, such as reaching the maximum number of iterations or the Q value converges, the global optimal solution is output as the solution to the problem, so that the drone 11 can perform obstacle avoidance according to the state-action corresponding to the Q value. The obstacle avoidance system guides the search process of the particle swarm optimization through the strategy learned by the Q-Learning algorithm, thereby improving the efficiency and accuracy of the search. Then, the global search capability of the particle swarm optimization algorithm is used to conduct an extensive search in the solution space to find the global optimal solution or the approximate optimal solution, thereby achieving a good obstacle avoidance function.

[0054] Preferably, the drone 11 management system is also provided with an energy management system to optimize battery usage and extend flight time, especially in low temperature environments.

[0055] The present invention also discloses a method for using the drone line-releasing device, comprising the following steps:

[0056] Step 1: Survey the construction area and obtain geographic information and related data;

[0057] Step 2: Perform pre-flight inspection and debugging on the drone 11 to ensure the working status of the drone 11;

[0058] Step 3: connect the drone 11 to the wire-releasing device 2, install the guide rope 4 in the wire-releasing device 2, connect the other end of the guide rope 4 to the conductor 6 released by the tension wire-releasing device 3, and set the marking device 5 on the designated tower;

[0059] Step 4: According to the survey results, the geographical information and line-laying requirements are input into the central control unit, thereby controlling the flight path and line-laying position of the UAV 11;

[0060] Step 5: The staff monitors the flight status and cable laying progress of the drone 11 in real time through the central control unit, and makes fine adjustments as needed to carry out the laying of the cable 6.

[0061] Step 5.1: The drone 11 flies according to the path preset by the central control unit;

[0062] Step 5.2: When the drone 11 approaches the identification device 5, the camera captures the image information of the tower area, and the visual processing module processes and analyzes the image to identify the location of the tower;

[0063] Step 5.3: Based on the recognition result, the drone 11 automatically adjusts the flight attitude and path so that it can accurately fly over the tower and complete the deployment of the cable 6.

[0064] Step 6: After the drone 11 has completed the line release, it returns to the take-off point.

[0065] The beneficial effect of the present invention is that, compared with the prior art, when line engineering construction personnel use drones to lay out wires, they install a wire-laying device on the drone, and then connect the end of the guide rope to the wire-laying device, and the other end of the guide rope is connected to the cable deployed by the tension wire-laying device. The drone carries the wire-laying device connected to the guide rope, and adjusts the wire-laying speed to be consistent with the flight speed, so that the guide rope and the cable are in a straightened state, and the sag is not too large to hang on the trees below, and even affect the traffic; after the end of the guide rope is fixed on the target tower by the wire-laying device, the guide rope is tightened by the tension wire-laying device, so that the guide rope is tensioned along the entire line, and the sag of the guide rope and the cable is increased, and they will not fall to the ground or the woods, effectively reducing the impact on traffic and forest obstacles, and greatly improving the reliability of the wire-laying work.

[0066] At the same time, identification devices are set up on the poles or fixed cross bars where cables need to be installed, so that drones can be accurately identified and guided to fly.

[0067] The drone, wire-paying device and wire-tensioning device are all connected to the central control unit via signals, making it convenient for staff to operate. At the same time, the visual positioning system and high-precision positioning system in the drone work together to ensure accurate positioning of the drone in complex terrain. The inertial measurement unit helps the drone maintain stability in complex environments.

[0068] At the same time, a tension sensor is set in the wire-releasing device to monitor the tension status of the guide rope and the cable in real time; the tension sensor sets the tension threshold by preset to protect the guide rope and the cable from overload.

[0069] The obstacle avoidance system of the UAV combines the strategy learning ability of Q-Learning and the global search ability of particle swarm optimization, which improves the performance and efficiency of the algorithm. It introduces the experience sharing strategy to enhance the learning ability between particles and balance the global and local search capabilities of the algorithm. It uses Q-Learning to update the parameters of the particle swarm algorithm online, which improves the adaptability and robustness of the algorithm, enabling the UAV to avoid obstacles more accurately and quickly.

[0070] The image unit of the drone is highly separable through deep convolution and residual connection. Mobile_ResNet can significantly reduce the amount of calculation and the number of parameters while ensuring high accuracy. It also has great scalability and can adjust the network depth and width according to specific application requirements to balance the computational overhead and model performance. It combines the stability of ResNet and the efficiency of MobileNet. Mobile_ResNet is more efficient, more stable and has significant performance.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A UAV line-releasing device, comprising: The unmanned aerial vehicle (11), the central control unit, the wire-releasing device (2), the tension wire-releasing device (3), the guide rope (4) and the marking device (5) are characterized in that: A wire-releasing device (2) is installed at the bottom of the drone (11). The wire-releasing device (2) is clamped to one end of a guide rope (4). The other end of the guide rope (4) is connected to a cable (6) released by a tension wire-releasing device (3). The tension wire-releasing device (3) is arranged on the ground. By controlling the wire-releasing speed of the tension wire-releasing device (3) to be consistent with the flight speed of the drone (11), the guide rope (4) and the cable (6) are in a straightened state, thereby preventing the sag height of the cable (6) from affecting the wire-releasing. The drone (11), the wire-releasing device (2) and the tension wire-releasing device (3) are all connected to a central control unit by signal. At the same time, an identification device (5) is arranged on the pole tower where the cable needs to be installed, so that the drone (11) can identify the preset position.

2. The UAV line-releasing device according to claim 1, characterized in that: The unmanned aerial vehicle (11) has a built-in management system, which includes a positioning unit, and a central control unit includes a map system. The positioning unit and the map system of the central control unit are connected by signal, so that the staff can control the flight path and the line-laying position of the unmanned aerial vehicle (11) through the central control unit, wherein the positioning unit includes a high-precision positioning system, a visual positioning system and an inertial measurement system. The high-precision positioning system is positioned by GPS or GLONASS, the visual positioning system VPS and the high-precision positioning system are operated in combination, and the inertial measurement unit uses an IMU to provide flight attitude and heading information.

3. The UAV line-releasing device according to claim 2, characterized in that: The management system further comprises: an image unit, and the central control unit further comprises: a monitoring system. The monitoring system and the image unit in the management system of the unmanned aerial vehicle (11) are connected by electrical signals, so that the staff can obtain the flight status and line laying progress of the unmanned aerial vehicle (11) in real time. The image unit comprises: a front camera, a rear camera and a visual processing module. The unmanned aerial vehicle (11) is provided with a front camera and a rear camera for comprehensively capturing image information of a preset pole tower. The visual processing module is responsible for processing and analyzing the image captured by the camera for automatic flight control of the unmanned aerial vehicle.

4. The UAV line-releasing device according to claim 3, characterized in that: The visual processing module comprises: an input layer, a convolution layer, a residual block, a global average pooling layer and an output layer. The visual processing module uses Mobile_ResNet to process images. The input layer is connected to both the front camera and the rear camera. The input layer receives the image taken by the camera and performs preprocessing. The preprocessed image is then transmitted to the convolution layer. The convolution layer uses depthwise separable convolution and standard convolution to jointly extract image features of the pole tower and the identification device (5). The convolution layer contains multiple convolution kernels for capturing features of different scales and directions. At the same time, a residual block is introduced to construct a deep network. The residual block contains multiple convolution layers and jump connections, so that input data can be directly transmitted to the output layer, avoiding the problem of gradient disappearance. The image with extracted features is transmitted to the global average pooling layer for global average pooling and extracting global features. Finally, the image is transmitted to the output layer. The output layer uses a fully connected layer and an activation function to perform classification or regression tasks.

5. The UAV line-releasing device according to claim 1, characterized in that: The wire-releasing device (2) has a built-in release structure and a tension sensor. The release structure allows the guide rope (4) and the cable (6) to be stably extended. The tension sensor and the tension wire-releasing device (3) are connected by signal. Then, the tension sensor protects the guide rope (4) and the cable (6) from overload through the tension wire-releasing device (3).

6. The drone line-releasing device according to claim 5, characterized in that: The tension sensor has a preset tension threshold value, which is divided into three levels: a safe range, a warning range, and a dangerous range. When the tension of the guide rope (4) is within the safe range, the tension wire-releasing device (3) normally releases the cable (6); when the cable tension reaches the warning range, the tension wire-releasing device (3) automatically stops releasing the cable (6) and sends a signal alarm to the central control unit to alert the operator; when the tension of the guide rope (4) reaches the dangerous range, the tension wire-releasing device (3) automatically stops releasing the cable (6) and sends an emergency alarm to the central control unit.

7. The UAV line-releasing device according to claim 1, characterized in that: A physical sensor is installed on the guide rope (4) for real-time monitoring of stress changes and vibration conditions of the guide rope (4).

8. The UAV line-releasing device according to claim 1, characterized in that: The management system of the drone (11) is also provided with an obstacle avoidance system. The obstacle avoidance system is a fusion global planning algorithm based on the Q-Learning algorithm and the particle swarm optimization algorithm. The system creates a Q table to store the Q value of each state-action pair, fuses Q-Learning with PSO and performs iterative updates, shares experience and adjusts parameters, then extracts the global optimal solution, outputs the global optimal solution as a solution to the problem, and enables the drone (11) to perform obstacle avoidance movement according to the state-action corresponding to the Q value.

9. A method for using a drone line-releasing device, based on the drone line-releasing device according to any one of claims 1 to 8, characterized in that: Survey the construction area and obtain geographic information and data; Performing pre-flight inspection and debugging on the UAV (11); The drone (11) is connected to the wire-releasing device (2), a guide rope (4) is installed in the wire-releasing device (2), the guide rope (4) is connected to the wire (6) to be deployed, and a marking device (5) is set on a designated tower; According to the survey results, geographic information and line-laying requirements are input through the central control unit, thereby controlling the flight path and line-laying position of the drone (11); The staff monitors the flight status and line release progress of the drone (11) in real time through the central control unit and makes fine adjustments as needed; After the drone (11) has completed the line release, it returns to the take-off point.

10. The method for using the drone line-releasing device according to claim 9, characterized in that: The staff monitors the flight status and line-release progress of the drone (11) in real time through the central control unit, and makes fine adjustments as needed, including: The drone (11) flies according to a path preset by the central control unit; When the drone (11) approaches the identification device (5), the camera captures image information of the pole tower area, and the visual processing module processes and analyzes the image to identify the position of the pole tower; According to the recognition result, the drone (11) automatically adjusts the flight attitude and path, so that the drone (11) flies over the tower and completes the deployment of the cable (6).

Images