A laser obstacle clearing robot construction method and device based on AI vision dynamic aiming

Abstract

The application discloses a kind of laser clearing robot construction method and device based on AI vision dynamic aiming, belong to electric power clearing equipment technical field.The device contains mobile carrier, AI vision aiming, laser cutting, automatic control and power supply five big components: mobile carrier adopts caterpillar type structure, can adapt complex terrain;AI vision aiming component can accurately identify the branch invading 1 meter safety area of feeder and calculate cutting point;Laser cutting component can complete accurate cutting of branch;Automatic control component automatically formulates cutting scheme and from high to low cutting order, autonomously controls robot movement, aiming and cutting, avoids branch frame insertion, pressure broken feeder;Power supply component provides stable power for field operation.Its construction method covers component assembly, system commissioning, optimization debugging and operation maintenance.The application realizes clearing operation fully automatic, precision, safety, solves the problem of low efficiency, poor safety, easy to damage feeder of traditional clearing, can effectively guarantee the safe operation of transmission line.

Description

Technical Field

[0001] This invention belongs to the field of power obstacle clearing equipment technology, specifically relating to a method and device for constructing a laser obstacle clearing robot based on AI visual dynamic aiming. Background Technology

[0002] Transmission lines are the core carriers of power transmission, and their safe and stable operation is directly related to the reliability of the power supply system. During the long-term operation of transmission lines, the growth of trees around the lines can easily cause branches to intrude into the feeder's safety zone. When branches come into contact with or are too close to the feeder, they can easily cause short circuits, tripping, and other faults, and in severe cases, even large-scale power outages, affecting industrial production and residents' lives. Therefore, timely removal of branches that intrude into the feeder's safety zone is an important task to ensure the safe operation of transmission lines.

[0003] Existing methods for clearing obstructions from power transmission lines are mainly divided into two categories: manual clearing and mechanical clearing. Manual clearing requires workers to climb trees or use high-altitude work equipment to cut branches with handheld cutting tools. This method is inefficient, labor-intensive, and poses safety hazards such as falls and electric shocks, especially in mountainous areas and complex terrain, where the safety and feasibility of manual clearing are significantly reduced. Mechanical clearing mainly uses large-scale clearing equipment, which is bulky, inflexible, and cannot adapt to the complex terrain around power transmission lines. It also mostly requires manual assistance to aim at the cutting point and determine the cutting sequence, making fully automated clearing impossible. In addition, problems such as branches getting stuck or breaking feeders are common during the cutting process, resulting in insufficient safety and accuracy in clearing.

[0004] With the development of AI vision technology and laser cutting technology, laser obstacle removal robots are gradually being applied to the field of power transmission line obstacle removal. However, existing laser obstacle removal robots still have many shortcomings: First, they lack precise AI vision dynamic aiming function, cannot automatically identify the safe distance between tree branches and feeders, and have difficulty accurately locating the branches to be cut and the cutting point; second, they cannot automatically formulate cutting plans and cutting order, requiring manual intervention and resulting in low work efficiency; third, unreasonable cutting order during the cutting process can easily lead to the cut branches sticking together or falling and breaking feeders, posing a safety hazard; and fourth, their mobility is insufficient, making them unable to adapt to complex terrain operations, which limits their application scope.

[0005] Therefore, developing a laser obstacle removal robot and its construction method that can move autonomously with tracks, dynamically aim with AI vision, automatically formulate cutting plans and reasonable cutting sequences, accurately remove branches that intrude into the safety area of ​​the feeder, and avoid branches from sticking up or breaking the feeder, has become an urgent technical problem to be solved in the field of power obstacle removal equipment. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and device for constructing a laser obstacle removal robot based on AI vision dynamic aiming. This invention solves the problems of low efficiency, poor safety, and insufficient accuracy of existing obstacle removal methods, as well as the inability of existing laser obstacle removal robots to operate fully automatically, unreasonable branch cutting sequence, and easy damage to feeder lines. It enables fully automatic, precise, and safe operation of tree branch obstacle removal on power transmission lines.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A laser obstacle removal robot device based on AI vision dynamic aiming includes a mobile carrier component, an AI vision aiming component, a laser cutting component, an automatic control component, and a power supply component.

[0009] The mobile carrier assembly adopts a tracked self-propelled structure, adaptable to the complex terrain around power transmission lines, such as hillsides, gullies, and areas overgrown with weeds. It includes a tracked chassis, a drive motor, and a steering mechanism. The tracked chassis uses high-strength, wear-resistant tracks, effectively improving ground traction and preventing slippage in complex terrain. The drive motor is a DC servo motor with adjustable output power, connected to the tracked chassis via gear transmission, providing power for the chassis's movement. The steering mechanism uses a differential steering structure, enabling flexible steering of the tracked chassis, facilitating precise robot access to the obstacle-clearing area without manual assistance, thus reducing labor intensity.

[0010] The AI ​​vision targeting component is the core of achieving accurate obstacle clearance, including an optical lens, an automatic gimbal, and an AI vision recognition module. The optical lens uses a high-definition industrial camera with a resolution of no less than 1080P and a frame rate of no less than 25fps, capable of capturing high-definition images of the tree canopy in real time and transmitting them to the AI ​​visual recognition module. The automatic gimbal uses a servo-driven gimbal with a rotation angle range of 0°-360° horizontally and -90°-90° vertically, with a rotation accuracy of ±0.1°. The optical lens and laser emitter are fixedly mounted on the automatic gimbal, enabling omnidirectional scanning and dynamic aiming by following the rotation of the automatic gimbal, ensuring that all branches intruding into the safety area can be captured. The AI ​​visual recognition module has a built-in YOLOv8-based visual recognition model that has been trained and optimized with a large number of tree canopy, branch, and feeder samples from different environments and tree species. It can accurately distinguish between feeders and branches, identify the tree canopy outline in real time, and calculate the distance between branches and feeders. When a branch intrudes into the feeder within a 1-meter safety radius, it is automatically marked as a cutting area, and the optimal cutting point is calculated and determined based on the thickness, growth, and spatial position of the branch, providing data support for the formulation of the cutting plan.

[0011] The laser cutting assembly, comprising a laser generator, a laser emitter, and a focusing module, is used for precise cutting of tree branches. The laser generator is a fiber laser generator with adjustable power based on the branch thickness, offering high cutting efficiency and a clean cut, adaptable to cutting branches of varying thicknesses. The laser emitter is fixedly connected to an automatic gimbal and synchronized with the optical lens, adjusting the emission angle according to the gimbal's rotation to ensure precise laser beam alignment with the cutting point. The focusing module uses a convex lens focusing structure to focus the laser beam into a small spot, increasing the energy density of the laser cutting, ensuring rapid branch severing, and reducing debris generated during the cutting process. Furthermore, the laser cutting assembly includes a laser protection module; a protective shell prevents laser leakage and injury, while a beam shield prevents debris from splashing onto the feeder surface and causing damage.

[0012] The automatic control component is the core control unit for the robot's fully automated operation, including a scheme planning module, a motion control module, an aiming control module, and a wireless communication module. The scheme planning module is signal-connected to the AI ​​vision recognition module. Based on the tree canopy image and information about the branches to be cut obtained by the AI ​​vision recognition module, including their location, thickness, and distance from the feeder, it can automatically formulate a cutting scheme and cutting sequence. The cutting sequence strictly follows the principle of cutting from high to low. At the same time, considering the spatial position of the branches, it prioritizes cutting the branches closest to the feeder and the most threatening ones, avoiding cutting the feeder along the branches' falling trajectory, ensuring that the cut branches do not get stuck or break the feeder. The motion control module is signal-connected to the mobile carrier component. It can control the movement, stopping, and turning of the tracked chassis according to the operation requirements, realizing the robot's autonomous movement. The aiming control module is signal-connected to the automatic gimbal, AI vision recognition module, and laser cutting component. Based on the cutting scheme formulated by the scheme planning module, it can control the automatic gimbal to drive the optical lens and laser emitter to aim at the target cutting point. During the cutting process, the cutting point is dynamically changed according to the cutting progress to ensure that all branches to be cut are accurately cut. The wireless communication module has multiple communication modes, which can realize the reception of remote control commands and the real-time transmission of operation data. It supports remote start and stop of the robot, adjustment of cutting parameters, and modification of cutting schemes, which facilitates remote monitoring of the operation process by the staff and timely intervention in case of abnormalities.

[0013] The power supply components include a lithium battery pack and a charging module. The lithium battery pack uses high-capacity lithium iron phosphate batteries with a capacity of not less than 100Ah and a battery life of not less than 8 hours, which can meet the needs of long-term field operations. The charging module supports fast charging, which can quickly charge the lithium battery pack. It also has overcharge, over-discharge, and short-circuit protection functions to ensure safe and stable power supply and adapt to field operation environments without external power sources.

[0014] Furthermore, the tracked chassis is also equipped with a shock absorption module. The shock absorption module adopts a spring shock absorption structure, which can reduce the vibration of the robot during movement in complex terrain and avoid the vibration affecting the accuracy of AI vision aiming and the stability of laser cutting.

[0015] Furthermore, the AI ​​visual recognition module also has a real-time tracking function, which can dynamically track the positional changes of branches during the cutting process, adjust the cutting point in time, avoid cutting deviations caused by branch shaking, and ensure cutting accuracy.

[0016] This invention also provides a method for constructing a laser obstacle removal robot based on AI vision dynamic aiming, comprising the following steps:

[0017] Step 1: Assembly of mobile carrier components

[0018] High-strength, wear-resistant tracks are selected to manufacture the tracked chassis. The drive motor and steering mechanism are assembled and connected to the tracked chassis. The transmission coordination between the drive motor and the tracked chassis is adjusted to ensure that the tracked chassis can move smoothly and steer flexibly, with the moving speed adjustable between 0.1-0.5 m / s. Shock-absorbing modules are installed on the tracked chassis to reduce vibration during movement. Installation interfaces are reserved for fixing AI vision aiming components, laser cutting components, automatic control components, and power supply components to ensure that each component is installed securely.

[0019] Step 2: Assembly of AI vision aiming component and laser cutting component

[0020] The high-definition industrial camera and laser emitter are fixedly mounted on a servo-driven gimbal. The rotation accuracy of the automatic gimbal is adjusted to ensure that the optical lens and laser emitter are synchronized and the rotation angle is precisely controllable. The AI ​​vision recognition module is connected to the optical lens signal, and the trained YOLOv8 vision recognition model is loaded. The recognition accuracy of the model is adjusted to ensure that the model can accurately identify tree crowns, branches, and feeders, accurately determine the area where branches intrude into the feeder within a 1-meter safety radius, and calculate and determine the optimal cutting point with a recognition accuracy of no less than 98%. The fiber laser generator and convex lens focusing module are connected to the laser emitter. The laser power and focusing effect are adjusted to ensure that the laser beam can be accurately focused and can quickly cut branches of different thicknesses. The laser protection module is installed and its protective performance is adjusted to ensure that the laser does not leak and that debris does not fly onto the feeder.

[0021] Step 3: Deployment of Automated Control Components

[0022] The scheme planning module, motion control module, aiming control module, and wireless communication module are integrated and installed in a control box, which is fixedly mounted on the tracked chassis. A wired + wireless dual signal connection is established between the automatic control components and the mobile carrier components, AI vision aiming components, and laser cutting components to ensure stable signal transmission. The collaborative performance of each module is tested: the motion control module can precisely control the movement, stopping, and steering of the tracked chassis; the aiming control module can control the automatic gimbal to drive the optical lens and laser emitter to precisely aim at the cutting point; the scheme planning module can automatically generate a reasonable cutting scheme and a cutting sequence from high to low based on AI vision recognition data; and the wireless communication module can realize remote control command reception and real-time transmission of operation data.

[0023] Step 4: Power supply component assembly and system commissioning

[0024] Install the high-capacity lithium iron phosphate battery pack and fast-charging module in the reserved positions on the tracked chassis, complete the wiring connection with each power component, and test the power supply stability to ensure that the lithium battery pack has a range of no less than 8 hours. The charging module has overcharge, over-discharge, and short-circuit protection functions. Start the entire robot system and conduct joint debugging tests: control the tracked chassis to move to the simulated work area, scan the simulated tree canopy through the AI ​​vision recognition module to verify the accuracy of safe area intrusion judgment and cutting point calculation; generate a cutting plan through the scheme planning module to verify the rationality of the branch cutting sequence and ensure that the cut branches do not get stuck or break the simulated feeder; debug the dynamic aiming and cutting point changing functions of the automatic gimbal, start the laser cutting component, and verify the cutting accuracy and efficiency; conduct safety tests to verify the protective performance of the laser protection module and ensure operational safety.

[0025] Step 5: System Optimization and On-site Debugging

[0026] Based on the issues identified during joint debugging and testing, system optimization was implemented: The parameters of the visual recognition model were optimized to improve recognition accuracy and reduce false and missed recognitions; the algorithm of the scheme planning module was adjusted to optimize the cutting scheme and branch cutting sequence, further preventing branches from being planted or breaking feeder lines; the aiming accuracy of the automatic gimbal was calibrated to ensure precise laser cutting; and the mobility of the tracked chassis was optimized to improve adaptability to complex terrain. The robot was transported to an actual power line clearing operation site for on-site debugging: adapting to the site terrain, tree canopy morphology, and feeder line layout, calibrating cutting parameters, branch cutting sequence, and aiming accuracy; testing the stability of wireless communication to ensure reliable remote control and data transmission; and simulating actual clearing operations to verify the robot's fully automated operation capabilities and clearing effects, ensuring it meets actual operational requirements.

[0027] Step 6: System Operation and Maintenance

[0028] During obstacle removal operations, staff send work instructions via a remote control terminal, and the robot moves autonomously to the work area via its tracked chassis. The AI ​​vision recognition module scans the tree canopy with an optical lens, acquiring real-time images of the canopy, identifying branches encroaching on the feeder within a 1-meter safety radius, calculating and determining the cutting points for each branch, and transmitting the data to the scheme planning module. Based on the received data, the scheme planning module automatically formulates a cutting plan and a cutting sequence from high to low, prioritizing the cutting of branches closest to the feeder and posing the greatest threat. According to the cutting plan, the aiming control module controls the automatic gimbal to move the optical lens and laser emitter to aim at the first cutting point. The laser cutting component is activated, outputting a laser beam of appropriate power to complete the branch cutting. After cutting, the automatic gimbal moves the laser emitter dynamically to the next cutting point, repeating the cutting process until all branches threatening the feeder's safety are cut. During the operation, the wireless communication module transmits cutting images, work progress, and other data to the remote control terminal in real time, allowing staff to monitor the operation in real time and remotely stop the operation if any abnormalities occur. After the operation is completed, the staff cleans the robot to remove the debris generated during cutting; checks the operating status of each component and replaces any worn parts; and charges the lithium battery pack to ensure that the robot can be put into normal use next time.

[0029] The beneficial effects of this invention are:

[0030] 1. Flexible mobility and adaptable to complex terrain: Adopting a tracked self-propelled mobile structure and equipped with shock absorption modules, it can move smoothly on hillsides, gullies, and irregular sites around power transmission lines without the need for manual assistance, greatly improving the flexibility and feasibility of obstacle removal operations and expanding the scope of obstacle removal operations.

[0031] 2. AI-powered visual dynamic aiming with high accuracy: Through the collaborative work of the AI ​​visual recognition module and the automatic gimbal, it can accurately identify tree crowns, branches and feeders, accurately determine the area where branches have intruded into the feeder within a 1-meter safety radius, calculate and determine the optimal cutting point, and has a dynamic tracking function that can adjust the cutting point in time according to the swaying of branches to ensure accurate aiming and avoid miscutting or missing cuts.

[0032] 3. Fully automated operation with high efficiency: The scheme planning module can automatically formulate cutting schemes and cutting sequences based on AI visual recognition data without manual intervention. At the same time, the automatic control module can control the robot's movement, aiming, cutting, and cutting point changes, realizing the full automation of obstacle clearing operations, greatly improving work efficiency and reducing the labor intensity of workers.

[0033] 4. Reasonable and safe cutting sequence: The cutting sequence strictly follows the principle of cutting from high to low, and the cutting sequence is optimized in combination with the spatial position of the branches to ensure that the cut branches do not get stuck and the falling trajectory avoids the feeder line, thus avoiding crushing the feeder line. At the same time, the laser protection module can prevent laser leakage and debris splashing, further improving the safety of the operation and ensuring the safe and stable operation of the transmission line.

[0034] 5. Remote controllability and easy monitoring: Equipped with wireless communication function, it supports remote control and real-time transmission of work data. Staff can remotely monitor the work process from a safe area and intervene in time when abnormalities occur, avoiding safety hazards caused by high-altitude and close-range operations and improving the safety of operations. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the overall structure of the laser obstacle removal robot of the present invention;

[0036] Figure 2 This is a block diagram showing the module connection of the automatic control component of the present invention;

[0037] Figure 3 This is a flowchart illustrating the workflow of the obstacle removal operation of the present invention.

[0038] Figure 4 This is a flowchart illustrating the recognition and cutting point calculation process of the AI ​​visual recognition module of the present invention.

[0039] Figure 1 Markers: 1-High-definition camera, 2-Robotic arm, 3-Power supply compartment, 4-Communication module, 5-Laser module. Detailed Implementation

[0040] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments to facilitate a clearer understanding of the present invention. However, the present invention is not limited to the following embodiments.

[0041] like Figure 1 This embodiment provides a laser obstacle removal robot device based on AI vision dynamic aiming, including a mobile carrier component, an AI vision aiming component, a laser cutting component, an automatic control component, and a power supply component.

[0042] The mobile carrier component adopts a self-moving structure; the drive motor is a DC servo motor with an output power of 1000W, which is connected to the tracked chassis through gear transmission, and the moving speed can be adjusted between 0.1-0.5m / s; the steering mechanism adopts a differential steering gear, which can realize the flexible steering of the tracked chassis; the shock absorption module adopts a spring shock absorption structure, which can reduce the vibration during the robot's movement and ensure the stability of AI vision aiming and laser cutting.

[0043] The AI ​​vision aiming component includes an optical lens, an automatic gimbal, and an AI vision recognition module. The optical lens uses a high-definition industrial camera with a resolution of 1080P and a frame rate of 30fps, capable of capturing high-definition images of the tree canopy in real time. The automatic gimbal is a servo-driven gimbal with a rotation angle range of 0°-360° horizontally and -90°-90° vertically, with a rotation accuracy of ±0.1°. The optical lens and laser emitter are fixedly mounted on the automatic gimbal and can be synchronized. The AI ​​vision recognition module uses an industrial control board with a built-in trained YOLOv8 vision recognition model. Trained and optimized with over 10,000 samples of tree canopies, branches, and feeders from different environments and tree species, it achieves a recognition accuracy of 99%. It can accurately distinguish between feeders and branches, calculate the distance between branches and feeders in real time, and automatically mark areas to be cut when the distance is less than 1 meter, then calculates and determines the optimal cutting point.

[0044] The laser cutting assembly includes a laser generator, a laser emitting head, a focusing module, and a laser protection module. The laser generator is a fiber laser generator with a power adjustment range of 1000W-5000W, capable of cutting tree branches with a diameter of 0.5-15cm. The laser emitting head is fixedly connected to an automatic pan-tilt head and synchronized with the optical lens. The focusing module uses a convex lens focusing structure, resulting in a laser spot diameter of 0.1mm after focusing, improving cutting efficiency. The laser protection module includes a protective shell and a beam shield. The protective shell is made of high-temperature resistant and flame-retardant material to prevent laser leakage and injury, while the beam shield prevents cutting debris from splashing onto the feed line surface.

[0045] The automatic control components include a scheme planning module, a motion control module, an aiming control module, and a wireless communication module. The scheme planning module uses a microcontroller connected to the AI ​​vision recognition module, enabling it to automatically formulate cutting schemes and the cutting sequence from high to low. The motion control module is connected to the drive motor and steering mechanism, controlling the movement, stopping, and steering of the tracked chassis. The aiming control module is connected to the automatic gimbal and laser generator, controlling the automatic gimbal's dynamic aiming and laser cutting. The wireless communication module is a dual-mode module with a transmission distance of at least 1km, supporting remote control and real-time transmission of work data.

[0046] The power supply components include a lithium battery pack and a charging module. The lithium battery pack uses high-capacity lithium iron phosphate batteries, providing a 10-hour battery life; the charging module is a fast-charging module, with a charging time of no more than 2 hours, and features overcharge, over-discharge, and short-circuit protection functions.

[0047] This embodiment also provides a method for constructing a laser obstacle removal robot based on AI visual dynamic aiming, including the following steps:

[0048] Step 1: Assembly of mobile carrier components

[0049] The tracked chassis is manufactured, and the drive motor and steering mechanism are assembled and connected to the tracked chassis. The transmission coordination is adjusted to ensure that the tracked chassis can move smoothly and steer flexibly. Shock absorption modules are installed to reduce movement vibration. Installation interfaces are reserved for fixing various components.

[0050] Step 2: Assembly of AI vision aiming component and laser cutting component

[0051] Fix the optical lens and laser emitter on the automatic gimbal and adjust the synchronization accuracy; connect the AI ​​visual recognition module to the optical lens, load the YOLOv8 visual recognition model, and adjust the recognition accuracy to ensure that the 1-meter safe distance and cutting point can be accurately determined; connect the laser generator, focusing module and laser emitter and adjust the laser power and focusing effect; install the laser protection module and adjust the protection performance.

[0052] Step 3: Deployment of Automated Control Components

[0053] The scheme planning module, motion control module, aiming control module, and wireless communication module are integrated and installed in the control box and fixed on the tracked chassis; signal connections between each module and other components are established, and the collaborative working performance is debugged to ensure that each module operates normally.

[0054] Step 4: Power supply component assembly and system commissioning

[0055] Install the lithium battery pack and charging module, complete the wiring connection, and test the power supply stability; start the system and conduct joint debugging in the simulated work area to verify the accuracy and reliability of functions such as AI recognition, cutting scheme, aiming and cutting, and movement, and optimize system parameters.

[0056] Step 5: System Optimization and On-site Debugging

[0057] The system was optimized based on the joint debugging issues to improve recognition accuracy, cutting rationality, and aiming accuracy; the robot was transported to the actual work site to adapt to the site environment, calibrate parameters, simulate obstacle clearing operations, and verify the fully automated operation capability and obstacle clearing effect.

[0058] Step 6: System Operation and Maintenance

[0059] During operation, the robot moves to the work area on its own. The AI ​​vision recognition module scans the tree canopy to identify the branches to be cut and the cutting points. The scheme planning module formulates the cutting scheme and the cutting sequence. The aiming control module controls the automatic gimbal to aim, and the laser cutting component starts cutting until the obstacle is cleared. After the operation, the robot is cleaned and inspected, and the lithium battery pack is charged.

[0060] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A laser obstacle clearing robot device based on AI vision dynamic aiming, characterized in that, This includes mobile carrier components, AI vision aiming components, laser cutting components, automatic control components, and power supply components, among which: The mobile carrier component adopts a tracked self-propelled structure, including a tracked chassis, a drive motor, and a steering mechanism. The drive motor is connected to the tracked chassis, and the steering mechanism is used to control the steering of the tracked chassis. It can achieve smooth movement in complex terrain, including hillsides, gullies, and irregular sites around power transmission lines, and drive the entire robot to the obstacle clearing operation area. The AI ​​vision aiming component includes an optical lens, an automatic gimbal, and an AI vision recognition module. The optical lens is fixedly connected to the automatic gimbal, which can drive the optical lens to rotate at multiple angles in the horizontal and vertical directions, completing a full-range scan of the tree canopy and dynamic aiming at the target branch. The AI ​​vision recognition module has a built-in trained vision recognition model, which can scan and recognize the outline of the tree canopy through the tree canopy image captured by the optical lens, accurately determine the area where the branch has invaded the feeder's safe distance radius of 1 meter, and calculate and determine the optimal cutting point for the branch. The laser cutting assembly includes a laser generator, a laser emitting head, and a focusing module. The laser emitting head is fixedly connected to an automatic gimbal and is synchronously linked with an optical lens. It can adjust the emission angle following the rotation of the automatic gimbal to achieve precise aiming at the cutting point. The focusing module is used to focus the laser beam to improve cutting efficiency. The laser generator can adjust the laser power according to the cutting requirements. The automatic control component includes a scheme planning module, a motion control module, and an aiming control module. The scheme planning module is signal-connected to the AI ​​vision recognition module and can automatically formulate a cutting scheme and cutting sequence based on the tree canopy image and information on branches intruding into the safe area obtained by the AI ​​vision recognition module. The cutting sequence follows the principle of cutting from high to low to ensure that the cut branches do not interfere with or damage the feeder. The motion control module is signal-connected to the mobile carrier component and controls the movement, stopping, and turning of the tracked chassis. The aiming control module is signal-connected to the automatic gimbal and the AI ​​vision recognition module and controls the automatic gimbal to drive the optical lens and laser emitter to aim at the target cutting point and dynamically change the cutting point according to the cutting progress. The power supply components include a lithium battery pack and a charging module, which provide stable power to all components of the entire robot device and are suitable for outdoor operating environments without external power sources.

2. The apparatus of claim 1, wherein The visual recognition model in the AI ​​visual recognition module adopts the YOLOv8 object detection model based on deep learning. It has been trained and optimized with a large number of tree canopy, branch, and feeder samples. It can accurately distinguish between feeders and branches with a recognition accuracy of no less than 98%. It can also output the distance data between branches and feeders in real time. When the distance is less than 1 meter, it is automatically marked as the area to be cut.

3. The apparatus of claim 1, wherein The automatic gimbal is a servo-driven gimbal with a rotation angle range of 0°-360° horizontally and -90°-90° vertically, and a rotation accuracy of ±0.1°. It can achieve precise positioning and dynamic tracking of the cutting point, ensuring that the laser emitter and the cutting point are always aligned.

4. The apparatus of claim 1, wherein, The cutting scheme formulated by the scheme planning module, in addition to following the cutting order from high to low, also takes into account the thickness, growth, distance from the feeder and spatial position of the branches, giving priority to cutting the branches that are closest to the feeder and pose the greatest threat to the feeder, while avoiding the feeder in the trajectory of falling branches, further avoiding the situation of crushing the feeder or branches being stuck.

5. The apparatus of claim 1, wherein, The laser cutting assembly also includes a laser protection module, which includes a protective shell and a beam shield to prevent laser leakage from injuring people and to prevent debris generated during cutting from splashing onto the feeder surface and causing damage.

6. The apparatus of claim 1, wherein, The automatic control component also includes a wireless communication module, which can receive remote control commands and transmit operation data (cutting image, cutting position, operation progress) in real time, and supports remote start and stop of the robot, adjustment of cutting parameters and modification of cutting scheme.

7. A laser obstacle clearing robot construction method based on AI vision dynamic aiming, characterized in that, Includes the following steps: Step 1: Assemble the mobile carrier components. Connect the drive motor, steering mechanism and tracked chassis. Debug the transmission coordination between the drive motor and tracked chassis to ensure that the tracked chassis can move smoothly and turn flexibly, adapting to the complex terrain around the power transmission line. Reserve installation interfaces on the tracked chassis for fixing the AI ​​vision aiming component, laser cutting component and power supply component. Step 2: Assemble the AI ​​vision aiming component and laser cutting component. Fix the optical lens and laser emitter onto the automatic gimbal, and adjust the rotation accuracy of the automatic gimbal to ensure that the optical lens and laser emitter are synchronized and accurately aimed. Connect the AI ​​vision recognition module to the optical lens signal to complete the loading and debugging of the vision recognition model, ensuring that the model can accurately identify the tree crown, branches, and feed line, accurately determine the area where branches have invaded the feed line within a 1-meter safety radius, and calculate the cutting point. Connect the laser generator, focusing module, and laser emitter, and adjust the laser power and focusing effect to ensure cutting efficiency and cutting quality. Step 3: Deploy the automatic control components. Integrate the scheme planning module, motion control module, aiming control module, and wireless communication module into the control box and establish signal connections with the mobile carrier component, AI vision aiming component, and laser cutting component. Debug the collaborative performance of each module to ensure that the motion control module can accurately control the movement of the tracked chassis, the aiming control module can control the automatic gimbal to dynamically aim, and the scheme planning module can automatically generate cutting schemes and cutting sequences based on AI vision recognition data. Step 4: Power supply component assembly and system integration testing. Install the lithium battery pack and charging module in the reserved positions on the tracked chassis, complete the wiring connections with each power component, and test the power supply stability to ensure that the lithium battery pack's endurance meets the needs of field operations. Start the entire robot system and conduct integration testing. Control the tracked chassis to move to the simulated work area, and use the AI ​​vision recognition module to scan the simulated tree canopy to verify the accuracy of safe area intrusion judgment and cutting point calculation. Generate a cutting plan through the scheme planning module to verify the rationality of the branch cutting sequence (from high to low) and ensure that the cut branches do not pierce or break the simulated feeder. Test the dynamic aiming and cutting point changing functions of the automatic gimbal to ensure that laser cutting is precise and efficient. Step 5: System optimization and on-site debugging. Based on the problems found in the joint debugging test, optimize the recognition accuracy of the visual recognition model, the rationality of the cutting scheme of the scheme planning module, and the aiming accuracy of the automatic gimbal; transport the robot to the actual power transmission line obstacle clearing operation site for on-site debugging, adapt to the site terrain, tree crown shape and feeder layout, and further calibrate the cutting parameters, cutting sequence and aiming accuracy to ensure that the robot can stably complete the fully automatic laser obstacle clearing operation; Step 6: System Operation and Maintenance. During obstacle removal, the robot moves autonomously to the work area via its tracked chassis. The AI ​​vision recognition module scans the tree canopy through an optical lens, identifies branches encroaching on the feeder within a 1-meter safety radius, and calculates and determines the cutting point. The scheme planning module automatically formulates a cutting plan and the cutting sequence from high to low based on the recognition data. The aiming control module controls the automatic gimbal to drive the optical lens and laser emitter to aim at the cutting point. The laser cutting component starts to complete the cutting, automatically changing the cutting point during the cutting process until all branches threatening the safety of the feeder are cut off. After the operation is completed, all components are cleaned and inspected, and the lithium battery pack is charged to ensure that the robot can be put into normal use next time.

8. The method of claim 7, wherein, In step 2, the training process of the visual recognition model includes: collecting tree canopy, branch, and feeder samples from different environments (sunny, cloudy, and rainy days), different tree species, and different growth states; labeling the samples (labeling the outline of the branch, the position of the feeder, and the distance between the branch and the feeder); dividing the labeled samples into training, validation, and test sets; inputting them into the YOLOv8 model for training; optimizing the recognition accuracy by adjusting the model parameters until the recognition accuracy of the model on the test set is not less than 98% before loading it into the AI ​​visual recognition module.

9. The method of claim 7, wherein, Step 4 of the joint debugging test also includes a safety test. The protection performance of the laser protection module is tested to ensure that the laser does not leak. The rationality of the cutting sequence is verified by simulating tree branch cutting to ensure that the falling trajectory of the cut tree branch avoids the feeder and that there is no situation where the feeder is blocked or broken.

10. The method of claim 7, wherein, Step 5, on-site debugging also includes wireless communication testing to ensure that remote control commands can be transmitted accurately, work data can be fed back in real time, and remote emergency intervention can be supported. When an abnormal situation occurs, the robot operation can be stopped remotely.

Images