Aircraft laser cladding in-situ repair method and system based on mobile robot

Abstract

The invention discloses an aircraft laser cladding in-situ repair scheme based on a mobile robot, and belongs to the technical field of metal plating, and the aircraft laser cladding in-situ repair scheme comprises the steps that the mobile robot determines the position of a damaged part of a metal structure and damage form information; the upper computer plans a regularization processing path of the damaged part; the movable robot controls a mechanical polishing module arranged on a mechanical arm to regularize the damaged part according to the welding path, and a filling restoration model corresponding to the regularized damaged part is determined through interaction of a machine vision module and an upper computer; the upper computer plans a cladding path and cladding parameters according to the filling restoration model, the material of the damaged part and the current posture information of the mechanical arm of the movable robot; the movable robot controls a laser generator of the laser cladding module to perform cladding operation on a damaged part and calls the molten pool monitoring module to monitor a molten pool picture, molten pool features are extracted based on the molten pool picture, a cladding path is adjusted, in-situ repairing can be achieved, efficiency is high, and positioning is accurate.

Description

Technical Field

[0001] This invention relates to the field of metal coating technology, and in particular to a method and system for in-situ repair of aircraft by laser cladding based on a mobile robot. Background Technology

[0002] During various flight missions, aircraft equipment loads increase, leading to more stress corrosion cracks and wear on metal structures. Furthermore, external impacts can easily cause damage such as holes or defects in the aircraft's metal structure. Traditionally, damage from cracks, wear, or holes is repaired by replacing parts. However, this method is time-consuming and puts significant pressure on spare parts availability. To quickly restore damaged aircraft, in-situ repair is necessary—that is, rapid repair of the damaged parts at the site of the damage.

[0003] With the development of laser cladding technology, it has been gradually applied to equipment remanufacturing, but it still faces two problems: First, existing laser cladding equipment is mainly installed on gantry frames or fixed in workshops, and its poor mobility cannot meet the actual emergency repair requirements of aircraft structural damage modes that are not uniform and the damage locations are highly random; Second, existing laser cladding technology mainly uses the "teach first, then clad" method, which is complex to operate and has low repair efficiency in the face of the randomness of the damage locations, greatly limiting its application in aircraft structural emergency repair.

[0004] It is evident that existing laser cladding solutions for metal structure damage have two limitations: firstly, they cannot achieve in-situ repair; secondly, they cannot achieve precise positioning and repair of multiple damage modes, and their repair efficiency is low. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for in-situ repair of aircraft by laser cladding based on a mobile robot, which can solve at least one of the many problems mentioned above in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a mobile robot-based in-situ repair method for aircraft laser cladding. The method is applied to an aircraft laser cladding in-situ repair system, which includes a mobile robot integrating a laser cladding module, a molten pool monitoring module, a mechanical grinding module, and a machine vision module, and a host computer. The method includes:

[0008] The mobile robot interacts with the host computer through a machine vision module to determine the location and morphology of damage to the metal structure.

[0009] The host computer plans a regularized processing path for the damaged area based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot arm.

[0010] The mobile robot controls the mechanical grinding module on the robotic arm to perform regular processing on the damaged area according to the regularized processing path of the damaged area.

[0011] The mobile robot interacts with the host computer through a machine vision module to determine the filling repair model corresponding to the damaged area after regularization.

[0012] The host computer plans the cladding path and cladding parameters based on the filled repair model, the material of the damaged area, and the current posture information of the mobile robotic arm.

[0013] The mobile robot controls the laser generator of the laser cladding module to perform cladding operations on the damaged area according to the cladding path, the cladding parameters, and the welding process parameters. During the cladding operation, the robot calls the molten pool monitoring module to monitor the molten pool image, extracts molten pool features based on the molten pool image, and adjusts the cladding path based on the molten pool features. The laser generator of the laser cladding module, the mechanical grinding module, and the molten pool monitoring module are all located at the front end of the robotic arm.

[0014] Optionally, the laser cladding module includes a laser unit and an auxiliary unit. The laser unit includes a laser generator, and the auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler.

[0015] Optionally, the cooler is a dual-temperature dual-pump series water chiller, and the cooler is equipped with an intelligent temperature control module to control the low-temperature end of the factory water temperature to be 25°C.

[0016] The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

[0017] Optionally, the molten pool monitoring module includes: a high-speed camera, a light shield, and an image acquisition card; the high-speed camera is set next to the test bench of the laser cladding module, and the center of the high-speed camera lens is aligned with the cladding channel to acquire the molten pool image in real time during the cladding operation;

[0018] The light-shielding sheet is placed in front of the high-speed camera lens;

[0019] The image acquisition card is used to convert the electrical signals obtained from the image information acquired by the high-speed camera into digital signals for recognition by the host computer.

[0020] Optionally, the step of the mobile robot interacting with the host computer through a machine vision module to determine the location and morphology of damage to the metal structure includes:

[0021] The machine vision module captures and acquires image information, and sends the image information to the host computer; wherein, the machine vision module is disposed at the front end of the robotic arm;

[0022] The host computer locates the damaged area of ​​the aircraft's metal structure based on the image information and feeds back the damaged area information to the mobile robot; wherein, the image information includes the morphological information of the damaged area;

[0023] The mobile robot invokes the machine vision module to scan the damaged area with a laser, obtains the first point cloud information of the damaged area, and sends the point cloud information to the host computer.

[0024] Optionally, the step of the host computer planning a regularized processing path for the damaged area based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot arm includes:

[0025] The host computer constructs a first three-dimensional model corresponding to the damaged area based on the first point cloud information.

[0026] The host computer generates a regularized processing path for the damaged area based on the first 3D model, the posture information of the mobile robot arm, and the damage morphology information of the damaged area, and sends the regularized processing path for the damaged area to the mobile robot.

[0027] Optionally, the step of the host computer constructing a first three-dimensional model corresponding to the damaged area based on the first point cloud information includes:

[0028] The host computer preprocesses the first point cloud information to filter out invalid and outlier points, and obtains the second point cloud information.

[0029] The second point cloud information is filtered out using a voxel grid filtering algorithm to obtain the third point cloud information;

[0030] A preset point cloud reconstruction algorithm is used to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model.

[0031] Optionally, the step of the host computer preprocessing the first point cloud information to filter out invalid and outlier points to obtain the second point cloud information includes:

[0032] For each point in the first point cloud information, determine whether the coordinates of the point in the first dimension are within a preset first dimension value range;

[0033] If not, the point is considered invalid and filtered out.

[0034] Traverse each point in the first point cloud information and calculate the average distance between the point and each adjacent point; determine whether the average distance is within a preset distance range threshold.

[0035] If not, identify the point and each point adjacent to the point as outliers and filter out all outliers.

[0036] The first point cloud information after filtering out invalid and discrete points is determined as the second point cloud information.

[0037] Optionally, the step of using a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model includes:

[0038] Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models;

[0039] The model with the highest score among the multiple 3D models is selected as the first 3D model.

[0040] This invention also provides an in-situ repair system for aircraft laser cladding based on a mobile robot. The system includes a mobile robot and a host computer. The mobile robot includes a control cabinet, a machine vision module, a laser cladding module, a molten pool monitoring module, and a mechanical grinding module.

[0041] The control cabinet is used to call the machine vision module to interact with the host computer to determine the location and damage morphology of the damaged parts of the metal structure.

[0042] The host computer is used to plan a regularized processing path for the damaged area based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot arm.

[0043] The control cabinet is also used to control the mechanical grinding module on the robotic arm to perform regular processing on the damaged area according to the regularized processing path of the damaged area.

[0044] The control cabinet is also used to call the machine vision module to interact with the host computer and determine the filling repair model corresponding to the damaged area after regularization.

[0045] The host computer is also used to plan the cladding path and cladding parameters based on the filled repair model, the material of the damaged area, and the current posture information of the mobile robotic arm.

[0046] The control cabinet is also used to control the laser generator of the laser cladding module to perform cladding operation on the damaged area according to the cladding path, the cladding parameters and the welding process parameters, and to call the molten pool monitoring module to monitor the molten pool image during the cladding operation, extract molten pool features based on the molten pool image, and adjust the cladding path based on the molten pool features; wherein, the laser generator of the laser cladding module, the mechanical grinding module and the molten pool monitoring module are all located at the front end of the robotic arm.

[0047] Optionally, the laser cladding module includes a laser unit and an auxiliary unit. The laser unit includes a laser generator, and the auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler.

[0048] Optionally, the cooler is a dual-temperature dual-pump series water chiller, and the cooler is equipped with an intelligent temperature control module to control the low-temperature end of the factory water temperature to be 25°C.

[0049] The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

[0050] Optionally, the molten pool monitoring module includes: a high-speed camera, a light shield, and an image acquisition card; the high-speed camera is set next to the test bench of the laser cladding module, and the center of the high-speed camera lens is aligned with the cladding channel to acquire the molten pool image in real time during the cladding operation;

[0051] The light-shielding sheet is placed in front of the high-speed camera lens;

[0052] The image acquisition card is used to convert the electrical signals obtained from the image information acquired by the high-speed camera into digital signals for recognition by the host computer.

[0053] Optionally, the control cabinet includes:

[0054] The first invocation module is used to invoke the machine vision module to capture and acquire image information, and send the image information to the host computer; wherein, the machine vision module is disposed at the front end of the robotic arm;

[0055] The host computer includes:

[0056] The positioning module is used to locate the damaged part of the aircraft's metal structure based on the image information and feed the damaged part information back to the control cabinet; wherein, the image information includes the morphological information of the damaged part;

[0057] The control cabinet also includes:

[0058] The second calling module is used to call the machine vision module to obtain the first point cloud information of the damaged area by laser scanning, and send the point cloud information to the host computer.

[0059] Optionally, the host computer further includes:

[0060] The construction module is used to construct a first three-dimensional model corresponding to the damaged area based on the first point cloud information;

[0061] The path planning module is used to generate a regularized processing path for the damaged part based on the first three-dimensional model, the posture information of the mobile robot arm, and the damage morphology information of the damaged part, and send the regularized processing path for the damaged part to the mobile robot.

[0062] Optionally, the building module is specifically used for:

[0063] The first point cloud information is preprocessed to filter out invalid and outlier points, resulting in the second point cloud information.

[0064] The second point cloud information is filtered out using a voxel grid filtering algorithm to obtain the third point cloud information;

[0065] A preset point cloud reconstruction algorithm is used to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model.

[0066] Optionally, when the construction module preprocesses the first point cloud information to filter out invalid and outlier points to obtain the second point cloud information, it is specifically used for:

[0067] For each point in the first point cloud information, determine whether the coordinates of the point in the first dimension are within a preset first dimension value range;

[0068] If not, the point is considered invalid and filtered out.

[0069] Traverse each point in the first point cloud information and calculate the average distance between the point and each adjacent point; determine whether the average distance is within a preset distance range threshold.

[0070] If not, identify the point and each point adjacent to the point as outliers and filter out all outliers.

[0071] The first point cloud information after filtering out invalid and discrete points is determined as the second point cloud information.

[0072] Optionally, when the construction module uses a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model, it is specifically used for:

[0073] Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models;

[0074] The model with the highest score among the multiple 3D models is selected as the first 3D model.

[0075] This application discloses a mobile robot-based in-situ laser cladding repair scheme for aircraft. The mobile robot interacts with a host computer via a machine vision module to determine the location and morphology of damage to the metal structure. The host computer plans a standardized processing path for the damaged area based on the location, morphology, and current posture of the robot's arm. Following this path, the mobile robot controls a mechanical grinding module on its arm to perform standardized processing on the damaged area. The robot then interacts with the host computer again via the machine vision module to determine the corresponding filler repair model. Based on the filler repair model, the material of the damaged area, and the current posture of the robot's arm, the host computer plans the cladding path and parameters. The mobile robot controls the laser generator of the laser cladding module to perform cladding operations on the damaged area according to the cladding path, parameters, and welding process parameters. During the cladding process, a molten pool monitoring module is invoked to monitor the molten pool, extracting features and adjusting the cladding path accordingly. This mobile robot-based in-situ laser cladding repair solution for aircraft offers several advantages. First, it enables in-situ repair of aircraft metal structures by directly locating, regularizing, and laser cladding damaged areas using a mobile robot. Second, it eliminates the need for a "teach-then-clad" approach, thus improving cladding efficiency, shortening the repair cycle, and reducing the difficulty of repair. Third, by combining machine vision modules with host computer modeling to determine the damage location, it can adapt to various damage patterns and accurately locate damage. Attached Figure Description

[0076] Figure 1 This is a flowchart illustrating the steps of an in-situ repair method for aircraft laser cladding based on a mobile robot, according to an embodiment of this application.

[0077] Figure 2 This is a schematic diagram illustrating the structure of an in-situ repair system for aircraft laser cladding based on a mobile robot, according to an embodiment of this application.

[0078] Figure 3 This is a schematic diagram illustrating a mobile repair platform according to an embodiment of this application;

[0079] Figure 4 This is a schematic diagram illustrating the principle of surface point cloud slicing processing in an embodiment of this application;

[0080] Figure 5 This is a schematic diagram illustrating the voxel grid filtering principle of an embodiment of this application;

[0081] Figure 6 This is a schematic diagram illustrating the three-dimensional reconstruction effect of the surface damage point cloud of a 316L steel plate according to an embodiment of this application.

[0082] Figure 7 This is a structural block diagram illustrating an in-situ repair system for aircraft laser cladding based on a mobile robot, according to an embodiment of this application. Detailed Implementation

[0083] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0084] This application proposes a mobile robot-based in-situ laser cladding repair scheme for aircraft. It utilizes a mobile robot integrating a laser cladding module, a molten pool monitoring module, a mechanical grinding module, and a machine vision module, interacting with a host computer to locate, regularize, and perform in-situ laser cladding repair of damaged areas on the aircraft's metal structure. During operation, the mobile robot freely moves to the vicinity of the damaged aircraft metal structure. The mechanical grinding module, driven by a robotic arm and based on the laser principle, regularizes the damaged area. The cladding module, driven by the robotic arm, performs cladding on the damaged area, and the molten pool monitoring module monitors the cladding progress during the process, adjusting the cladding path in real time based on the monitoring results. This achieves in-situ laser cladding repair of damaged areas on the aircraft's metal structure. This solution overcomes the inefficient traditional laser cladding repair process, which relies excessively on manual labor and involves a "teach-before-clad" approach, achieving automated and intelligent operation of the repair process without programming or teaching. Based on the "eye in hand" vision system and the external laser molten pool monitoring module, the path adaptive adjustment during the cladding process is realized through line laser scanning to reduce the impact of thermal deformation during the cladding process; through molten pool monitoring during the laser cladding process, process parameters are adjusted in real time to ensure cladding quality.

[0085] The following description, in conjunction with the accompanying drawings, details the in-situ repair solution for aircraft laser cladding based on mobile robots provided in this application through specific embodiments and application scenarios.

[0086] As attached Figure 1As shown, the in-situ repair method for aircraft laser cladding based on a mobile robot according to an embodiment of this application includes the following steps:

[0087] Step 10 1: The mobile robot interacts with the host computer through the machine vision module to determine the location and morphology of the damage to the metal structure.

[0088] The in-situ repair method for aircraft laser cladding based on mobile robots provided in this embodiment is executed by an in-situ repair system for aircraft metal structure laser cladding plating. This system includes a mobile robot and a host computer that integrates a laser cladding module, a molten pool monitoring module, a mechanical grinding module, and a machine vision module.

[0089] Figure 2 This is a schematic diagram illustrating the structure of an in-situ repair system for aircraft laser cladding based on a mobile robot, according to an embodiment of this application. Figure 2 As shown, this system is based on "robot + machine vision" technology. It integrates a robot, laser cladding module, machine vision module, melt pool monitoring module, and mechanical grinding module onto a mobile mechanism (such as a mobile repair platform), constructing the mobile robot shown in this embodiment. The mobile robot interacts with a host computer to perform in-situ laser cladding repair of aircraft metal structures. The robotic arm of the mobile robot is equipped with the laser cladding module, machine vision module, melt pool monitoring module, and mechanical grinding module, which act as the human's eyes and hands. The host computer is equipped with a vision software system that can process images and point clouds, build 3D models, and perform damage localization and cladding path planning for aircraft metal structures. The control cabinet is the "brain" of the mobile robot, used to control the movement of the mobile repair platform, the movement of the robotic arm, and the operation of the laser cladding module, melt pool monitoring module, mechanical grinding module, and machine vision module.

[0090] This application embodiment is based on "visual robot" technology, integrating a robotic arm, laser, cladding head, visual sensor, molten pool monitoring system, and control unit (i.e., the control cabinet mentioned below) onto the AGV chassis, designing a mobile multi-degree-of-freedom laser additive repair platform with multiple repair modes working in tandem. For example... Figure 3 As shown, the mobile repair platform can be a wheeled platform with a maximum speed of 10 km / h, possessing a certain obstacle-crossing capability. It can be remotely controlled to ensure a flexible and rapid response to actual damage and repair requirements. The repair platform mainly comprises three aspects: hardware system, control loop, and software system design.

[0091] The mobile robot system consists of an ABB IRB4600 six-DOF industrial robot, an IRC5 control system, and a FlexPendant teach pendant. Its working range is 2.05m, enabling a wide range of motion to meet the demands of complex on-site repair environments. The effective arm load capacity is 30kg, ensuring the strength requirements of each module mounted at the end of the robotic arm. For convenient and rapid on-site repair, the vehicle body features a steering wheel structure with six independent drive axes, remote control and RFID guidance, and can move in any lateral and diagonal direction, with a braking and parking function. Furthermore, to meet the repair needs of different structural parts of the aircraft, the vehicle body should have a lifting function with a lifting height of 2m to meet the emergency repair needs of the aircraft's upper surface, upper tail, and other parts.

[0092] The machine vision module, based on the characteristics of vision sensors, combines structured light sensors and line laser sensors and mounts them on the front end of the robotic arm to construct an eye-in-hand system, enabling data scanning and acquisition. This module mainly includes: a structured light sensor, a line laser sensor, and a flange.

[0093] Structured light sensors project structured light onto the surface of a workpiece, and a camera acquires the image information to detect surface information. They feature a large scanning range and high scanning speed. By acquiring images of damaged areas using structured light sensors, rapid damage localization can be achieved. An exemplary model of structured light sensor is the SA-T 1000, with a near-field FOV of 700×600mm and a far-field FOV of 4000×3000mm, enabling rapid damage localization over a wide viewing angle. It also communicates with a host computer via gigabit Ethernet to ensure real-time data transmission.

[0094] In one exemplary structure, the laser cladding module may include a laser unit and an auxiliary unit. The laser unit includes a laser generator; the auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler. The cooler is a dual-temperature, dual-pump water-cooled unit, and an intelligent temperature control module is installed inside the cooler to control the low-temperature end of the factory water temperature to 25°C. The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

[0095] Lasers output high-energy-density laser beams to melt metal alloy powders, and their performance directly affects the laser cladding effect. In practical applications, to ensure cladding quality, fiber lasers are used, offering both continuous and modulated modes, with a laser power of 6 kW. During laser cladding, the laser continuously outputs high power, converting most of the electrical energy into heat. The internal laser module generates a significant amount of heat, and excessively high temperatures can damage the laser and optical components. Based on the aforementioned laser cooling requirements, a dual-temperature, dual-pump water chiller is selected. The factory-set low-temperature end water temperature is 25°C, and an internal intelligent temperature control module enables precise temperature control, provides excellent anti-interference capabilities, and supports Modbus-RTU communication with a maximum transmission distance of at least 1000 meters. The powder feeding rate plays a decisive role in the dilution rate of the cladding channel. In order to ensure the cladding quality, a scraper-type powder feeder is preferred, with a flow rate range of 1L / min-20L / min, a pressure range of -0.1Mpa-1Mpa, and a powder particle size of 90 mesh-300 mesh.

[0096] In one exemplary structure, the molten pool monitoring module may include: a high-speed camera, a light-shielding plate, and an image acquisition card; the high-speed camera is set next to the test bench of the laser cladding module, with the center of the high-speed camera lens aligned with the cladding channel, so as to acquire the molten pool image in real time during the cladding operation; the light-shielding plate is laid in front of the high-speed camera lens; the image acquisition card is used to convert the electrical signal after the image information acquired by the high-speed camera into a digital signal for recognition by the host computer.

[0097] In laser additive repair, every moment of the melting and solidification of the molten pool is crucial to the final repair quality and effect. Therefore, a molten pool monitoring module is needed to monitor the cladding process and the state of the molten pool. The main purposes are: first, monitoring during process parameter adjustments; and second, data acquisition during the repair process for easy playback and analysis of problems. The high brightness and instantaneous nature of the laser cladding process pose challenges to molten pool monitoring. To address this, a high-speed camera is used to capture images of the molten pool, obtaining a two-dimensional optical image. The molten pool is monitored and photographed online, and machine vision technology is used to process the images to obtain the molten pool edge and parameters. By comparing parameter changes over time, alarms or shutdowns are triggered promptly in case of abnormalities.

[0098] Since laser cladding visual monitoring is conducted in complex lighting environments, high-speed cameras, due to their high-speed shooting capabilities, are suitable for capturing fast-moving or instantaneous events. They feature high resolution, high frame rate, and real-time viewing and analysis capabilities, making them ideal for real-time monitoring of the molten pool during the cladding process. Based on the selection of camera parameters such as frame rate, resolution, target size, and wavelength range, a CCD camera was chosen and externally mounted next to the laser cladding test bench, with the camera center aligned with a single cladding channel to achieve real-time image acquisition of the cladding process.

[0099] Laser cladding is a high-energy thermoforming technology with a molten pool temperature reaching approximately 1600℃, resulting in strong thermal radiation. Furthermore, the fiber laser can cause overexposure during high-speed photography. Therefore, filters are needed to block the cladding laser light. However, the addition of filters results in a darker image and insufficient resolution. Therefore, a low-wavelength laser illumination device with sufficient power is also required to ensure the high-speed camera can clearly capture the surface morphology of the molten pool. During the cladding acquisition process, the image information is converted into electrical signals by the CCD camera, and then converted into digital signals by an image acquisition card before being recognized by the host computer. To ensure the smooth operation of the cladding process, an image acquisition card is used to achieve dynamic image capture.

[0100] In actual implementation, the machine vision module, laser cladding module, melt pool monitoring module, and mechanical grinding module can all be installed at the front end of the robotic arm of the mobile robot.

[0101] In one optional embodiment, the mobile robot interacts with the host computer through a machine vision module to determine the location and morphology of damage to the metal structure, which may include the following sub-steps:

[0102] Sub-step 1: The mobile robot captures image information through the machine vision module and sends the image information to the host computer.

[0103] Sub-step 2: The host computer locates the damaged parts of the aircraft's metal structure based on the image information and feeds back the information of the damaged parts to the mobile robot.

[0104] The image information includes morphological information of the damaged area.

[0105] Sub-step 3: The mobile robot calls the machine vision module to scan the damaged area with a laser, obtain the first point cloud information of the damaged area, and send the point cloud information to the host computer.

[0106] Line laser sensors utilize the high directionality and brightness of laser light to measure the three-dimensional information of an object by transmitting a single line of laser light. Composed of a laser emitter and a camera, they feature simple structure, high precision, and easy image processing. Based on these characteristics, after the structured light sensor has located the damage site, this project uses a line laser sensor to scan the damaged area and capture real-time images of the line laser's state. The changes in the laser stripe information are then used to obtain the three-dimensional point cloud information of the damaged area.

[0107] Step 102: The host computer plans a regularized processing path for the damaged area based on the location information, morphological information, material of the damaged area, and current posture information of the mobile robot arm.

[0108] In one optional embodiment, planning a regularized processing path for the damaged area may include the following sub-steps:

[0109] Sub-step 1021: The host computer constructs the first three-dimensional model corresponding to the damaged area based on the first point cloud information.

[0110] The vision software system consists of a camera acquisition module, a system calibration module, a point cloud processing module, and auxiliary software. Through vision algorithms, it can automatically identify holes and damage, automatically plan paths for regularized processing of damaged areas, plan paths for laser cladding, and perform automatic programming. The camera acquisition module is used to acquire image information from the camera; the system calibration module is used to unify the camera's image coordinates to the robot's base coordinates; and the point cloud processing module is used to analyze the three-dimensional point cloud information, i.e., the first point cloud information, acquired by the line laser sensor.

[0111] The vision software system runs on a host computer. During operation, the system first performs image capture, color setting, and image saving via the camera. Then, it completes system calibration through camera calibration, light plane calibration, and hand-eye calibration, unifying the camera's image coordinates to the robot's base coordinates. This is a prerequisite for the robot to accurately measure damaged objects. The point cloud processing module includes point cloud filtering, point cloud visualization, and point cloud size and color parameter functions. Common point cloud data processing methods include point cloud data preprocessing and 3D point cloud reconstruction. Based on the data processed by the point cloud processing module, mesh processing systems such as Meshlab are used to edit, render, and filter the 3D triangular mesh of the point cloud file. Simultaneously, 3D modeling of the damaged area is achieved using reverse engineering software, generating layered slices to lay the foundation for generating cladding repair paths. A schematic diagram of the curved surface point cloud slicing processing principle is shown below. Figure 4 As shown.

[0112] An optional method for a host computer to construct a first 3D model corresponding to the damaged area based on the first point cloud information may include the following sub-steps:

[0113] Sub-step 1: The host computer preprocesses the first point cloud information to filter out invalid and outlier points, and obtains the second point cloud information;

[0114] An example of how a host computer can preprocess the first point cloud information to filter out invalid and outlier points and obtain the second point cloud information is as follows:

[0115] For each point in the first point cloud information, determine whether the point's coordinates in the first dimension are within the preset first dimension value range; if not, the point is considered invalid; if so, it is considered valid and invalid points are filtered out. Traverse each point in the first point cloud information and calculate the average distance between the point and each of its neighboring points; determine whether the average distance is within a preset distance range threshold; if not, identify the point and each of its neighboring points as outliers and filter out all outliers; the first point cloud information after filtering out invalid points and discrete points is identified as the second point cloud information.

[0116] It should be noted that the specific values ​​of the preset first dimension value range and the preset distance range threshold can be flexibly set by those skilled in the art, and no specific restrictions are imposed on them in this embodiment.

[0117] Sub-step 1 involves point cloud data preprocessing: Two methods, pass-through filtering and statistical filtering, can be used to filter out irrelevant and outlier points from the point cloud library. The pass-through filtering algorithm removes irrelevant points by specifying a dimension (X, Y, Z) and its value range, traversing each point in the point cloud, and determining whether the point is within the value range, removing points outside the range. The statistical filtering algorithm traverses each point in the point cloud, calculates the average distance to all neighboring points, and deletes outliers by comparing the average distance to a standard range threshold, retaining the filtered point cloud data. Specifically, outlier removal is achieved using statistical filters from the point cloud library.

[0118] Sub-step 2: Use a voxel grid filtering algorithm to filter out the second point cloud information to obtain the third point cloud information;

[0119] Sub-step 2 involves further simplification of the preprocessed point cloud data. A large number of redundant data points increase algorithm runtime and thus affect 3D reconstruction efficiency; therefore, further simplification of the preprocessed point cloud data is necessary. For example, a voxel grid filtering algorithm can be used. This algorithm divides the point cloud data into fixed-size voxel grids, calculates the centroid of all points within each grid, deletes grids with no point cloud data, and approximates all points within each grid as centroids. The set of all voxel centroids represents the voxel grid filtered point cloud data. This method preserves the shape characteristics of the point cloud data and retains its original geometric structure. A schematic diagram of the voxel grid filtering principle is attached. Figure 5 As shown.

[0120] Sub-step 3: Use a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model.

[0121] An optional method for using a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model can be as follows:

[0122] Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models; the model with the highest score is selected from the multiple 3D models and determined as the first 3D model.

[0123] The preset point cloud reconstruction algorithms may include, but are not limited to: Delaunay, Possion, and rolling ball method.

[0124] Commonly used point cloud 3D reconstruction algorithms include Delaunay, Possion, and the rolling sphere method. Delaunay is a triangulation algorithm that, given a point set, inserts each point into a selected region containing that point set, searches and deletes neighboring triangles to form Delaunay cavities, and then connects each point to each vertex in the cavity to form a new triangular mesh. Possion is a reconstruction method based on Poisson meshes, which fits a surface indicator function to point clouds with normal data, extracts isosurfaces by setting a threshold, and achieves 3D reconstruction of the point cloud. The rolling sphere method is a local region growing 3D reconstruction algorithm. It first defines a sphere, which is defined within a seed triangle for rotation and rolling until it touches the next point. The edge of the sphere and the point form a triangle, and the surface triangle is reconstructed through continuous iteration. For point cloud data preprocessed and simplified due to structural damage, 3D reconstruction of the point cloud surface was performed using three point cloud reconstruction algorithms: Delaunay, Possion, and the rolling sphere method. The reconstruction effects were compared and analyzed to determine the optimal 3D point cloud reconstruction algorithm. Figure 6 This is a schematic diagram illustrating the three-dimensional reconstruction effect of the surface damage point cloud of a 316L steel plate according to an embodiment of this application. Figure 6 (a) is a schematic diagram of the reconstruction effect of the Delaunay reconstruction method. Figure 6 (b) is a schematic diagram of the reconstruction effect of the Possion reconstruction method. Figure 6 (c) is a schematic diagram of the reconstruction effect of the rolling ball method.

[0125] Step 1022: The host computer generates a regularized processing path for the damaged area based on the first 3D model, the posture information of the mobile robot arm, and the damage morphology information of the damaged area, and sends the regularized processing path for the damaged area to the mobile robot.

[0126] Step 103: The mobile robot controls the mechanical grinding module on the robotic arm to perform regular processing on the damaged area according to the regularized processing path of the damaged area.

[0127] In this embodiment, a large-scene scan of the entire aircraft is performed using a structured light sensor to identify and locate the damaged areas of the aircraft's metal structure. Subsequently, a robotic arm moves to drive a line laser to scan the damaged areas and obtain damage point cloud data. Based on the shape of the damage, the software selects appropriate regular patterns and uses milling process parameters from the mechanical grinding process library to perform regularization processing on the damaged areas.

[0128] Step 104: The mobile robot interacts with the host computer through the machine vision module to determine the filling repair model corresponding to the damaged area after regularization.

[0129] After the damaged area is regularized by the mechanical grinding module on the robotic arm, the three-dimensional coordinate information of the regularized area needs to be obtained by line laser scanning. Then, a reverse modeling process is performed using 3D reconstruction software to obtain the filling restoration model. The method for obtaining the three-dimensional coordinate information in this step is the same as that used in step 102 when constructing the first 3D model of the damaged area, and will not be repeated here. The method for generating the filling restoration model corresponding to the damaged area based on the obtained three-dimensional coordinate information can refer to the construction method of the first 3D model.

[0130] Step 105: The host computer plans the cladding path and cladding parameters based on the filling repair model, the material of the damaged area, and the current posture information of the mobile robot arm.

[0131] In the actual implementation process, based on the obtained model of the filling restoration, and considering factors such as the material to be repaired, the typical laser cladding process library embedded in the software platform system is selected and called. The typical process information includes laser power, moving speed, defocusing amount, powder feeding speed, and the height and width of a single cladding layer. Through the filling restoration model and the single cladding layer information, the cladding parameters such as the number of cladding layers, the number of passes, and the overlap rate are determined. Based on the slicing processing of the filling restoration model, the cladding path is automatically planned, thereby determining the pose information of the cladding head.

[0132] Step 106: The mobile robot controls the laser generator of the laser cladding module to perform cladding operation on the damaged area according to the cladding path and cladding parameters. During the cladding operation, the robot calls the molten pool monitoring module to monitor the molten pool image, extracts molten pool features based on the molten pool image, and adjusts the cladding path based on the molten pool features.

[0133] In this step, based on the automatically planned cladding path and cladding head pose information, the optimal process is invoked to perform the cladding operation. During the cladding process, a line laser scans the existing weld pool features to capture the thermal deformation generated during cladding, thereby adjusting the cladding path in real time. By setting up a high-speed off-axis camera, the molten pool is captured in real time during the cladding process, and the molten pool features (length, width, and area) are extracted based on software algorithms. Based on the changes in the molten pool features, the process parameters during the cladding process are adjusted in real time to ensure cladding quality.

[0134] The in-situ repair method for aircraft laser cladding based on mobile robots provided in this application has several advantages. First, it enables in-situ repair of aircraft metal structures by directly locating, regularizing, and laser cladding damaged areas using a mobile robot. This overcomes the problems of high difficulty, long cycle, and heavy spare parts supply associated with traditional replacement repairs. Second, it eliminates the need for a "teach-then-clad" approach, thus improving the efficiency of cladding damaged areas and shortening the repair cycle. Third, by combining machine vision modules with host computer modeling to determine the damage location, it enables flexible repair of damaged areas. This flexible repair, through automated and intelligent operation, solves the repair challenges faced by aircraft damaged areas that are random and of varying sizes. It is adaptable to various damage modes and can accurately locate damage.

[0135] Figure 7 The structural block diagram of an in-situ repair system for aircraft laser cladding based on a mobile robot, as described in this application, is shown in the embodiment.

[0136] The aircraft laser cladding in-situ repair system based on a mobile robot provided in this application embodiment includes a mobile robot 201 and a host computer 202. The mobile robot 201 includes a control cabinet 2011, a machine vision module 2012, a laser cladding module 2013, a molten pool monitoring module 2014, and a mechanical grinding module 2015.

[0137] The control cabinet 2011 is used to call the machine vision module 2012 to interact with the host computer 202 to determine the location and damage morphology information of the damaged parts of the metal structure.

[0138] The host computer 202 is used to plan a regularized processing path for the damaged part based on the location information of the damaged part, the damage morphology information, and the current posture information of the mobile robot arm.

[0139] The control cabinet 2011 is also used to control the mechanical grinding module on the robotic arm to perform regular processing on the damaged parts according to the welding path;

[0140] The control cabinet 2011 is also used to call the machine vision module 2012 to interact with the host computer 202 to determine the filling repair model corresponding to the damaged part after the regularization process.

[0141] The host computer 202 is also used to plan the cladding path and cladding parameters based on the filled repair model, the material of the damaged area and the current posture information of the mobile robot arm;

[0142] The control cabinet 2011 is also used to control the laser generator of the laser cladding module to perform cladding operation on the damaged area according to the cladding path and the cladding parameters, and to call the molten pool monitoring module to monitor the molten pool image during the cladding operation, extract molten pool features based on the molten pool image, and adjust the cladding path based on the molten pool features; wherein, the laser generator of the laser cladding module, the mechanical grinding module and the molten pool monitoring module are all located at the front end of the robotic arm.

[0143] Optionally, the laser cladding module includes a laser unit and an auxiliary unit. The laser unit includes a laser generator, and the auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler.

[0144] Optionally, the cooler is a dual-temperature dual-pump series water chiller, and the cooler is equipped with an intelligent temperature control module to control the low-temperature end of the factory water temperature to be 25°C.

[0145] The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

[0146] Optionally, the molten pool monitoring module includes: a high-speed camera, a light shield, and an image acquisition card; the high-speed camera is set next to the test bench of the laser cladding module, and the center of the high-speed camera lens is aligned with the cladding channel to acquire the molten pool image in real time during the cladding operation;

[0147] The light-shielding sheet is placed in front of the high-speed camera lens;

[0148] The image acquisition card is used to convert the electrical signals obtained from the image information acquired by the high-speed camera into digital signals for recognition by the host computer.

[0149] Optionally, the control cabinet includes:

[0150] The first invocation module is used to invoke the machine vision module to capture and acquire image information, and send the image information to the host computer; wherein, the machine vision module is disposed at the front end of the robotic arm;

[0151] The host computer includes:

[0152] The positioning module is used to locate the damaged part of the aircraft's metal structure based on the image information and feed the damaged part information back to the control cabinet; wherein, the image information includes the morphological information of the damaged part;

[0153] The control cabinet also includes:

[0154] The second calling module is used to call the machine vision module to obtain the first point cloud information of the damaged area by laser scanning, and send the point cloud information to the host computer.

[0155] Optionally, the host computer further includes:

[0156] The construction module is used to construct a first three-dimensional model corresponding to the damaged area based on the first point cloud information;

[0157] The path planning module is used to generate a regularized processing path for the damaged part based on the first three-dimensional model, the posture information of the mobile robot arm, and the damage morphology information of the damaged part, and send the regularized processing path for the damaged part to the mobile robot.

[0158] Optionally, the building module is specifically used for:

[0159] The first point cloud information is preprocessed to filter out invalid and outlier points, resulting in the second point cloud information.

[0160] The second point cloud information is filtered out using a voxel grid filtering algorithm to obtain the third point cloud information;

[0161] A preset point cloud reconstruction algorithm is used to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model.

[0162] Optionally, when the construction module preprocesses the first point cloud information to filter out invalid and outlier points to obtain the second point cloud information, it is specifically used for:

[0163] For each point in the first point cloud information, determine whether the coordinates of the point in the first dimension are within a preset first dimension value range;

[0164] If not, the point is considered invalid and filtered out.

[0165] Traverse each point in the first point cloud information and calculate the average distance between the point and each adjacent point; determine whether the average distance is within a preset distance range threshold.

[0166] If not, identify the point and each point adjacent to the point as outliers and filter out all outliers.

[0167] The first point cloud information after filtering out invalid and discrete points is determined as the second point cloud information.

[0168] Optionally, when the construction module uses a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model, it is specifically used for:

[0169] Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models;

[0170] The model with the highest score among the multiple 3D models is selected as the first 3D model.

[0171] The in-situ repair system for aircraft laser cladding based on a mobile robot provided in this application has the following advantages: First, it enables in-situ repair of aircraft metal structures by directly locating, regularizing, and laser cladding damaged areas using a mobile robot. Second, it eliminates the need for a "teach-then-clad" approach, thus improving the cladding efficiency and shortening the repair cycle. Third, by combining a machine vision module with a host computer to model and determine the damage location, it can adapt to various damage patterns and accurately locate damage.

[0172] This invention also provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus.

[0173] Memory, used to store computer programs;

[0174] When the processor executes the program stored in the memory, it implements each step of the in-situ repair method for aircraft laser cladding based on a mobile robot, which is executed by the host computer in the above method embodiment.

[0175] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0176] The communication interface is used for communication between the aforementioned terminal and other devices.

[0177] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0178] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0179] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to implement the method steps executed by the host computer as described in any of the above embodiments.

[0180] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0181] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for in-situ repair of aircraft using laser cladding based on a mobile robot, characterized in that, An application is made to an in-situ repair system for aircraft laser cladding, the system comprising a mobile robot and a host computer integrating a laser cladding module, a molten pool monitoring module, a mechanical grinding module, and a machine vision module, the method comprising: The mobile robot interacts with the host computer through a machine vision module to determine the location and morphology of damage to the metal structure. The host computer plans a regularized processing path for the damaged area based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot arm. The mobile robot controls the mechanical grinding module on the robotic arm to perform regular processing on the damaged area according to the regularized processing path of the damaged area. The mobile robot interacts with the host computer through a machine vision module to determine the filling repair model corresponding to the damaged area after regularization. The host computer plans the cladding path and cladding parameters based on the filled repair model, the material of the damaged area, and the current posture information of the mobile robotic arm. The mobile robot controls the laser generator of the laser cladding module to perform cladding operations on the damaged area according to the cladding path, the cladding parameters, and the welding process parameters. During the cladding operation, the robot calls the molten pool monitoring module to monitor the molten pool image, extracts molten pool features based on the molten pool image, and adjusts the cladding path based on the molten pool features. The laser generator of the laser cladding module, the mechanical grinding module, and the molten pool monitoring module are all located at the front end of the robotic arm.

2. The method according to claim 1, characterized in that, The laser cladding module includes a laser unit and an auxiliary unit. The laser unit includes a laser generator. The auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler. Furthermore, the cooler is a dual-temperature dual-pump series water chiller, and the cooler is equipped with an intelligent temperature control module to control the low-temperature end of the factory water temperature to be 25℃. The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

3. The method according to claim 1, characterized in that, The molten pool monitoring module includes a high-speed camera, a light shield, and an image acquisition card. The high-speed camera is set next to the test bench of the laser cladding module, and the center of the high-speed camera lens is aligned with the cladding channel to capture the molten pool image in real time during the cladding operation. The light-shielding sheet is placed in front of the high-speed camera lens; The image acquisition card is used to convert the electrical signals obtained from the image information acquired by the high-speed camera into digital signals for recognition by the host computer.

4. The method according to claim 1, characterized in that, The steps for the mobile robot to interact with the host computer through a machine vision module to determine the location and morphology of damage to the metal structure include: The machine vision module captures and acquires image information, and sends the image information to the host computer; The host computer locates the damaged area of ​​the aircraft's metal structure based on the image information and feeds back the damaged area information to the mobile robot; wherein, the image information includes the morphological information of the damaged area; The mobile robot invokes a machine vision module to scan the damaged area with a laser, obtaining first point cloud information of the damaged area, and sends the point cloud information to the host computer. Further, the host computer, based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot's robotic arm, plans a regularized processing path for the damaged area, including: The host computer constructs a first three-dimensional model corresponding to the damaged area based on the first point cloud information. The host computer generates a regularized processing path for the damaged area based on the first 3D model, the posture information of the mobile robot arm, and the damage morphology information of the damaged area, and sends the regularized processing path for the damaged area to the mobile robot.

5. The method according to claim 4, characterized in that, The step of the host computer constructing a first three-dimensional model corresponding to the damaged area based on the first point cloud information includes: The host computer preprocesses the first point cloud information to filter out invalid and outlier points, and obtains the second point cloud information. The second point cloud information is filtered out using a voxel grid filtering algorithm to obtain the third point cloud information; A preset point cloud reconstruction algorithm is used to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model; Furthermore, the step of the host computer preprocessing the first point cloud information to filter out invalid and outlier points to obtain the second point cloud information includes: For each point in the first point cloud information, determine whether the coordinates of the point in the first dimension are within a preset first dimension value range; If not, the point is considered invalid and filtered out. Traverse each point in the first point cloud information and calculate the average distance between the point and each adjacent point; determine whether the average distance is within a preset distance range threshold. If not, identify the point and each point adjacent to the point as outliers and filter out all outliers. The first point cloud information after filtering out invalid points and discrete points is determined as the second point cloud information; Furthermore, the step of using a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model includes: Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models; The model with the highest score among the multiple 3D models is selected as the first 3D model.

6. A mobile robot-based in-situ laser cladding repair system for aircraft, characterized in that, The system includes a mobile robot and a host computer. The mobile robot includes a control cabinet, a machine vision module, a laser cladding module, a molten pool monitoring module, and a mechanical grinding module. The control cabinet is used to call the machine vision module to interact with the host computer to determine the location and damage morphology of the damaged parts of the metal structure. The host computer is used to plan a regularized processing path for the damaged area based on the location information of the damaged area, the damage morphology information, and the current posture information of the mobile robot arm. The control cabinet is also used to control the mechanical grinding module on the robotic arm to perform regular processing on the damaged area according to the regularized processing path of the damaged area. The control cabinet is also used to call the machine vision module to interact with the host computer and determine the filling repair model corresponding to the damaged area after regularization. The host computer is also used to plan the cladding path and cladding parameters based on the filled repair model, the material of the damaged area, and the current posture information of the mobile robotic arm. The control cabinet is also used to control the laser generator of the laser cladding module to perform cladding operation on the damaged area according to the cladding path, the cladding parameters and the welding process parameters, and to call the molten pool monitoring module to monitor the molten pool image during the cladding operation, extract molten pool features based on the molten pool image, and adjust the cladding path based on the molten pool features; wherein, the laser generator of the laser cladding module, the mechanical grinding module and the molten pool monitoring module are all located at the front end of the robotic arm.

7. The system according to claim 6, characterized in that, The laser cladding module includes a laser unit and an auxiliary unit. The laser unit includes a laser generator. The auxiliary unit includes an argon gas protector, a powder feeder, a dust collector, and a cooler. Furthermore, the cooler is a dual-temperature dual-pump series water chiller, and the cooler is equipped with an intelligent temperature control module to control the low-temperature end of the factory water temperature to be 25℃. The powder feeder is a scraper-type powder feeder with a flow rate range of 1L / min to 20L / min, a pressure range of -0.1Mpa to 1Mpa, and a powder particle size of 90 mesh to 300 mesh.

8. The system according to claim 6, characterized in that, The molten pool monitoring module includes a high-speed camera, a light shield, and an image acquisition card. The high-speed camera is set next to the test bench of the laser cladding module, and the center of the high-speed camera lens is aligned with the cladding channel to capture the molten pool image in real time during the cladding operation. The light-shielding sheet is placed in front of the high-speed camera lens; The image acquisition card is used to convert the electrical signals obtained from the image information acquired by the high-speed camera into digital signals for recognition by the host computer.

9. The system according to claim 6, characterized in that, The control cabinet includes: The first invocation module is used to invoke the machine vision module to capture and acquire image information, and send the image information to the host computer; wherein, the machine vision module is disposed at the front end of the robotic arm; The host computer includes: The positioning module is used to locate the damaged part of the aircraft's metal structure based on the image information and feed the damaged part information back to the control cabinet; wherein, the image information includes the morphological information of the damaged part; The control cabinet also includes: The second calling module is used to call the machine vision module to scan the damaged area with a laser to obtain the first point cloud information of the damaged area, and send the point cloud information to the host computer. Furthermore, the host computer also includes: The construction module is used to construct a first three-dimensional model corresponding to the damaged area based on the first point cloud information; The path planning module is used to generate a regularized processing path for the damaged part based on the first three-dimensional model, the posture information of the mobile robot arm, and the damage morphology information of the damaged part, and send the regularized processing path for the damaged part to the mobile robot.

10. The system according to claim 9, characterized in that, The building module is specifically used for: The first point cloud information is preprocessed to filter out invalid and outlier points, resulting in the second point cloud information. The second point cloud information is filtered out using a voxel grid filtering algorithm to obtain the third point cloud information; A preset point cloud reconstruction algorithm is used to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model; Furthermore, when the construction module preprocesses the first point cloud information to filter out invalid and outlier points to obtain the second point cloud information, it is specifically used for: For each point in the first point cloud information, determine whether the coordinates of the point in the first dimension are within a preset first dimension value range; If not, the point is considered invalid and filtered out. Traverse each point in the first point cloud information and calculate the average distance between the point and each adjacent point; determine whether the average distance is within a preset distance range threshold. If not, identify the point and each point adjacent to the point as outliers and filter out all outliers. The first point cloud information after filtering out invalid points and discrete points is determined as the second point cloud information; Furthermore, when the construction module uses a preset point cloud reconstruction algorithm to perform point cloud 3D reconstruction on the third point cloud information to obtain the first 3D model, it is specifically used for: Multiple preset point cloud reconstruction algorithms are used to perform point cloud 3D reconstruction on the third point cloud information to obtain multiple 3D models; The model with the highest score among the multiple 3D models is selected as the first 3D model.

Images