Robot inspection system for irregular long and narrow curved surface inner wall

Abstract

The invention belongs to the technical field of nondestructive testing and robots, and relates to an irregular long and narrow curved surface inner wall robot inspection system which comprises a bionic joint module, a detection module, a 3D vision module and a control module. The detection module and the 3D vision module are installed on the bionic joint module, and the control module is connected with the bionic joint module, the detection module and the 3D vision module. The 3D vision module is used for carrying out data acquisition on the irregular long and narrow curved surface inner wall structure; the bionic joint module is used for moving on the inner wall of the irregular long and narrow curved surface; the detection module is used for scanning and detecting the irregular long and narrow curved surface inner wall structure; and the control module is used for controlling the 3D vision module to collect data of the irregular long and narrow curved surface inner wall structure, generating a detection path, controlling the motion posture and path of the bionic joint module, controlling the detection module to scan and detect, and identifying and classifying defects. According to the invention, flexible detection can be realized in a narrow space, and high adaptability, high precision and autonomous operation capability are realized.

Description

Technical Field

[0001] This invention belongs to the field of nondestructive testing and robotics, specifically relating to a robotic inspection system for irregular, narrow, curved inner walls. Background Technology

[0002] In the fields of aircraft, nuclear power, and chemical engineering inspection and maintenance, there are many irregular, narrow, and curved internal wall structures. For example, the aircraft air intake, as a critical component of the engine, directly impacts flight safety by detecting foreign matter (such as metal debris and foreign object accumulation) and structural defects (such as cracks and corrosion). Traditional inspection methods mainly rely on manual visual inspection or handheld flaw detection equipment, requiring the disassembly of some components and the erection of scaffolding, resulting in low efficiency, high cost, and numerous blind spots. Especially for narrow and curved air intake structures, manual inspection not only has limited operating space but also poses safety risks such as personnel falling and contact with hazardous residues. Existing automated inspection equipment mostly uses wheeled or tracked mobile mechanisms, which can partially replace manual labor, but are prone to jamming and difficulty in attitude adjustment in complex geometric pipes. Furthermore, the low sensor integration makes it difficult to simultaneously achieve high-definition imaging, defect location, and foreign matter removal functions. In addition, conventional non-destructive testing techniques (such as ultrasonic and eddy current testing) have unstable coupling effects on curved structures and lack the ability to fuse and analyze multimodal test data in real time, making it difficult to meet the high standards of modern aviation maintenance in terms of testing accuracy and efficiency.

[0003] The cross-sectional shape and size of the inner wall of the irregular, narrow, curved surface continuously vary, including circular and square cross-sections. The maximum and minimum dimensions differ by nearly three times. The span lengths of different curved surfaces vary, ranging from approximately 4 meters to nearly 10 meters. Current technologies primarily rely on the following methods for detecting the inner walls of irregular, narrow, curved surfaces:

[0004] Manual inspection: Operators inspect using endoscopes or handheld devices, which is inefficient, has many blind spots, and poses safety hazards.

[0005] Rigid robotic arm inspection: Limited by the robotic arm's degrees of freedom and range of motion, it is difficult to adapt to complex curvature changes.

[0006] Traditional crawling robots use wheeled or tracked structures, which are prone to slipping on smooth or highly curved surfaces and lack multi-dimensional perception capabilities.

[0007] Poor adaptability: Existing equipment has difficulty adapting to both narrow spaces (length-to-diameter ratio ≥ 10:1) and irregular curved surfaces (abrupt changes in radius of curvature).

[0008] Low detection accuracy: It cannot compensate for pose errors in real time, resulting in distorted sensor data.

[0009] Insufficient autonomy: It relies on preset paths and cannot dynamically avoid obstacles or adjust detection strategies.

[0010] Therefore, there is an urgent need for a non-destructive testing robot suitable for irregular, narrow, curved inner walls to solve the above-mentioned technical problems. Summary of the Invention

[0011] The purpose of this invention is to provide a robotic inspection system for irregular, narrow, curved inner walls. This system can operate in narrow and complex spaces, possessing high adaptability, high precision, and autonomous operation capabilities. It also features flexible motion detection in confined spaces, multi-sensor collaborative detection, and intelligent analysis functions, thus solving the core problems of poor environmental adaptability, single detection dimensions, and high risks associated with manual operation in existing technologies.

[0012] Technical solution to achieve the purpose of this invention:

[0013] A robotic inspection system for irregular, elongated curved inner walls includes: a bionic joint module, a detection module, a 3D vision module, and a control module. The detection module and the 3D vision module are mounted on the bionic joint module, and the control module is connected to the bionic joint module, the detection module, and the 3D vision module respectively. The 3D vision module is used to acquire data on the irregular, elongated curved inner wall structure. The bionic joint module is used to move on the irregular, elongated curved inner wall. The detection module is used to scan and detect the irregular, elongated curved inner wall structure. The control module is used to control the 3D vision module to acquire data on the irregular, elongated curved inner wall structure, generate a detection path, control the movement posture and route of the bionic joint module, control the detection module to scan and detect, and perform defect identification and classification.

[0014] Furthermore, the bionic joint module includes a leg module, which comprises an active leg module and a driven leg module. The active and driven leg modules are symmetrically arranged and connected by three evenly spaced connecting guide rods to ensure the accuracy of the active and driven leg modules at both ends. A transmission shaft is also provided between the active and driven leg modules to realize synchronous movement between the driven and active leg modules. The driven leg module has the same structure as the active leg module, the only difference being that the active leg module is the active end and the driven leg module is the driven end.

[0015] Furthermore, the active outrigger module includes three outriggers, which are evenly spaced and connected to one edge of the outrigger bracket. A first lead screw is located at the center of one end of the outrigger bracket, and an adjusting motor module is located at the center of the other end of the outrigger bracket. The first lead screw is driven and connected to the adjusting motor module, and a first lead screw nut is located on the first lead screw. The adjusting motor module drives the first lead screw to rotate, causing the first lead screw nut to move axially along the first lead screw. Each of the three outriggers is respectively equipped with an adjusting motor assembly, a synchronous pulley and belt, and a second lead screw for adjustment. The motor assembly is connected to the synchronous pulley and belt drive, and the second lead screw is also connected to the synchronous pulley and belt drive. The second lead screw nut is located on the second lead screw. Adjusting the motor assembly drives the second lead screw to rotate through the synchronous pulley and belt, causing the second lead screw nut to move along the axial direction of the second lead screw. Guide rods are also provided on the three support legs, allowing the second lead screw nut to move along the guide rods. The first lead screw nut and the second lead screw nut are adjustablely connected through a connecting rod. Each support leg can be controlled independently, ensuring that each wheel fits tightly against the inner wall, providing better passability and smoother passage over uneven inner walls.

[0016] Furthermore, one of the three legs of the active support leg module is equipped with a drive module and an active wheel at its tail end, and the drive module is connected to the active wheel in a drive connection; the other two legs are respectively equipped with driven wheels at their tail ends.

[0017] Furthermore, the detection module includes a swing head, a telescopic rod, a quick-change connector, a scanner, a rotating module, and a probe mounting bracket; the bottom of the swing head is connected to the bionic joint module, the rotating module is rotatably connected to the swing head, the telescopic rod is connected to the rotating module, the quick-change connector is connected to the telescopic rod, the probe mounting bracket is connected to the quick-change connector, and the scanner is mounted on the probe mounting bracket.

[0018] Furthermore, the scanner includes an ultrasonic probe with a curved surface fitting design, which achieves quantitative defect detection through water film coupling; the scanner also includes an eddy current sensor embedded in a flexible substrate, which is suitable for identifying cracks in composite materials and can effectively improve detection efficiency and safety; the detection module can be equipped with a cleaning probe to simultaneously complete the cleaning of foreign objects and defect location.

[0019] Furthermore, the 3D vision module includes a 3D LiDAR, a surround-view camera, and a laser rangefinder. The 3D LiDAR is used to scan the irregular, narrow, curved inner wall to acquire three-dimensional point cloud data of the scanned environment and send the three-dimensional point cloud data to the control module. The surround-view camera is used to synchronously acquire images of the scanned environment of the 3D LiDAR, acquire corresponding two-dimensional image data, generate a panoramic image with high-resolution texture and color information, and send the panoramic image to the control module. The laser rangefinder is used to perform high-precision distance measurement on specific feature points of the irregular, narrow, curved inner wall to acquire absolute reference distance data and send the distance data to the control module.

[0020] Furthermore, the control module includes a controller, a driver, an I / O module, and a video server;

[0021] The controller sends control commands for acquiring the irregular, narrow, curved inner wall structure to the 3D vision module via the IO module, controlling the 3D vision module to acquire the structure of the irregular, narrow, curved inner wall. It also receives the acquired data from the 3D vision module via the IO module, generating a 3D digital model of the irregular, narrow, curved inner wall. Based on this 3D digital model, the controller calculates the detection path, generates motion control commands and scanning control commands for the detection path, sends the motion control commands to the driver, and sends the scanning control commands to the video server.

[0022] The driver receives motion control commands for the detection path, generates motion control signals for the detection path, and sends them to the bionic joint module through the IO module to control the motion posture and path of the bionic joint module.

[0023] The video server receives the detection path scanning control command, generates the detection path scanning control signal, and sends it to the detection module through the IO module to control the detection module to scan and detect. The video server also receives the scan detection data from the detection module through the IO module, processes and analyzes the scan detection data, generates defect identification and classification results, and sends the defect identification and classification results to the controller.

[0024] Furthermore, the controller adjusts the detection path based on the defect identification and classification results, updates the detection path motion control command and the detection path scanning control command, sends the updated detection path motion control command to the driver, and sends the detection path scanning control command to the video server.

[0025] Furthermore, the controller receives data acquired by the 3D vision module and generates a three-dimensional digital model of the irregular, narrow, curved inner wall, including:

[0026] The controller receives three-dimensional point cloud data sent by the 3D LiDAR and generates a reference point cloud model representing the geometric shape of the inner wall of the irregular narrow curved surface.

[0027] The controller receives panoramic images sent by the surround-view camera, and maps their texture information onto the baseline point cloud model through data fusion, thereby generating a textured color point cloud model.

[0028] The controller receives distance data sent by the laser rangefinder, and performs accuracy correction on the key dimensions of the color point cloud model based on the absolute reference distance data, generating a three-dimensional digital model of the irregular narrow curved inner wall after data fusion and accuracy verification.

[0029] The beneficial technical effects of this invention are as follows:

[0030] 1. The present invention provides a robotic inspection system for irregular narrow curved inner walls, which solves the problems of poor environmental adaptability, single detection dimension and high risk of manual operation in traditional detection methods by integrating autonomous motion, multimodal sensing, intelligent analysis and cleaning functions. It achieves blind spot detection in narrow and curved air intakes, and simultaneously completes the cleaning of foreign objects and precise location of defects, thereby improving detection efficiency and safety and reducing maintenance costs.

[0031] 2. The present invention provides a robotic inspection system for irregular narrow curved inner walls, which has a composite motion mechanism: a bionic joint design of a bionic joint module, which is adapted to irregular narrow curved space with a diameter of 50mm to 500mm.

[0032] 3. The present invention provides a robot inspection system for irregular narrow curved inner walls, which can model the curved surface in real time: the 3D vision module integrates laser and visual data and dynamically updates the detection path.

[0033] 4. The present invention provides a robotic inspection system for irregular narrow curved inner walls, which has autonomous decision-making capabilities: a lightweight neural network (controller) is embedded in the control module to realize defect classification (cracks, corrosion, peeling) and priority determination.

[0034] 5. The present invention provides an irregular, narrow, curved inner wall robot inspection system with the ability to overcome obstacles: each leg of the bionic joint module can be adjusted individually, making it flexible and versatile, and able to avoid obstacles in narrow spaces.

[0035] 6. The present invention provides a robot inspection system for irregular narrow curved inner walls, which has flexible detection capabilities: the detection module uses a swivel head and telescopic rod structure to install the scanning probe, which can adapt to irregular curved surfaces and places with protrusions or pits, thereby improving the scanning accuracy.

[0036] 7. The present invention provides a robotic inspection system for irregular, narrow, curved inner walls, which has the ability to quickly replace probes: the probes of the detection module are installed using a quick-change module, realizing the diversity of probes and the comprehensiveness of functions. Attached Figure Description

[0037] Figure 1 A schematic diagram of the overall structure of a robot inspection system for irregular, narrow, curved inner walls provided by the present invention;

[0038] Figure 2 A schematic diagram of a biomimetic joint module structure in a robot inspection system for an irregular, narrow, curved inner wall provided by the present invention;

[0039] Figure 3 A schematic diagram of the support module structure in a robot inspection system for irregular, narrow, curved inner walls provided by the present invention;

[0040] Figure 4 This is a partially enlarged structural diagram of the leg module in a robot inspection system for irregular, narrow, curved inner walls provided by the present invention.

[0041] Figure 5 A schematic diagram of the detection module structure in a robot inspection system for irregular, narrow, curved inner walls provided by the present invention;

[0042] Figure 6 The block diagram of the control module in a robot inspection system for irregular, narrow, curved inner walls provided by the present invention.

[0043] In the diagram: 1. Bionic joint module; 2. 3D vision module; 3. Detection module; 4. Illumination lamp; 5. Active support leg module; 6. Driven support leg module; 7. Connecting guide rod; 8. Drive shaft; 9. Adjustment motor module; 10. First lead screw; 11. Connecting rod; 12. First lead screw nut; 13. Drive module; 14. Drive wheel; 15. Adjustment motor assembly; 16. Synchronous pulley and belt; 17. Second lead screw nut; 18. Second lead screw; 19. Guide rod; 20. Driven wheel; 21. Swing head; 22. Telescopic rod; 23. Quick-change connector; 24. Scanner; 25. Rotation module; 26. Probe mounting bracket. Detailed Implementation

[0044] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0045] like Figure 1 As shown, the present invention provides a robot inspection system for irregular narrow curved inner walls, including a bionic joint module 1, a detection module 3, a 3D vision module 2, and a control module; the detection module 3 and the 3D vision module 2 are installed on the bionic joint module 1, and the control module is connected to the bionic joint module 1, the detection module 3, and the 3D vision module 2 respectively.

[0046] The 3D vision module is used to acquire data on the irregular, narrow, curved inner wall structure; the bionic joint module is used to move on the irregular, narrow, curved inner wall and provide reliable stability to ensure that the detection module can perform stable detection; the detection module is used to scan and detect the irregular, narrow, curved inner wall structure, and can safely and quickly replace the detection probe to achieve diversity of detection targets; the control module is used to control the 3D vision module to acquire data on the irregular, narrow, curved inner wall structure, calculate and provide the detection path, dynamically avoid obstacles and optimize the detection path, control the movement posture and route of the bionic joint module 1, control the detection module 3 to scan and detect, and perform defect identification, classification (cracks, corrosion, peeling) and priority determination to improve detection efficiency and safety.

[0047] The system of the present invention also includes a lighting lamp 4, which is mounted on the bionic joint module 1.

[0048] like Figure 2-4 As shown, the bionic joint module 1 includes a leg module, an adjustment motor module 9, and a drive module 13.

[0049] The outrigger module includes an active outrigger module 5 and a driven outrigger module 6, which are symmetrically arranged. The active outrigger module 5 and the driven outrigger module 6 are connected by three evenly spaced connecting guide rods 7 to ensure the accuracy of the active outrigger module 5 and the driven outrigger module 6 at both ends. A drive shaft 8 is also provided between the active outrigger module 5 and the driven outrigger module 6 to realize the synchronous movement of the driven outrigger module 6 and the active outrigger module 5.

[0050] The active outrigger module 5 includes three outriggers, which are evenly spaced and connected to one edge of the outrigger bracket. The first lead screw 10 is located at the center of one end of the outrigger bracket, and the adjusting motor module 9 is located at the center of the other end of the outrigger bracket. The first lead screw 10 is driven and connected to the adjusting motor module 9, and the first lead screw nut 12 is located on the first lead screw 10. The adjusting motor module 9 drives the first lead screw 10 to rotate, so that the first lead screw nut 12 moves axially along the first lead screw 10.

[0051] Each of the three support legs is equipped with an adjusting motor assembly 15, a synchronous pulley and belt 16, and a second lead screw 18. The adjusting motor assembly 15 is driven by the synchronous pulley and belt 16, and the second lead screw 18 is also driven by the synchronous pulley and belt 16. A second lead screw nut 17 is mounted on the second lead screw 18. The adjusting motor assembly 15 drives the second lead screw 18 to rotate via the synchronous pulley and belt 16, causing the second lead screw nut 17 to move axially along the second lead screw 18. Guide rods 19 are also provided on the three support legs, allowing the second lead screw nut 17 to move along the guide rods 19. The first lead screw nut 12 and the second lead screw nut 17 are adjustablely connected via a connecting rod 11. Each support leg can be controlled independently, ensuring each wheel fits tightly against the inner wall, providing better passability and smoother movement over uneven inner walls.

[0052] The active support leg module 5 has a drive module 13 on one of its three legs and an active wheel 14 at its tail end. The drive module 13 is connected to the active wheel 14 in a drive connection. The other two support legs are each equipped with a driven wheel 20 at their tail ends.

[0053] By installing a lead screw 18 on each outrigger to adjust the position of the outrigger fixing point, the position of the fixing point of each outrigger can be controlled independently, thereby realizing the ability to adjust the attitude and overcome obstacles in the pipeline with multiple degrees of freedom.

[0054] The driven outrigger module 6 and the active outrigger module 5 have the same structure, the only difference being that the active outrigger module 5 is the active end and the driven outrigger module 6 is the driven end.

[0055] like Figure 5 As shown, the detection module 3 includes a swing head 21, a telescopic rod 22, a quick-change connector 23, a scanner 24, a rotating module 25, and a probe mounting bracket 26. The bottom of the swing head 21 is connected to the bionic joint module. The rotating module 25 is rotatably connected to the swing head 21. The telescopic rod 22 is connected to the rotating module 25. The quick-change connector 23 is connected to the telescopic rod 22. The probe mounting bracket 26 is connected to the quick-change connector 23. The scanner 24 is mounted on the probe mounting bracket 26. The combination of the swing head 21, the rotating module 25, and the telescopic rod 22 drives the scanner 24 to rotate and rise in multiple directions to adapt to the complex environment of the inner wall. The quick-change connector 23 enables the rapid replacement of the scanner 24.

[0056] The detection module can also be equipped with a cleaning probe to simultaneously clean up foreign objects and locate defects. The scanner 24 includes an ultrasonic probe with a curved surface fitting design, which achieves quantitative detection of defects through water film coupling. The scanner 24 also includes an eddy current sensor, which is embedded in a flexible substrate and is suitable for identifying cracks in composite materials, effectively improving detection efficiency and safety.

[0057] 3D vision module 2 includes a 3D LiDAR, a surround-view camera, and a laser rangefinder;

[0058] 3D LiDAR is used to scan the inner wall of an irregular, narrow curved surface, acquire three-dimensional point cloud data of the scanned environment, and send the three-dimensional point cloud data to the controller of the control module to generate a reference point cloud model that characterizes the geometric shape of the inner wall of the irregular, narrow curved surface.

[0059] The surround-view camera is used to synchronously acquire images of the scanning environment of the 3D LiDAR, obtain the corresponding two-dimensional image data, generate a panoramic image with high-resolution texture and color information, and send the panoramic image to the controller of the control module. Through data fusion, its texture information is mapped onto the reference point cloud model, thereby generating a textured color point cloud model.

[0060] The laser rangefinder is used to perform high-precision distance measurement on specific feature points (such as welds, depressions, or corroded areas) of the inner wall of an irregular, narrow curved surface. It acquires absolute reference distance data, sends the distance data to the controller of the control module, and corrects the key dimensions of the color point cloud model based on the absolute reference distance data. It then generates a three-dimensional digital model of the inner wall of the irregular, narrow curved surface after data fusion and accuracy verification.

[0061] like Figure 6 As shown, the control module includes a controller, a driver, an I / O module, and a video server.

[0062] The controller sends control commands for acquiring the irregular, narrow, curved inner wall structure to the 3D vision module via the IO module, controlling the 3D vision module to acquire the structure of the irregular, narrow, curved inner wall. It also receives the acquired data from the 3D vision module via the IO module, generating a 3D digital model of the irregular, narrow, curved inner wall. Based on this 3D digital model, the controller calculates the optimal detection path using a dynamic path planning algorithm, generating motion control commands and scanning control commands for the detection path. The motion control commands are then sent to the driver, and the scanning control commands are sent to the video server.

[0063] The controller receives data from the 3D vision module and generates a 3D digital model of the irregular, narrow, curved inner wall, including:

[0064] The controller receives three-dimensional point cloud data sent by the 3D LiDAR and generates a reference point cloud model representing the geometric shape of the inner wall of the irregular narrow curved surface.

[0065] The controller receives panoramic images sent by the surround-view camera, and maps their texture information onto the baseline point cloud model through data fusion, thereby generating a textured color point cloud model.

[0066] The controller receives distance data sent by the laser rangefinder, and performs accuracy correction on the key dimensions of the color point cloud model based on the absolute reference distance data, generating a three-dimensional digital model of the irregular narrow curved inner wall after data fusion and accuracy verification.

[0067] The driver receives the motion control command for the detection path, generates the motion control signal for the detection path, and sends it to the bionic joint module 1 through the IO module to control the motion posture and path of the bionic joint module 1.

[0068] The video server receives the detection path scanning control command, generates the detection path scanning control signal, and sends it to the detection module 3 through the IO module to control the detection module 3 to scan and detect. The video server also receives the scanning detection data from the detection module 3 through the IO module, processes and analyzes the scanning detection data, generates defect identification and classification results, and sends the defect identification and classification results to the controller. The controller adjusts the detection path according to the defect identification and classification results, updates the detection path motion control command and the detection path scanning control command, sends the updated detection path motion control command to the driver, and sends the detection path scanning control command to the video server.

[0069] In one specific implementation, the video server receives the scanning and detection data from the detection module 3 through the edge computing unit, processes and analyzes the scanning and detection data, uses a defect identification algorithm to mark suspicious areas, generates defect identification and classification results, and sends the defect identification and classification results to the controller; at the same time, it can also trigger high-precision re-inspection, switch to the high-frequency probe, and re-scan and detect.

[0070] The control module also includes a lower-level machine for receiving instructions from the upper-level machine. The upper-level machine generates control instructions for acquiring irregular, narrow, curved inner wall structures and sends these instructions to the lower-level machine via a PC.

[0071] The control module also includes a power supply for powering the control module.

[0072] This invention provides a robotic inspection system for irregular, narrow, curved inner walls. Before fully automated crawling on the inner wall, the zero position of the locator needs to be determined. This allows for 100% inspection of all heat transfer pipes through path planning, and effective anti-collision strategies are planned throughout the inspection process. The robot is placed at the inlet of the pipe to be inspected, and a self-test program is initiated to initialize sensors and communication links. Before starting formal movement, the robot initializes its various motors. Then, the 3D vision module scans the inner wall of the initial section using a 3D LiDAR to generate a reference point cloud model. After image acquisition by a surround-view camera and distance correction by a laser rangefinder, a three-dimensional digital model is generated. The control module then calculates the optimal inspection path using a dynamic path planning algorithm, controls the bionic joint module to adjust the robot's posture, and begins crawling towards the narrow, curved inner wall. The detection module performs scanning inspection, with an ultrasonic phased array probe scanning at a preset step distance. Data is uploaded to the edge computing unit in real time, and a defect identification algorithm marks suspicious areas, triggering a high-precision re-inspection (switching to a high-frequency probe). When the robot encounters a sudden obstacle, it activates a three-level obstacle avoidance strategy (deceleration → detour → emergency suction braking). After communication is interrupted, it enables local storage and attempts to resume transmission from the point of interruption.

[0073] The present invention has been described in detail above with reference to the accompanying drawings and embodiments. However, the present invention is not limited to the above embodiments, and various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention. All contents not described in detail in the present invention can be derived from existing technologies.

Claims

1. An irregular elongated curved surface inner wall robot inspection system, characterized in that, It comprises: a bionic joint module (1), a detection module (3), a 3D vision module (2), and a control module; The detection module (3) and the 3D vision module (2) are installed on the bionic joint module (1), and the control module is connected with the bionic joint module (1), the detection module (3), and the 3D vision module (2) respectively. The 3D vision module is used for data collection of the irregular narrow curved inner wall structure. The bionic joint module is used for movement of the irregular narrow curved inner wall. The detection module is used for scanning detection of the irregular narrow curved inner wall structure. The control module is used for controlling the 3D vision module to collect data of the irregular narrow curved inner wall structure, generating a detection path, controlling the movement posture and route of the bionic joint module (1), and controlling the scanning detection of the detection module (3) to identify and classify defects.

2. A robot inspection system for irregularly shaped elongated curved inner walls as recited in claim 1, wherein, The bionic joint module (1) comprises a supporting leg module, which comprises a driving supporting leg module (5) and a driven supporting leg module (6). The driving supporting leg module (5) and the driven supporting leg module (6) are symmetrically arranged, and the driving supporting leg module (5) and the driven supporting leg module (6) are connected through three uniformly spaced connecting guide rods (7) to ensure the accuracy of the two ends of the driving supporting leg module (5) and the driven supporting leg module (6). A transmission shaft (8) is further arranged between the driving supporting leg module (5) and the driven supporting leg module (6) to realize synchronous movement of the driven supporting leg module (6) and the driving supporting leg module (5). The structure of the driven supporting leg module (6) and the driving supporting leg module (5) is completely the same, and the only difference is that the driving supporting leg module (5) is the driving end and the driven supporting leg module (6) is the driven end.

3. A robot inspection system for irregularly shaped elongated curved inner walls as recited in claim 2, wherein, The active leg module (5) comprises three legs, the three legs are evenly spaced and connected to the edge of one end of the leg support, the first lead screw (10) is arranged at the center of one end of the leg support, the adjusting motor module (9) is arranged at the center of the other end of the leg support, the first lead screw (10) is drivingly connected with the adjusting motor module (9), and the first lead screw nut (12) is arranged on the first lead screw (10); the adjusting motor module (9) drives the first lead screw (10) to rotate, so that the first lead screw nut (12) moves along the first lead screw (10) in the axial direction; the three legs are respectively provided with an adjusting motor assembly (15), a synchronous belt pulley and a belt (16), and a second lead screw (18), the adjusting motor assembly (15) is drivingly connected with the synchronous belt pulley and the belt (16), the second lead screw (18) is drivingly connected with the synchronous belt pulley and the belt (16), and the second lead screw nut (17) is arranged on the second lead screw (18); the adjusting motor assembly (15) drives the second lead screw (18) to rotate through the synchronous belt pulley and the belt (16), so that the second lead screw nut (17) moves along the second lead screw (18) in the axial direction; the three legs are also provided with guide rods (19), so that the second lead screw nut (17) moves along the guide rods (19); the first lead screw nut (12) and the second lead screw nut (17) are adjustably connected through the connecting rod (11); each leg can be controlled independently, so that each wheel is closely attached to the inner wall, has better passability, and passes through the uneven inner wall more stably.

4. A robot inspection system for irregularly shaped elongated curved inner walls according to claim 3, wherein, One of the three legs of the active leg module (5) is provided with a driving module (13), and a driving wheel (14) is arranged at the tail end, and the driving module (13) is drivingly connected with the driving wheel (14); the tail ends of the other two legs are respectively provided with driven wheels (20).

5. The irregularly shaped elongated curved inner wall robotic inspection system of claim 1, wherein, The detection module (3) comprises a swing head (21), an extension rod (22), a quick-change joint (23), a scanner (24), a rotating module (25) and a probe fixing frame (26); the swing head (21) is connected to the bionic joint module at the bottom, the rotating module (25) is rotatably connected to the swing head (21), the extension rod (22) is connected to the rotating module (25), the quick-change joint (23) is connected to the extension rod (22), the probe fixing frame (26) is connected to the quick-change joint (23), and the scanner (24) is installed on the probe fixing frame (26).

6. A robot inspection system for irregularly shaped elongated curved inner walls according to claim 5, wherein, The scanner (24) comprises an ultrasonic probe, the ultrasonic probe adopts a curved surface fitting design, and quantitative detection of defects is realized through water film coupling; the scanner (24) further comprises an eddy current sensor, the eddy current sensor is embedded in a flexible substrate, is suitable for composite material crack identification, and can effectively improve detection efficiency and safety; the detection module can carry a cleaning probe and simultaneously complete cleaning of excess materials and defect positioning.

7. The irregularly shaped elongated curved inner wall robotic inspection system of claim 1, wherein, The 3D vision module (2) comprises a 3D laser radar, a surround-view camera and a laser range finder; the 3D laser radar is used for scanning the irregular long and narrow curved inner wall to obtain three-dimensional point cloud data of the scanned environment and sending the three-dimensional point cloud data to the control module; the surround-view camera is used for synchronously collecting images of the scanning environment of the 3D laser radar to obtain corresponding two-dimensional image data, generating a panoramic image with high-resolution texture and color information and sending the panoramic image to the control module; and the laser range finder is used for high-precision ranging on specific feature points of the irregular long and narrow curved inner wall to obtain absolute reference distance data and sending the distance data to the control module.

8. A robot inspection system for irregularly shaped elongate curved inner walls according to claim 7, wherein, The control module comprises a controller, a driver, an IO module and a video server. The controller sends irregular long and narrow curved inner wall structure collection control instructions to the 3D vision module through the IO module, controls the structure collection of the irregular long and narrow curved inner wall of the 3D vision module, receives the collection data of the 3D vision module through the IO module and generates an irregular long and narrow curved inner wall three-dimensional digital model; according to the irregular long and narrow curved inner wall three-dimensional digital model, a detection path is calculated, detection path motion control instructions and detection path scanning control instructions are generated, the detection path motion control instructions are sent to the driver and the detection path scanning control instructions are sent to the video server; The driver receives the detection path motion control instructions, generates detection path motion control signals, sends the detection path motion control signals to the bionic joint module (1) through the IO module and controls the motion posture and route of the bionic joint module (1); The video server receives the detection path scanning control instructions, generates detection path scanning control signals, sends the detection path scanning control signals to the detection module (3) through the IO module, controls the scanning detection of the detection module (3), receives the scanning detection data of the detection module (3) through the IO module, processes and analyzes the scanning detection data, generates defect identification and classification results and sends the defect identification and classification results to the controller.

9. A robot inspection system for irregularly shaped elongated curved inner walls according to claim 8, wherein, The controller adjusts the detection path according to the defect identification and classification results, updates the detection path motion control instructions and the detection path scanning control instructions, sends the updated detection path motion control instructions to the driver and sends the detection path scanning control instructions to the video server.

10. The irregularly shaped elongated curved inner wall robotic inspection system of claim 8, wherein, The controller receives the collection data of the 3D vision module and generates an irregular long and narrow curved inner wall three-dimensional digital model, which comprises: The controller receives the three-dimensional point cloud data sent by the 3D laser radar and generates a reference point cloud model representing the geometric appearance of the irregular long and narrow curved inner wall; The controller receives the panoramic image sent by the surround-view camera, maps the texture information of the panoramic image onto the reference point cloud model through data fusion, thereby generating a color point cloud model with texture; and The controller receives the distance data sent by the laser range finder, performs precision correction on the key dimensions of the color point cloud model according to the absolute reference distance data and generates an irregular long and narrow curved inner wall three-dimensional digital model after data fusion and precision verification.

Images