An online multi-angle image taking detection device based on AI machine vision

Abstract

The utility model relates to a kind of online multi-angle image-taking detection equipment based on AI machine vision, image acquisition component is installed on annular guide rail by connecting plate, annular guide rail is open annular structure, annular guide rail is rotatably installed on base plate by multiple guide wheels, driving mechanism for driving annular guide rail rotates around its central axis is installed on base plate, light source is installed on annular guide rail, machine replaces manual, can be carried out online instant detection in dangerous force test environment, personnel safety risk is completely eliminated, and transient leakage that evaporates easily can be captured;Adopt machine vision and AI algorithm to determine, exclude human subjective factor;The rotary motion of annular guide rail combines the independent adjustment of camera angle, realizes the coverage without dead angle to the outer surface of the workpiece to be detected, multiple acquisition components can work in parallel, improve detection efficiency;Open annular guide rail design makes equipment can easily fit into the workpiece of complex pipeline layout, installation is convenient, and applicability is wide.

Description

Technical Field

[0001] This utility model relates to the field of automated inspection technology, and in particular to an online multi-angle image acquisition and inspection device based on AI machine vision. Background Technology

[0002] In high-tech and heavy industrial sectors such as aerospace, energy and power, and petrochemicals, the operational safety of critical pressure-bearing equipment such as core pumps, high-pressure valves, and pipeline flanges is paramount. When these devices operate under harsh conditions such as high temperature, high pressure, and vibration for extended periods, their sealing components are at risk of leaking media (such as fuel, lubricating oil, and coolant) due to fatigue, corrosion, and other factors. Even minor leaks, if not detected and addressed promptly, can not only lead to resource waste and environmental pollution but also potentially trigger catastrophic accidents such as fires and explosions, threatening human lives and causing enormous property damage. Therefore, efficient and reliable leak detection for these devices is a crucial step in ensuring production safety.

[0003] Currently, for liquid leak detection in such equipment, especially online detection during testing or operation, the mainstream method still heavily relies on manual operation. The typical detection process is as follows: after the equipment completes a phase of pressurization or stress testing, it is first shut down, depressurized, and cooled. Once the equipment is stable and the environment is safe, inspection personnel, carrying specialized tools (such as UV flashlights), enter the site to check for potential leak points one by one. Specifically, for liquids containing fluorescent tracers, inspection personnel use UV light to illuminate the equipment surface and visually observe for any fluorescent reaction to determine if a leak exists.

[0004] However, this traditional manual inspection method has a series of inherent and insurmountable drawbacks, mainly in the following aspects: Poor safety and inability to achieve true online detection: Leaks are most likely to occur during stress tests or normal operation of equipment. However, the environment at this stage is often accompanied by high temperature, high pressure, high decibel noise, or the presence of toxic and harmful gases, posing a serious threat to the personal safety of testing personnel. Therefore, for safety reasons, testing must wait until the equipment is completely shut down and environmental parameters return to normal before it can be carried out. This results in a severe time disconnect between the testing action and the leakage process, a typical "offline" detection that cannot capture the instantaneous state of the leak.

[0005] Poor timeliness and the risk of missed detection due to medium evaporation: There is a significant time window (usually several minutes or even hours) between equipment shutdown and personnel arrival for inspection. For volatile leaked media (such as fuels, some solvents) or minute "leakage," the leaked liquid may have completely evaporated within this time window. This makes it impossible for inspectors to observe any traces of the leak, resulting in serious missed detections and leaving huge safety hazards.

[0006] The subjective nature of the testing standards leads to poor consistency in results: Test results rely entirely on the experience, sense of responsibility, eyesight, and even the mental state of the testing personnel. Different personnel may have different standards for judging fluorescence intensity and trace morphology, and even the same person may make misjudgments when fatigued. This subjectivity results in a lack of objectivity and repeatability, making it impossible to establish unified and reliable judgment standards, thus introducing uncertainty into the safety status assessment of equipment.

[0007] Low testing efficiency impacts overall operational efficiency: Manual testing requires downtime, waiting, and step-by-step troubleshooting, a time-consuming and labor-intensive process that significantly extends the equipment's testing or maintenance cycle and reduces production efficiency. This problem is particularly pronounced for equipment with complex structures and numerous testing points.

[0008] To address these issues, some experimental technical solutions have emerged in the industry, such as laying oil-absorbing pads on the equipment surface or using leak detection lines. However, these methods are slow to respond, cannot pinpoint leaks accurately, and still require manual intervention. Therefore, there is an urgent need for an online leak detection device that can replace manual labor and provide automated, intelligent, and highly timely leak detection while the equipment is in operation. Summary of the Invention

[0009] To overcome the shortcomings of existing technologies, this utility model provides an online multi-angle imaging and detection device based on AI machine vision. It realizes machine replacement of manual labor, can perform online real-time detection in dangerous force testing environments, completely eliminates personnel safety risks, and can capture instantaneous leaks that are easily evaporated. The device uses machine vision and AI algorithms for judgment, which eliminates subjective human factors, and the detection standards are objective, uniform, and reliable.

[0010] To solve the above-mentioned technical problems, this utility model provides the following technical solution: An online multi-angle image acquisition and detection device based on AI machine vision includes: an image acquisition component, a light source, an annular guide rail, and a substrate. The image acquisition component is mounted on the annular guide rail via a connecting plate. The annular guide rail has an open annular structure. The annular guide rail is rotatably mounted on the substrate via multiple guide wheels. A drive mechanism for driving the annular guide rail to rotate around its central axis is mounted on the substrate. The light source is mounted on the annular guide rail.

[0011] Furthermore, the image acquisition component includes a camera and an angle adjustment motor for driving the camera, the camera being mounted on the angle adjustment motor via a swing arm.

[0012] Furthermore, the driving mechanism includes a motor, a gear, and a gear ring fixedly mounted on the base plate. The gear meshes with the gear ring, and the gear ring is coaxially connected to the annular guide rail.

[0013] Furthermore, the guide wheel is a roller bearing, and the rim of the guide wheel contacts the track of the annular guide rail.

[0014] Furthermore, the image acquisition component also includes a swing arm that connects the output shaft of the angle adjustment motor to the camera.

[0015] Furthermore, the light source is one of ultraviolet light, visible light, or infrared light.

[0016] Furthermore, the opening width of the annular guide rail is wide enough to allow pipes or cables of the workpiece to be inspected to pass through.

[0017] Furthermore, the image acquisition components and the light source are in multiple groups, and the image acquisition components and the light source are arranged at intervals along the circumference of the annular guide rail.

[0018] Furthermore, the cameras of the multiple sets of image acquisition components can be adjusted independently.

[0019] Furthermore, the substrate is provided with mounting holes or clamping structures for fixing the entire device to an external platform.

[0020] 1. Safety and online detection: It realizes the replacement of manual labor by machines, and can perform online real-time detection in dangerous stress testing environments, completely eliminating personnel safety risks, and can detect instantaneous leaks that are easy to evaporate.

[0021] 2. Unified testing standards: The use of machine vision and AI algorithms for judgment eliminates subjective human factors, ensuring that the testing standards are objective, unified, and the results are reliable.

[0022] 3. High efficiency and high coverage: The rotational motion of the ring guide rail combined with the independent adjustment of the camera angle achieves coverage of the outer surface of the workpiece to be inspected without any blind spots. Multiple sets of acquisition components can work in parallel, which greatly improves the inspection efficiency.

[0023] 4. Compact structure and easy installation: The unique open ring guide rail design allows the equipment to easily fit into workpieces with complex pipeline layouts, making installation convenient and widely applicable. Attached Figure Description

[0024] Figure 1 This is an exploded view of the present invention; Figure 2 This is a schematic diagram of the overall structure of this utility model; Figure 3 This is a schematic diagram of the image acquisition component in this utility model; Figure 4 This is a schematic diagram of the guide wheel in this utility model.

[0025] The serial numbers in the figure are: 1. Image acquisition component; 2. Light source; 3. Gear ring; 4. Circular guide rail; 5. Guide wheel; 6. Gear; 7. Base plate; 8. Drive motor; 9. Detected workpiece; 11. Connecting plate; 12. Angle adjustment motor; 13. Swing arm; 14. Drive camera; 15. Camera. Detailed Implementation

[0026] To make the technical means, creative features, and achieved objectives and effects of this utility model easier to understand, the present utility model is further described below with reference to specific embodiments. However, the following embodiments are merely preferred embodiments of this utility model and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments described herein without creative effort are all within the protection scope of this utility model. Unless otherwise specified, the experimental methods in the following embodiments are conventional methods, and the materials and reagents used in the following embodiments are commercially available unless otherwise specified.

[0027] Combination Figures 1 to 4 As shown, this utility model provides an online multi-angle imaging and inspection device based on AI machine vision, including a base plate 7 as the overall mounting base. A ring guide rail 4 is rotatably mounted on the base plate 7 via multiple guide wheels 5. The guide wheels 5 are mounted on the base plate 7 via shafts. The ring guide rail 4 has an open ring structure. The opening width of the ring guide rail 4 is set to allow the pipes or cables of the workpiece 9 to be inspected to pass through, so as to realize the installation method of the device being inserted from the side of the workpiece. A drive mechanism is also mounted on the base plate 7. The drive mechanism includes a drive motor 8 fixed on the base plate 7, a gear 6 directly driven by the drive motor 8, and a gear ring 3 meshing with the gear 6. The gear 6 is mounted on the output end of the drive motor 8 via a coupling. The gear ring 3 is coaxially fixedly connected to the ring guide rail 4. When the drive motor 8 drives the gear 6, the gear 6 can drive the ring guide rail 4 to rotate continuously or stepwise around its central axis by meshing with the teeth on the gear ring 3.

[0028] At least one image acquisition assembly 1 is installed on the annular guide rail 4. The image acquisition assembly 1 includes an angle adjustment motor 12 fixed to the annular guide rail 4 via a connecting plate 11, and a camera 14 connected to the output shaft of the angle adjustment motor 12 via a swing arm 13. A camera 15 is installed at the front end of the camera 14. The angle adjustment motor 12 is used to drive the camera 14 to pitch and swing around an axis parallel to the tangent of the annular guide rail 4 to adjust the image acquisition angle of the camera 14 relative to the workpiece 9 to be inspected. A light source 2 is also installed on the annular guide rail 4 to provide illumination for the camera 14 to acquire images.

[0029] In this embodiment, the guide wheel 5 is a roller bearing. The rim of the guide wheel 5 forms a V-shaped groove or an arc-shaped groove. The track of the annular guide rail 4 has a raised profile that matches the groove. Through the cooperation of the groove and the raised profile, the radial and axial limits of the annular guide rail 4 are achieved. The inner surface of the groove of the guide wheel 5 rim forms a surface contact with the raised profile of the track of the annular guide rail 4, which not only increases the contact area but also improves the load-bearing capacity and stability. The contact surface is precision machined to ensure a moderate coefficient of friction, ensuring smooth rolling while avoiding slippage and providing stable rolling support. When the annular guide rail 4 is subjected to radial force, the two inclined surfaces of the groove of the guide wheel 5 will simultaneously contact the raised part of the track of the annular guide rail 4, forming a bidirectional constraint. This design can effectively prevent the annular guide rail 4 from shifting and wobbling in the radial plane, ensuring the concentricity of the rotational motion. The bottom profile of the groove of the guide wheel 5 forms an axial constraint with the top of the raised part of the track of the annular guide rail 4. When the annular guide rail 4 is subjected to axial force, the bottom of the groove will generate contact resistance with the raised part of the track, preventing the annular guide rail 4 from moving in the axial direction.

[0030] The light source 2 is one of ultraviolet light source, visible light source or infrared light source. The shape of the light source 2 is one of ring light source, strip light source or surface light source, and it is fixedly installed facing the workpiece 9 to be tested, so as to provide illumination of a specific wavelength to excite fluorescence reaction or form uniform illumination.

[0031] In this embodiment, multiple sets of image acquisition components 1 and light sources 2 are provided and are spaced apart along the circumference of the annular guide rail 4. The cameras 14 of the multiple sets of image acquisition components 1 can independently adjust the motor 12 to different angles by their respective angles, so as to synchronously acquire images of multiple angles on the same circumferential surface of the workpiece 9 to be inspected or to segment and be responsible for image acquisition of different areas, thereby improving detection efficiency and coverage.

[0032] The base plate 7 is provided with multiple mounting holes or quick clamp structures for firmly fixing the device to an external platform or support frame. The base plate 7 is made of metal alloy or carbon fiber composite material to provide a stable support base while taking into account the requirements of lightweight. The annular guide rail 4 and the gear ring 3 are made of high carbon chromium bearing steel, alloy structural steel or aluminum alloy with hard anodized surface to ensure the wear resistance, strength and motion accuracy of the transmission components.

[0033] The wheel body of the guide wheel 5 is made of high carbon chromium bearing steel, and its outer ring is covered with a wear-resistant layer of polyurethane or engineering plastic to reduce operating noise and wear on the annular guide rail 4 while ensuring load-bearing strength. The cross-section of the annular guide rail 4 is one of I-shaped, V-shaped, T-shaped or rectangular to optimize the use of materials while ensuring structural rigidity.

[0034] In use, the substrate 7 is fixed to an external platform via its mounting holes or clamps to ensure overall stability. Utilizing the opening design of the annular guide rail 4, the workpiece 9 to be inspected is inserted through the side opening of the invention, positioning it at the center of the guide rail. The position of the workpiece 9 relative to the image acquisition device is adjusted to ensure that the image acquisition range of the image acquisition component 1 covers the area to be inspected. The contact between the guide wheel 5 and the annular guide rail 4 is checked for smoothness to avoid jamming. The power interface on the substrate 7 is connected to the control system, and the motor 8 in the drive mechanism is started. The meshing state of the gear 6 and the gear ring 3 is checked. The camera 14 and the light source 2 in the image acquisition component 1 are tested with power to ensure normal communication. The angle adjustment motor 12 is adjusted via software to initially set the image acquisition angle of the camera 14. The rotation mode of the annular guide rail 4 is set to continuous / stepped in the control interface. The brightness and switching sequence of the light source 2 (ultraviolet / visible / infrared light) are adjusted. The tilt angle of each camera 14 is independently adjusted by fixing the image acquisition component 1 in position via the connecting plate 11. The pitch angle is adjusted by the angle adjustment motor 12 driving the swing arm 13. The preview camera 15 displays the image, optimizing the image capture range. The motor 8 is started, driving the gear 6 and gear ring 3 to rotate the annular guide rail 4. The guide wheel 5 supports the smooth movement of the guide rail, and multiple image acquisition components 1 simultaneously acquire images. The light source 2 provides specific wavelength illumination, enhancing the identification of leak traces. Images are transmitted in real-time to the AI ​​processing system for analysis. After the AI ​​system marks the leak location, it controls the annular guide rail 4 to stop rotating.

[0035] The image acquisition and inspection device proposed in this utility model is connected to an image processing system. The image processing system has a built-in AI algorithm model for real-time analysis of multi-angle images acquired by the camera 14, and can automatically identify and judge leakage defects of the workpiece 9 to be inspected.

[0036] The foregoing has shown and described the basic principles, main features, and advantages of this utility model. Those skilled in the art should understand that this utility model is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the utility model. Various changes and modifications can be made to this utility model without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed utility model. The scope of protection of this utility model is defined by the appended claims and their equivalents.

Claims

1. An online multi-angle image acquisition and detection device based on AI machine vision, comprising: The image acquisition component (1), light source (2), annular guide rail (4) and substrate (7) are characterized in that the image acquisition component (1) is mounted on the annular guide rail (4) via a connecting plate (11), the annular guide rail (4) is an open annular structure, the annular guide rail (4) is rotatably mounted on the substrate (7) via multiple guide wheels (5), the substrate (7) is equipped with a drive mechanism for driving the annular guide rail (4) to rotate around its central axis, and the light source (2) is mounted on the annular guide rail (4).

2. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The image acquisition component (1) includes a camera (14) and an angle adjustment motor (12) for driving the camera (14), the camera (14) being mounted on the angle adjustment motor (12) via a swing arm (13).

3. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The drive mechanism includes a motor (8), a gear (6) and a gear ring (3) fixedly mounted on the base plate (7). The gear (6) meshes with the gear ring (3), and the gear ring (3) is coaxially connected to the annular guide rail (4).

4. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The guide wheel (5) is a roller bearing, and the rim of the guide wheel (5) is in contact with the track of the annular guide rail (4).

5. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 2, characterized in that, The image acquisition component (1) also includes a swing arm (13) that connects the output shaft of the angle adjustment motor (12) to the camera (14).

6. The online multi-angle imaging and detection device according to claim 1, characterized in that, The light source (2) is one of ultraviolet light source, visible light source or infrared light source.

7. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The opening width of the annular guide rail (4) is wide enough to allow the pipes or cables of the workpiece (9) to be inspected to pass through.

8. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The image acquisition component (1) and the light source (2) are each provided in multiple sets, and the image acquisition component (1) and the light source (2) are arranged at intervals along the circumference of the annular guide rail (4).

9. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 8, characterized in that, The cameras (14) of the multiple sets of image acquisition components (1) can be adjusted independently.

10. The online multi-angle image acquisition and detection device based on AI machine vision according to claim 1, characterized in that, The substrate (7) is provided with mounting holes or clamping structures for fixing the entire device to an external platform.

Images