An AI vision-based platform motion measurement method and device

Abstract

This invention belongs to the field of motion measurement device technology, specifically an AI vision-based platform motion measurement method and device, including a data acquisition module, a processing module, and a control module. The data acquisition module includes a vision acquisition unit, a measurement unit, and a mounting base. It also includes a heat sink, which is mounted on the mounting base and houses the vision acquisition unit. One end of the heat sink has a lens that matches the data acquisition unit, and vibration damping blocks are provided on the inner wall of the heat sink. A drive motor rotates a rotating rod, which in turn rotates a protective film via a rotating frame. The protective film rotates out of a cleaning box, and the cleaning rod rolls over the protective film to clean it, improving its cleanliness before installation. After rotating out, the protective film moves until it contacts the lens and stops, where the lens and protective film come into contact and adhere. By installing an anti-strong light protective film on the lens, the accuracy of motion measurement is improved while reducing the amount of abnormal data rejected by AI, thereby improving measurement efficiency.

Description

Technical Field

[0002] This invention belongs to the field of motion measurement device technology, specifically a platform motion measurement method and device based on AI vision. Background Technology

[0003] AI vision-based motion measurement devices, with their multi-sensor fusion and AI anti-interference optimization features, are widely adaptable to complex outdoor working conditions. These include health monitoring of infrastructure such as wind turbines, transmission lines, and large bridges, measuring structural dynamic deformation and vibration displacement. They operate stably for extended periods in outdoor environments with large temperature differences and frequent wind and rain. Outdoor measurements follow a closed-loop process: multi-source data collaborative acquisition, preprocessing and AI optimization, constraint fusion, motion estimation, and error correction. Anti-interference design is strengthened to address the characteristics of outdoor environments. The steps include synchronous multi-source data acquisition, data preprocessing and AI anti-interference optimization, constraint modeling and multi-source data fusion, motion parameter estimation and displacement compensation, and real-time error correction and adaptive adjustment. However, during prolonged use, strong light and overexposure can cause surface feature points, such as black positioning marks and metal edges, to blend into the background, significantly reducing the feature point detection rate. Additionally, halos and glare can lead to mismatches of feature points. Furthermore, strong light reflection can cause localized image brightness saturation, decreasing the accuracy of feature point coordinate extraction. All of these factors contribute to the loss of feature point details, reduced grayscale contrast, and decreased AI feature point detection and matching accuracy, affecting the data accuracy of the measurement device. Moreover, prolonged exposure to the elements can lead to dust accumulation, which physically obstructs the lens, reducing light transmittance and creating noise in the image, thus compromising the integrity of feature points and causing misidentification or tracking loss by the AI ​​algorithm. Additionally, water mist or droplets can cause light refraction, blurring of images, and blurring of feature point edges, reducing the synchronization of data from multiple camera arrays and consequently decreasing the accuracy of measurement data. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies and solve the aforementioned technical problems, this invention proposes a platform motion measurement method and device based on AI vision.

[0005] The technical solution adopted by this invention to solve its technical problem is as follows: This invention proposes a platform motion measurement method and device based on AI vision, including a data acquisition module, a processing module, and a control module; the data acquisition module includes a vision acquisition unit, a measurement unit, and a mounting base; the device also includes: A heat sink is mounted on a mounting base, and a vision acquisition unit is installed therein. A lens is provided at one end of the heat sink, which matches the data acquisition unit. A vibration damping block is provided on the inner wall of the heat sink. An inner sliding groove is provided on the inner side of the vibration damping block, and an inner sliding rod is slidably connected in the inner sliding groove by a spring. The inner sliding rod contacts the inner side of the lens. An outer sliding rod is slidably connected to the outer side of the vibration damping block. An electric linkage device is provided in the vibration damping block for linkage between the inner and outer sliding rods. The outer sliding rod is connected to an electric push rod on the vibration damping block. A rotating rod is rotatably connected to a vibration damping block. The rotating rod is connected to a drive motor mounted on the vibration damping block. A rotating frame is provided on the rotating rod, and a protective film is provided on the rotating frame. The protective film is located at one end of the lens. A cleaning box is provided on the vibration damping block, and a cleaning rotating rod is rotatably connected to one side of the cleaning box. The rotating frame is located inside the cleaning box.

[0006] Preferably, the heat dissipation box has a storage box at one end, and the storage boxes are located on both sides of the lens. A rotating shaft is rotatably connected inside the storage box. A cleaning motor is provided in the storage box above the lens. The cleaning motor is connected to the rotating shaft. A separator belt is wound on the rotating shaft. Separating grooves are evenly distributed on the separator belt. The lens is located at the center of the separating groove. A cleaning rod is slidably connected to the outer slide rod by a spring. One side of the cleaning rod contacts the rotating frame and the protective film. A top block is provided in the heat dissipation box away from the protective film. The top block contacts one side of the cleaning rod.

[0007] Preferably, the rotating frame is provided with a pressure block, which is connected to the rotating frame by a spring. A water storage bladder is provided between the pressure block and the rotating frame. One end of the water storage bladder is provided with a nozzle facing the lens, and one side of the pressure block's arc surface contacts one side of the cleaning rod.

[0008] Preferably, a roller is rotatably connected to the arc surface of the cleaning rod, and the roller contains a protective liquid.

[0009] Preferably, the rotating frame is provided with an airbag assembly, which is located on one side of the lens. The air nozzles on the airbag assembly all face the protective film, and the airbag assembly contacts the arc surface of the cleaning rod. The heat dissipation box is provided with a heating wire, which is close to the airbag assembly.

[0010] Preferably, the separating strip is provided with a wiping strip, and the wiping strip is close to the separating groove.

[0011] Preferably, the cleaning box is provided with a squeezing tube, and both ends of the squeezing tube are connected to flexible hoses. The portion of the flexible hose away from the squeezing tube is coiled around the heat sink in the visual acquisition unit. A squeezing block is slidably connected inside the squeezing tube by a spring. A squeezing rod is provided on the outer slide rod, and one side of the squeezing rod contacts the squeezing block.

[0012] Preferably, the inner wall of the extrusion tube is provided with a storage air bladder, and the storage air bladder contains refrigerant. The storage air bladder is connected to the extrusion tube through a temperature control valve at one end.

[0013] Preferably, the probe in the temperature control valve is wound around a support strip provided in the hose, the support strip is spiral-shaped, and the probe in the temperature control valve is close to the visual acquisition unit.

[0014] A platform motion measurement method based on AI vision, the measurement method comprising: S1: Acquisition: Capture a continuous sequence of images of the target platform and its surrounding environment using the acquisition module, and record the image timestamps synchronously; acquire the platform's acceleration and angular velocity data using the acquisition module to obtain real-time information on the platform's attitude changes and dynamic motion; acquire environmental parameters using auxiliary devices such as light sensors and temperature sensors; S2: Processing: Process the acquired images, including distortion correction, image registration, removal of invalid frames, and enhancement of feature point region contrast; filter high-frequency noise through filtering algorithms to correct sensor bias; use a trained AI model to perform error compensation on the data and correct the cumulative error caused by sensor noise and bias. S3: Fusion: Based on the geometric relationship and motion consistency law between feature points, a joint motion model is established; combined with physical motion law and spatiotemporal consistency constraints; Kalman filtering, deep learning fusion model and other algorithms are used to weighted fuse the processed visual feature point displacement data and posture data; S4: Compensation: Based on the fused dataset and constraint model, the six-degree-of-freedom motion parameters of the platform are calculated in real time to reconstruct the complete motion trajectory of the platform; based on the platform motion estimation results, the relative displacement measured by the visual sensor is compensated to obtain the true displacement data of the target; S5: Adaptive Adjustment: The system uses AI algorithms to identify environmental changes and automatically adjusts parameters, including camera exposure parameters, feature point detection thresholds, and sensor operating modes. Based on sensor data deviations, the system periodically triggers calibration processes to update model parameters and outputs measurement results such as platform motion trajectory, displacement, and attitude in real time.

[0015] The beneficial effects of this invention are as follows: 1. The AI ​​vision-based platform motion measurement method and device of the present invention includes a drive motor that drives a rotating rod to rotate, which in turn drives a protective film to rotate via a rotating frame. The protective film rotates out of the cleaning box, and the cleaning rod rolls over the protective film to clean it, improving the cleanliness of the protective film before installation. After rotating out, the protective film moves until it contacts the lens and stops, where the lens and the protective film come into contact and adhere to each other. By installing an anti-strong light protective film on the lens, the influence of strong light during the day on visual measurement is reduced, abnormal data is reduced, and the accuracy of motion measurement is improved. At the same time, the amount of abnormal data removed by AI is reduced, and the data processing efficiency is improved, thereby improving the measurement efficiency.

[0016] 2. The AI ​​vision-based platform motion measurement method and device described in this invention, since the protective film is stored in the cleaning box, and the opening of the cleaning box is sealed by the conventional protective film brush on the surface of the cleaning rod, the inside of the cleaning box is isolated from the outside world, preventing the protective film stored inside the cleaning box from being contaminated by external impurities; in addition, one half of the cleaning rod cleans the used protective film, and the other half cleans the stored protective film, and the cleaning rod is placed horizontally, so that the impurities removed by the cleaning rod fall away from the cleaning box, maintaining the cleanliness of the protective film and the lens. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a perspective view of the present invention; Figure 2 This is a diagram showing the inside of the cleaning box; Figure 3 This is a partial sectional view of the invention from the side view direction; Figure 4 It is a partial sectional view of the rotation axis from a top view; Figure 5 yes Figure 4 A magnified view of a portion of the image; Figure 6 It is a 3D view of the extruded tube; Figure 7 This is a partial cross-sectional view of the extruded tube; In the diagram: Acquisition module 1, heat dissipation box 11, lens 12, vibration damping block 13, inner slide groove 14, inner slide rod 15, outer slide rod 16, electric linkage device 17, electric push rod 18, rotating rod 19, drive motor 2, rotating frame 21, protective film 22, cleaning box 23, cleaning rotating rod 24, storage box 25, rotating shaft 26, cleaning motor 27, separating strip 28, separating groove 29, cleaning rod 3, top block 31, pressure block 32, water storage bladder 33, roller 34, air bladder assembly 35, wiping strip 36, squeezing tube 37, hose 38, squeezing block 39, squeezing rod 4, storage air bladder 41, support strip 42. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: To effectively solve the above problems, see the attached diagram in the instruction manual. Figure 1-7As shown, a platform motion measurement device based on AI vision includes a data acquisition module 1, a processing module, and a control module; the data acquisition module 1 includes a vision acquisition unit, a measurement unit, and a mounting base; the vision acquisition unit includes... Visual data: Using a single camera or a multi-camera array, images of the target platform and its environment are captured at a set frame rate, and timestamps are recorded synchronously. Multiple cameras use spatiotemporal synchronization technology to ensure that the shooting time is consistent. IMU data: Acquire platform acceleration and angular velocity data at a set sampling rate to supplement the lack of visual data in fast motion and low light scenarios; Environmental data: Outdoor environmental parameters are collected through light sensors, temperature and humidity sensors, and dust sensors to provide a basis for subsequent error correction; The description also includes: A heat sink 11 is mounted on a mounting base, and a vision acquisition unit is installed therein. A lens 12 is provided at one end of the heat sink 11, which is matched with the data acquisition unit. A vibration damping block 13 is provided on the inner wall of the heat sink 11. An inner sliding groove 14 is provided on the inner side of the vibration damping block 13, and an inner sliding rod 15 is slidably connected in the inner sliding groove 14 by a spring. The inner sliding rod 15 contacts the inner side of the lens 12. An outer sliding rod 16 is slidably connected to the outer side of the vibration damping block 13. An electric linkage device 17 is provided in the vibration damping block 13. The electric linkage device 17 is used for linkage between the inner sliding rod 15 and the outer sliding rod 16. The outer sliding rod 16 is connected to an electric push rod 18 on the vibration damping block 13. A rotating rod 19 is rotatably connected to a vibration damping block 13. The rotating rod 19 is connected to a drive motor 2 mounted on the vibration damping block 13. A rotating frame 21 is provided on the rotating rod 19, and a protective film 22 is provided on the rotating frame 21. The protective film 22 is located at one end of the lens 12. A cleaning box 23 is provided on the vibration damping block 13, and a cleaning rotating rod 24 is rotatably connected to one side of the cleaning box 23. The rotating frame 21 is located inside the cleaning box 23. The lens 12 is a conventional configuration installed at one end of the vision acquisition unit. The vibration damping block 13 and the heat sink 11 are connected and fixed to each other by conventional vibration damping rubber to reduce the vibration transmission of the vibration damping block 13 to the acquisition electrical components. The electric linkage device 17 is a conventional device for driving, including an electromagnet and an iron block. The electromagnet is energized and fixed on the outer slide rod 16, and the iron block is fixed on the inner slide rod 15. The gap between the inner slide rod 15 and the outer slide rod 16 is small and there is no magnetic shielding effect, so that the outer slide rod 16 drives the inner slide rod 15 to move synchronously in the slide groove by attracting the iron block through the electromagnet. That is, the reciprocating movement of the inner slide rod 15 can be achieved by the extension and retraction of the electric push rod 18. The protective film 22 is a conventional lens 12 film, and there are various types of protective films stored for the protective film 22, including protective films 22 for reducing direct sunlight at noon. Specific workflow: During daytime measurements, drive motor 2 drives rotating rod 19 to rotate. Rotating rod 19 drives protective film 22 to rotate via rotating frame 21. Protective film 22 rotates out from cleaning box 23. Cleaning rod 24 rolls over protective film 22 to clean it, improving the cleanliness of protective film 22 before installation. After rotating out, protective film 22 moves to contact lens 12 and stops. Lens 12 and protective film 22 come into contact and adhere to each other. By installing anti-strong light protective film 22 on lens 12, the impact of strong light on visual measurement during the day is reduced, abnormal data is reduced, and the accuracy of motion measurement is improved. At the same time, the amount of abnormal data removed by AI is reduced, and the data processing efficiency is improved, thereby improving measurement efficiency. At night, the protective film 22 is rotated in the manner described above, transferring the protective film 22 used during the day into the cleaning box 23, while the protective film 22 used at night, including high-definition types, is transferred out. The protective film 22 used at night is then installed on the lens 12 in the manner described above to complete the motion measurement work. By automatically changing different types of protective films 22 according to the visual detection needs of day and night, the measurement accuracy of the measuring device throughout the day is improved, thereby improving the precision of the measurement results. Different types of protective films 22, such as light brown and light gray, in addition to resisting strong light, can also specifically optimize the color reproduction of different outdoor scenes: for example, light brown protective film 22 can improve the color contrast between road markings and backgrounds in rainy weather, helping AI algorithms to more accurately identify colored signs. Furthermore, when the protective film 22 used during the day is replaced, its surface is covered with many impurities due to prolonged exposure to the outside. When the protective film 22 used during the day passes through the cleaning rod 24, the cleaning rod 24 wipes away the impurities on its surface, improving cleanliness and preventing it from affecting the next use. Moreover, since the protective film 22 is stored in the cleaning box 23, the opening of the cleaning box 23 is sealed by the regular protective film 22 brush on the surface of the cleaning rod 24, thus isolating the inside of the cleaning box 23 from the outside world and preventing the protective film 22 stored inside the cleaning box 23 from being contaminated by external impurities. In addition, one half of the cleaning rod 24 cleans the used protective film 22, and the other half cleans the stored protective film 22. The cleaning rod 24 is horizontal, so that the impurities removed by the cleaning rod 24 fall away from the cleaning box 23, maintaining the cleanliness of the protective film 22 and the lens 12. Furthermore, at night, if the temperature difference between day and night is large, the electrical components inside the heat dissipation box 11 will generate heat, causing a temperature difference between the inner and outer surfaces of the lens 12, resulting in water mist on the inner surface of the lens 12. At this time, the electric push rod 18 is activated to move the outer slide rod 16. The outer slide rod 16 drives the inner slide rod 15 to move synchronously through the electric linkage device 17. The inner slide rod 15 moves back and forth across the lens 12, so that the lens 12 can complete the water mist removal work while the heat dissipation box 11 is kept sealed, avoiding water mist from affecting the visual measurement work, thereby improving the accuracy of motion measurement and thus improving the practicality of the measuring device. Furthermore, in addition to wiping away water mist, the reciprocating movement of the inner slide bar 15 can dry any residual moisture on the inside of the lens 12, such as residual purified water after cleaning or trace amounts of water stains after the water mist evaporates. This prevents water stains from forming watermarks that affect light transmittance, ensuring image clarity and reducing interference from AI feature point extraction. In outdoor environments, the lens 12 is easily scratched by wind, sand, gravel, and rain, or corroded by acid rain and dust. The protective film 22, as a protective layer for the lens 12, can directly withstand these external damages, preventing damage to the lens 12 itself and reducing the cost of replacing the lens 12. Moreover, during the replacement of the lens 12, one side of the rotating frame 21 blocks the surface of the lens 12, preventing external contact with the lens 12, reducing the exposed area of ​​the lens 12, and preventing dust from getting on the lens 12, thereby improving the cleanliness of the lens 12 and thus improving the accuracy of visual measurement. In addition, the cleaning box 23 is located above the lens 12, so that when strong light shines downwards, the cleaning box 23 can block the strong light, improving the accuracy of visual measurement.

[0021] Example 2: Based on Embodiment 1, a storage box 25 is provided at one end of the heat dissipation box 11, and the storage boxes 25 are located on both sides of the lens 12. A rotating shaft 26 is rotatably connected inside the storage box 25. A cleaning motor 27 is provided in the storage box 25 above the lens 12. The cleaning motor 27 is connected to the rotating shaft 26. A separator strip 28 is wound on the rotating shaft 26. Separating grooves 29 are evenly distributed on the separator strip 28. The lens 12 is located at the center of the separating grooves 29. A cleaning rod 3 is slidably connected to the outer slide rod 16 by a spring. The side contact rotating frame 21 and protective film 22, the heat dissipation box 11 is provided with a top block 31 at a position away from the protective film 22, and the top block 31 contacts one side of the cleaning rod 3; the separator 28 is a strip tool commonly used for cleaning the lens 12, for example, a long strip with a flannel cloth on the inner surface for gently cleaning the lens 12; the cleaning motor 27 can also be mounted on the vibration damping block 13 to reduce the vibration transmitted to the heat dissipation box 11 when the cleaning motor 27 starts; the rotating shaft 26 located above the lens 12 is called the upper rotating shaft 26, and the other is called the lower rotating shaft 26; The rotating frame 21 is provided with a pressure block 32, which is connected to the rotating frame 21 by a spring. A water storage bladder 33 is provided between the pressure block 32 and the rotating frame 21. One end of the water storage bladder 33 is provided with a nozzle facing the lens 12. One side of the pressure block 32 is in contact with one side of the cleaning rod 3. A roller 34 is rotatably connected to the arc surface of the cleaning rod 3, and the roller 34 contains a protective liquid. The rotating frame 21 is equipped with an airbag assembly 35, which is located on one side of the lens 12. The air nozzles on the airbag assembly 35 are all facing the protective film 22. The airbag assembly 35 contacts the arc surface of the cleaning rod 3. The heat dissipation box 11 is equipped with a heating wire, which is close to the airbag assembly 35. The separator 28 is provided with a wiping strip 36, and the wiping strip 36 is close to the separator groove 29; Specific workflow: When replacing the protective film 22, the cleaning motor 27 is started to drive the upper rotating shaft 26 to rotate. The upper rotating shaft 26 winds up the separator belt 28. The inner wall of the separator groove 29 in the separator belt 28 gradually approaches and inserts between the protective film 22 and the lens 12 until the wind guard and the lens 12 are separated. At this time, the separator belt 28 rotates to a position where the separator groove 29 is far away from the lens 12, which plays a role in shielding and protecting the lens 12 during the replacement of the protective film 22, thereby improving the protection level of the lens 12, improving the cleanliness of the lens 12, and thus improving the measurement accuracy of the measuring device. When the separator strip 28 covers the lens 12, the drive motor 2 unlocks from the fixed state and drives the rotating frame 21 to start rotating from the fixed state. The used protective film 22 is transferred into the cleaning box 23, and the stored protective film 22 is transferred out of the cleaning box 23. When the protective film 22 moves to one side of the lens 12, the cleaning motor 27 once again winds up the separator strip 28 through the rotating shaft 26. The separator strip 28 passes through the lens 12 and the protective film 22, removing and carrying away the impurities on the surface of both, thus improving the cleanliness. When the center of the separator groove 29 moves to the center of the lens 12, the lens 12 contacts the protective film 22. The electric push rod 18 drives the cleaning rod 3 to move through the outer slide rod 16. When the cleaning rod 3 is stored, it is located on the top block 31 and is far away from the rotating frame 21, so as not to affect the rotation of the rotating frame 21. When the cleaning rod 3 moves away from the top block 31, the cleaning rod 3 is reset to contact the rotating frame 21 by the spring. The cleaning rod 3 continues to move and contacts the protective film 22, squeezing the protective film 22 onto the lens 12. The cleaning rod 3 scrapes from one end of the lens 12 to the other end, squeezing out the air bubbles between the lens 12 and the protective film 22, improving the adhesion of the protective film 22, thereby improving the effect of the protective film 22 and thus improving the accuracy of the measuring device. Furthermore, when replacing the protective film 22, the linkage between the outer slide bar 16 and the inner slide bar 15 allows the inner slide bar 15 to clean the inside of the lens 12 during the film replacement process, reducing the time that subsequent additional cleaning would obstruct visual measurement and improving cleaning efficiency, thereby improving the efficiency of visual measurement. Moreover, by setting up the storage box 25, the separator 28 is sealed and stored, preventing the separator 28 from being contaminated by impurities or rainwater, thereby improving the cleanliness of the separator 28, and thus improving the cleaning effect of the separator 28 and improving the clarity of the lens 12. As the cleaning rod 3 moves, its curved surface forces the pressure block 32 into the rotating frame 21. The pressure block 32 then compresses the water reservoir 33 within the rotating frame 21, causing it to spray water between the lens 12 and the protective film 22. This water flow washes away impurities from both surfaces, further improving cleanliness. After the water is sprayed, the water flows away from the lens 12 and the protective film 22 due to gravity. The moving cleaning rod 3 presses the protective film 22 onto the lens 12. At this point, both the rotating frame 21 and the lens 12 are fixed. The movement of the cleaning rod 3 then presses the protective film 22 onto the lens 12. Water between the protective film 22 and the lens is scraped away. The water flows and carries away the fine dust and oil on the surface, reducing dust particles trapped between the film and the lens. This avoids scratching the lens coating and allows the film to fully adhere to the curved surface of the lens, ensuring uniform light transmission. Water slowly drains from between the protective film 22 and the lens, removing the air between them. This prevents air from being trapped under the film during dry application, thus avoiding excessive air bubbles that could cause light refraction, blurred images, and affect the feature point recognition of the visual sensor, thereby improving the accuracy of visual measurements. Furthermore, by using water as an auxiliary bonding medium between the protective film 22 and the lens 12, no strong adhesive force is generated between the two when the protective film 22 separates from the lens 12, thus preventing the protective film 22 from being torn and protecting it, thereby improving the recycling rate. In addition, during separation, the insertion of the separator 28 assists in the separation of the protective film 22 from the lens 12, improving the protection of the protective film 22 and also promptly removing stains and other contaminants that occur when the protective film 22 and the lens 12 separate from each other, thereby improving cleanliness. By setting the roller 34, the sliding friction between the cleaning rod 3 and the rotating frame 21 and pressure block 32 is changed to rolling friction, reducing the impact of friction-generated vibration on the electrical appliances. Furthermore, the cleaning rod 3 rolls over the protective film 22 through the roller 34, preventing the protective film 22 from being damaged or shifted due to the cleaning rod 3 scraping, thereby improving the adhesion of the protective film 22 and thus improving the accuracy of visual measurement. In addition, the protective liquid inside the roller 34 is a conventional antistatic solution, and the roller 34 contains micropores and other holes for slow outflow. When the roller 34 rolls over the protective film 22, the roller 34 evenly coats the outer surface of the protective film 22 with the protective liquid, reducing friction during cleaning and preventing the generation of static electricity that attracts dust, thereby improving the accuracy of visual measurement. By setting up an airbag assembly 35, which is composed of multiple airbag assemblies 35, with the air nozzles on the airbag assembly 35 all facing the protective film 22, when the cleaning rod 3 passes through the protective film 22, the cleaning rod 3 squeezes the airbag assembly 35, and the airbag assembly 35 is squeezed to spray air onto the surface of the protective film 22, blowing away the impurities attached to the surface and improving the cleanliness. In addition, by setting up a heating wire, when the temperature is low at night, the heating wire is activated to heat the airbag assembly 35, which raises the temperature of the air blown onto the protective film 22, heating the protective film 22, making the protective film 22 softer and improving its adhesion. Furthermore, the increased temperature of the protective film 22 can also increase the temperature when the protective liquid is applied to the protective film 22, which helps the protective liquid form a film layer on the protective film 22 and improves the effect. Furthermore, as the separator 28 moves, the separator 28 carries a soft wiping strip 36 past the lens 12 and the protective film 22. The wiping strip 36 removes impurities and stains, improving the cleaning effect, thereby improving the clarity of the lens 12, improving the accuracy of visual measurement, and thus improving the measurement accuracy of the measuring device.

[0022] Example 3: Based on Embodiment 2, the cleaning box 23 is provided with a squeezing tube 37, and both ends of the squeezing tube 37 are connected to hoses 38. The portion of the hoses 38 away from the squeezing tube 37 is coiled around the heat sink in the visual acquisition unit. A squeezing block 39 is slidably connected inside the squeezing tube 37 by a spring. A squeezing rod 4 is provided on the outer slide rod 16, and one side of the squeezing rod 4 contacts the squeezing block 39. One-way valves are provided at both ends of the squeezing tube 37 and the hoses 38. The inner wall of the extrusion tube 37 is provided with a storage air bag 41, and the storage air bag 41 contains refrigerant. The storage air bag 41 is connected to the extrusion tube 37 through a temperature control valve at one end. The probe in the temperature control valve is wound around the support bar 42 provided in the hose 38. The support bar 42 is spiral-shaped, and the probe in the temperature control valve is close to the vision acquisition unit. Specific workflow: During the day when heat dissipation is required, the outer slide bar 16 moves back and forth without contacting the protective film 22, and at this time the electric linkage device 17 is not working. The outer slide bar 16 moves independently, and the outer slide bar 16 drives the squeezing rod 4 to reciprocate once. The squeezing rod 4 squeezes the squeezing block 39 into the squeezing tube 37 once. The squeezing tube 37 delivers water to the hose 38 through the one-way valve at one end. After the squeezing rod 4 moves away from the squeezing block 39, the squeezing block 39 slides back to its original position due to the influence of the spring. A negative pressure is formed in the squeezing tube 37, which draws water into the hose 38 through the one-way valve at the other end, realizing the water circulation effect in the hose 38. The hose 38 contacts the heat sink in the vision acquisition unit and transfers heat out, achieving the purpose of cooling the sealed heat dissipation box 11, reducing the temperature inside the heat dissipation box 11, and avoiding excessive temperature difference between the inside and outside of the lens 12 at night, which may cause a large amount of water mist or water droplets to appear, thereby improving the clarity of the lens 12 and thus improving the accuracy of vision measurement. By setting up a storage airbag 41, when the extrusion block 39 is squeezed into the extrusion tube 37, the extrusion block 39 squeezes the storage airbag 41. If the temperature control valve detects that the temperature of the circulating water in the hose 38 is too high, it will connect the storage airbag 41 and the extrusion tube 37 in one direction. The refrigerant in the storage airbag 41, such as saltpeter, will change the dielectric constant of water when the saltpeter dissolves in water. During the process of the change in dielectric constant, the interaction between molecules and the energy state change, and energy needs to be absorbed to lower the water temperature, thereby lowering the temperature of the circulating water, thereby improving the heat dissipation effect and reducing the degree of fogging of the lens 12. By setting up the support bar 42, the probe is brought close to the vision acquisition unit via the support bar 42, which can directly sense the real-time temperature of the core heat-generating component. This avoids the deviation between the ambient temperature and the component temperature caused by traditional long-distance temperature measurement. For example, when the outdoor temperature is high, the external temperature of the heat sink 11 is lower than the actual temperature of the vision acquisition unit, ensuring the accuracy of the temperature data and thus improving the cooling effect. In addition, the spiral support bar 42 provides a stable winding carrier for the probe, preventing the probe from shifting due to rotation during measurement, following the target movement, or the flow of the temperature control medium, thus ensuring that the probe is always close to the vision acquisition unit and maintaining the temperature measurement accuracy. At the same time, the support bar 42 can isolate the probe from direct friction with the inner wall of the hose 38, reducing probe wear and extending its service life. Moreover, the spiral structure can guide the temperature control medium in the hose 38 to form a spiral flow, prolonging the residence time of the medium around the visual acquisition unit, improving heat exchange efficiency, and avoiding insufficient heat exchange caused by the direct entry and exit of the temperature control medium; in addition, the support bar 42 itself has a certain degree of elasticity, which can buffer the impact of outdoor vibration on the probe, while reducing the impact of liquid impact generated when the temperature control medium flows on the probe, further ensuring the stability of temperature measurement data.

[0023] Example 4: A platform motion measurement method based on AI vision, the measurement method comprising: S1: Acquisition: Capture a continuous sequence of images of the target platform and its surrounding environment using acquisition module 1, and record the image timestamps synchronously; acquire acceleration and angular velocity data of the platform using acquisition module 1 to obtain real-time information on platform attitude changes and dynamic motion; acquire environmental parameters using auxiliary devices such as light sensors and temperature sensors; S2: Processing: Process the acquired images, including distortion correction, image registration, removal of invalid frames, and enhancement of feature point region contrast; filter high-frequency noise through filtering algorithms to correct sensor bias; use a trained AI model to perform error compensation on the data and correct the cumulative error caused by sensor noise and bias. S3: Fusion: Based on the geometric relationship and motion consistency law between feature points, a joint motion model is established; combined with physical motion law and spatiotemporal consistency constraints; Kalman filtering, deep learning fusion model and other algorithms are used to weighted fuse the processed visual feature point displacement data and posture data; S4: Compensation: Based on the fused dataset and constraint model, the six-degree-of-freedom motion parameters of the platform are calculated in real time to reconstruct the complete motion trajectory of the platform; based on the platform motion estimation results, the relative displacement measured by the visual sensor is compensated to obtain the true displacement data of the target; S5: Adaptive Adjustment: The system uses AI algorithms to identify environmental changes and automatically adjusts parameters, including camera exposure parameters, feature point detection thresholds, and sensor operating modes. Based on sensor data deviations, the system periodically triggers calibration processes to update model parameters and outputs measurement results such as platform motion trajectory, displacement, and attitude in real time.

[0024] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A platform motion measurement device based on AI vision, comprising a data acquisition module (1), a processing module, and a control module; the data acquisition module (1) includes a vision acquisition unit, a measurement unit, and a mounting base; characterized in that, The description also includes: A heat sink (11) is mounted on a mounting base. A lens (12) is provided at one end of the heat sink (11). A damping block (13) is provided on the inner wall of the heat sink (11). An inner groove (14) is provided on the inner side of the damping block (13), and an inner slide rod (15) is slidably connected in the inner groove (14) by a spring. An outer slide rod (16) is slidably connected on the outer side of the damping block (13). An electric linkage device (17) is provided in the damping block (13), and the outer slide rod (16) is connected to an electric push rod (18) on the damping block (13). A rotating rod (19) is rotatably connected to a damping block (13). The rotating rod (19) is connected to a drive motor (2) installed on the damping block (13). A rotating frame (21) is provided on the rotating rod (19), and a protective film (22) is provided on the rotating frame (21). The protective film (22) is located at one end of the lens (12). A cleaning box (23) is provided on the damping block (13), and a cleaning rotating rod (24) is rotatably connected to one side of the cleaning box (23). The rotating frame (21) is located inside the cleaning box (23).

2. The platform motion measurement device based on AI vision according to claim 1, characterized in that: The heat dissipation box (11) has a storage box (25) at one end, and the storage box (25) is located on both sides of the lens (12). The storage box (25) is rotatably connected to a rotating shaft (26). The storage box (25) located above the lens (12) is equipped with a cleaning motor (27). The cleaning motor (27) is connected to the rotating shaft (26). A separator strip (28) is wound on the rotating shaft (26). Separator grooves (29) are evenly distributed on the separator strip (28). The lens (12) is located at the center of the separator groove (29). A cleaning rod (3) is slidably connected to the outer slide rod (16) by a spring. One side of the cleaning rod (3) contacts the rotating frame (21) and the protective film (22). The heat dissipation box (11) is provided with a top block (31) at a position away from the protective film (22). The top block (31) contacts one side of the cleaning rod (3).

3. The platform motion measurement device based on AI vision according to claim 2, characterized in that: The rotating frame (21) is provided with a pressure block (32), which is connected to the rotating frame (21) by a spring. A water storage bladder (33) is provided between the pressure block (32) and the rotating frame (21). One end of the water storage bladder (33) is provided with a nozzle facing the lens (12), and one side of the pressure block (32) is in contact with one side of the cleaning rod (3).

4. The platform motion measurement device based on AI vision according to claim 3, characterized in that: The cleaning rod (3) has a roller (34) rotatably connected to its arc surface, and the roller (34) contains a protective liquid.

5. The platform motion measurement device based on AI vision according to claim 4, characterized in that: The rotating frame (21) is provided with an airbag assembly (35), and the airbag assembly (35) is located on one side of the lens (12). The air nozzles on the airbag assembly (35) are all facing the protective film (22). The airbag assembly (35) contacts the arc surface of the cleaning rod (3). The heat dissipation box (11) is provided with a heating wire, and the heating wire is close to the airbag assembly (35).

6. The platform motion measurement device based on AI vision according to claim 5, characterized in that: The separator (28) is provided with a wiping strip (36), and the wiping strip (36) is close to the separator groove (29).

7. The platform motion measurement device based on AI vision according to claim 1, characterized in that: The cleaning box (23) is provided with a squeezing tube (37), and the two ends of the squeezing tube (37) are connected to a hose (38). The part of the hose (38) away from the squeezing tube (37) is coiled on the heat sink in the visual acquisition unit. A squeezing block (39) is slidably connected inside the squeezing tube (37) by a spring. A squeezing rod (4) is provided on the outer slide rod (16), and one side of the squeezing rod (4) contacts the squeezing block (39).

8. The platform motion measurement device based on AI vision according to claim 7, characterized in that: The inner wall of the extrusion tube (37) is provided with a storage air bladder (41), and the storage air bladder (41) contains refrigerant. The storage air bladder (41) is connected to the extrusion tube (37) through a temperature control valve at one end.

9. The platform motion measurement device based on AI vision according to claim 8, characterized in that: The probe in the temperature control valve is wound around a support strip (42) set in a hose (38). The support strip (42) is spiral-shaped, and the probe in the temperature control valve is close to the visual acquisition unit.

10. A platform motion measurement method based on AI vision, wherein the measurement method uses the measuring device according to any one of claims 1-9, characterized in that: The measurement method includes: S1: Acquisition: The acquisition module (1) captures a continuous sequence of images of the target platform and its surrounding environment, and records the image timestamps synchronously; the acquisition module (1) is used to acquire the platform's acceleration and angular velocity data, and to obtain the platform's attitude change and dynamic motion information in real time; environmental parameters are acquired through auxiliary equipment such as light sensors and temperature sensors; S2: Processing: Process the acquired images, including distortion correction, image registration, removal of invalid frames, and enhancement of feature point region contrast; filter high-frequency noise through filtering algorithms to correct sensor bias; use a trained AI model to perform error compensation on the data and correct the cumulative error caused by sensor noise and bias. S3: Fusion: Based on the geometric relationship and motion consistency law between feature points, a joint motion model is established; combined with physical motion law and spatiotemporal consistency constraints; Kalman filtering, deep learning fusion model and other algorithms are used to weighted fuse the processed visual feature point displacement data and posture data; S4: Compensation: Based on the fused dataset and constraint model, the six-degree-of-freedom motion parameters of the platform are calculated in real time to reconstruct the complete motion trajectory of the platform; based on the platform motion estimation results, the relative displacement measured by the visual sensor is compensated to obtain the true displacement data of the target; S5: Adaptive Adjustment: The system uses AI algorithms to identify environmental changes and automatically adjusts parameters, including camera exposure parameters, feature point detection thresholds, and sensor operating modes. Based on sensor data deviations, the system periodically triggers calibration processes to update model parameters and outputs measurement results such as platform motion trajectory, displacement, and attitude in real time.

Images