Water film online detection method, system and equipment

Abstract

The invention relates to the technical field of industrial visual inspection and intelligent assembly, solves the technical problem of low accuracy and reliability in a special detection environment in the prior art, and particularly relates to a water film online detection method, system and equipment, which comprises the following steps of: acquiring an image at the frequency of 1Hz, and dynamically comparing the image with a previous image; detecting whether the water film has a broken sign and marking; detecting a rupture image with a rupture sign of the water film; and continuously collecting the image and comparing the image with the fracture image to judge whether the water film fracture phenomenon continuously exists or not. According to the method, multiple processes and multiple numerical parameters can be comprehensively considered, full-process and multi-numerical-parameter water film damage detection is realized, and the assembly accuracy is ensured. And meanwhile, the technical blank of non-contact and in-situ dynamic water film monitoring is filled, and reliable data support is provided for preventive maintenance in a high-risk industrial scene.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial visual inspection and intelligent assembly, and in particular to a water film online detection method, system and equipment. Background Art

[0002] In industrial production, the integrity of the water film on the surface of materials is one of the key indicators for evaluating corrosion protection, coating performance, and welding quality. For example, if the water film on the surface protective layer of equipment such as nuclear power plant pressure vessels and chemical pipelines ruptures in acidic or high-temperature and high-pressure environments, it may lead to accelerated local corrosion and threaten the safety of the equipment. Although traditional detection methods such as water film ultrasonic flaw detection systems can improve the penetration of sound waves through coupling agents and accurately identify internal cracks and pores (with a sensitivity of 50-100μm), they rely on manual operation of the probe positioning and cannot monitor the dynamic rupture process of the water film in real time.

[0003] For the detection of surface water film status, existing technologies mostly use static water film rupture test (Water Break Test), which indirectly evaluates cleanliness by observing the wettability of distilled water on the substrate surface. However, this method relies on manual visual judgment, is easily affected by operating experience, and lacks the ability to record quantitative data and dynamic tracking. Some industrial scenarios try to introduce mechanical tests (such as ASTM D1004 tear strength test) to evaluate the durability of membrane materials, but destructive sampling is required and in-situ continuous monitoring cannot be achieved.

[0004] As the demand for industrial automation increases, dynamic detection technologies based on machine vision are increasingly being applied to surface defect identification. For example, endoscopic camera systems are used to detect water supply network leaks in real time to capture pipe cracks, and dam crack monitoring is combined with underwater imaging technology for remote diagnosis. Such technologies offer new insights into water film rupture detection: high-frame-rate image acquisition and comparison algorithms can capture micron-level water film deformation characteristics, and combined with mobile robotic arm positioning, efficient multi-area scanning is achieved. However, existing systems still face two major challenges:

[0005] Insufficient environmental adaptability: The sensor is easily corroded in acidic or high-humidity environments, requiring the design of sealed protection and rapid retraction mechanisms;

[0006] High false positive rate: Instantaneous bubbles or light interference can easily trigger false alarms, requiring the integration of multi-frame image analysis and delayed verification to improve reliability. Summary of the Invention

[0007] In view of the shortcomings of the existing technology, the present invention provides a water film online detection method, system and equipment, which solve the technical problem of low accuracy and reliability of the existing technology under special detection environments.

[0008] To solve the above technical problems, the present invention provides the following technical solution: a water film online detection method, the method comprising the following steps:

[0009] S1. Move the inspection vehicle to the inspection position of the workpiece to be tested. After the sensor detects the workpiece to be tested, start the image acquisition program to collect images at a frequency of 1 Hz and perform dynamic comparison with the previous image.

[0010] S2, continuously collecting images and comparing and analyzing them with the previous image to detect whether the water film shows signs of rupture and mark them;

[0011] S3, detecting the rupture image of the water film showing signs of rupture, and sequentially storing and recording the rupture image, the time, location and area of the rupture;

[0012] S4, continue to collect images and compare them with the rupture image to determine whether the water film rupture phenomenon persists; if so, confirm the authenticity of the rupture phenomenon and proceed to step S5; if not, re-analyze the image and return to step S2 to determine again whether it is a false positive;

[0013] S5. After the inspection is completed, the inspection trolley is moved into the box cabin and vacuum extraction is performed to prevent further corrosion of the equipment.

[0014] Furthermore, the dynamic comparison includes: using the OpenCV difference algorithm to calculate the pixel difference between adjacent images in real time. If the difference value exceeds a threshold, it is marked as "detection abnormality" and an alarm is triggered.

[0015] Furthermore, in step S2, the specific process includes the following steps:

[0016] S21, sequentially comparing the current image acquired with the previous image, and determining abnormal images having differences or feature changes among the images;

[0017] S22, performing threshold processing on the grayscale value or color distribution of the abnormal image to identify possible crack or rupture areas;

[0018] S23. Marking the abnormal image, including the boundary box, time, and location of the crack or rupture area.

[0019] Furthermore, in step S21, the abnormal image is determined by using any one of pixel-level difference comparison, image feature extraction comparison, or structural similarity index, wherein:

[0020] Pixel-level difference comparison: Compare the pixel values of the two images and calculate the difference value D(x,y) between the pixels. If the difference value D(x,y) exceeds the preset threshold, it indicates that the water film has changed or broken, that is:

[0021] D(x,y)=|I1(x,y)-I2(x,y)|

[0022] Where I1(x,y) and I2(x,y) are the pixel values corresponding to the current image and the previous image respectively, where (x,y) is the pixel coordinate;

[0023] Image feature extraction and comparison: By extracting key feature points from the image and comparing the changes in key feature points between the current image and the previous image, it is possible to determine whether the water film has broken or deformed.

[0024] Structural Similarity Index: The lower the value of the SSIM, the greater the difference between the two images, which means that the water film has broken or deformed. The calculation formula of the SSIM(x,y) value is:

[0025]

[0026] Where μ x ,μ y are the means of the two images respectively; σ x ,σ y are their variances; σ xy is their covariance; C1 and C2 are constants to avoid division by zero errors.

[0027] Furthermore, the key feature points include edge points and corner points in the image.

[0028] Furthermore, in step S4, it includes:

[0029] Continue to collect and compare images, and continue monitoring within 2 seconds after the rupture is detected. If the rupture area continues to appear and changes consistently in several consecutive frames of images, it means that the rupture is persistent. Otherwise, it may be a misjudgment.

[0030] Furthermore, in step S4, the specific process includes:

[0031] S41. Use YOLOv8 network model instance segmentation to obtain water film image I t The water film boundary contour in

[0032] S42. Calculate the given water film boundary contour C t The local curvature of the i-th point in is calculated as:

[0033]

[0034] Among them, I t is the water film image (input image) of the tth frame; (x′ i ,y′ i) is the tangent direction of the boundary point; (x″ i ,y″ i ) is the second-order derivative of the boundary point, which represents the curvature of the boundary; κ i Reflects the local curvature change of the water film boundary; C t is the boundary contour of the water film in the tth frame;

[0035] S43, all local curvatures κ i Normalized to get the curvature distribution P(κ i ), the expression is:

[0036]

[0037] Where N is the total number of pixels on the water film boundary; P(κ) represents the distribution of different curvature values;

[0038] S44, using the information entropy formula according to the curvature distribution P(κ i ) Calculate the water film boundary perturbation entropy H that reflects the complexity of the water film boundary morphology t , the calculation formula is:

[0039]

[0040] If H t The continuous increase indicates that the water film boundary is changing dynamically, indicating that the rupture phenomenon persists;

[0041] If H t Small or decreasing fluctuations indicate that the water film has stabilized and may be a misjudgment;

[0042] S45, using the dynamic time warping algorithm to calculate the time used to represent the previous frame C t With the previous frame C t-1 The dynamic time warping distance D between similarities t , the calculation formula is:

[0043] D t =DTW(C t ,C t-1 )

[0044] If D t is larger, indicating that the water film morphology changes dramatically;

[0045] If D t is smaller, indicating that the water film boundary is more stable;

[0046] Among them, C t is the water film boundary contour of the tth frame; D t is the dynamic time warping distance between the t-th frame and the previous frame; DTW is used to calculate the similarity of two time series;

[0047] S46, based on dynamic time warping distance D t Define the time series stability factor S t , the expression is:

[0048]

[0049] If S t >T S , the rupture persists;

[0050] If S t <T S , then the rupture area is unstable and may be a misjudgment;

[0051] Among them, T S To set the threshold, the value needs to be determined through experiments based on the actual target, and this embodiment does not impose any limitation; N is the number of dynamic time warping distances;

[0052] S47, according to the water film boundary perturbation entropy H t and the time series stability factor S t Continuously detect the authenticity of water film rupture;

[0053] If H t >T H And S t >T S , it means that the water film rupture phenomenon persists at multiple time points and is confirmed to be rupture;

[0054] If H t <T H or S t <T S , it means that the fluctuation in the water film rupture area is small, which may be environmental noise or false detection, and needs to be re-analyzed.

[0055] A system for implementing the above-mentioned water film online detection method comprises:

[0056] The dynamic comparison module moves the inspection vehicle to the inspection position of the workpiece to be tested. After the sensor detects the workpiece to be tested, it starts the image acquisition program to collect images at a frequency of 1Hz and dynamically compares them with the previous image.

[0057] The rupture sign detection module continuously collects images and compares and analyzes them with the previous image to detect and mark whether the water film shows signs of rupture;

[0058] The rupture detection module detects the rupture image of the water film showing signs of rupture, and sequentially stores and records the rupture image, the time, location and area of the rupture;

[0059] The continuous detection module continues to collect images and compares them with the rupture images to determine whether the water film rupture phenomenon persists. If so, the authenticity of the rupture phenomenon is confirmed; if not, the image is re-analyzed to determine whether it is a false positive.

[0060] Reset the module. After the inspection is completed, move the inspection cart into the box cabin and perform vacuum extraction to prevent further corrosion of the equipment.

[0061] A device for implementing the above-mentioned water film online detection method comprises:

[0062] processor;

[0063] Memory;

[0064] robotic arm;

[0065] Box cockpit;

[0066] Detection control cabinet;

[0067] and one or more programs, wherein the one or more programs are stored in a memory and configured to be executed by the processor, and the programs are used for a computer to execute the water film online detection method.

[0068] By means of the above technical solution, the present invention provides a method, system and device for online water film detection, which have at least the following beneficial effects:

[0069] 1. This invention comprehensively considers multiple processes and multiple numerical parameters to detect water film damage across all processes and multiple numerical parameters, ensuring assembly accuracy. It also fills the technological gap in non-contact, in-situ dynamic water film monitoring, providing reliable data support for preventive maintenance in high-risk industrial scenarios.

[0070] 2. This invention adopts a multi-period cross-validation mechanism to significantly improve the accuracy of rupture determination. It integrates machine vision and deep learning algorithms to break through the limitations of traditional contact detection. The automated process increases detection efficiency by several times and improves detection accuracy to reduce rework losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0072] Figure 1 Flowchart of the water film online detection method of the present invention;

[0073] Figure 2 This is a schematic structural diagram of the box cabin in the present invention;

[0074] Figure 3 This is a schematic diagram of the detection vehicle in the box cabin in the present invention. DETAILED DESCRIPTION

[0075] To make the above-mentioned objectives, features, and advantages of the present invention more clearly understood, the present invention is further described below in detail with reference to the accompanying drawings and specific embodiments. This will enable a full understanding of how this application uses technical means to solve technical problems and achieve technical effects, and to implement the invention accordingly.

[0076] This example refers to the quantitative definition of material fracture behavior in ISO 34-1 and ASTM D1004, and establishes a fracture determination model based on image grayscale change and edge detection; draws on ultrafiltration membrane pool sealing technology and electromagnetic crack detection devices to optimize the equipment's corrosion resistance structure and signal anti-interference capabilities; and combines the positioning logic of the in-pipeline inspection robot to achieve full-process closed-loop control of mobile vehicle coordinate calibration, image acquisition, and result feedback. Please refer to Figure 1-Figure 3 This embodiment proposes an online water film detection method, which provides a new idea for complex assembly tasks in the field of intelligent assembly. It comprehensively considers multiple processes and multiple data parameters to achieve comprehensive detection and screening of water films in all processes and multiple data parameters, ensuring the integrity of water film quality and solving the problem of low accuracy and reliability of existing technologies in special detection environments. Please refer to Figure 1 , the method comprises the following steps:

[0077] S1. The inspection cart moves to the inspection position of the workpiece to be tested. After the sensor detects the workpiece, the image acquisition program starts to capture images at a frequency of 1Hz and dynamically compares them with the previous image. In this embodiment, the operator must use mechanical guide rails to push the inspection cart to the precise inspection position of the workpiece to be tested, ensuring that the inspection area is well-lit and free of obstructions and reflective interference. Adjustments must rely on the guide rail scale or laser-assisted positioning system to ensure a positioning error of ≤±0.5mm (in compliance with ASTM F22-2013).

[0078] Refer to the AGV operating instructions to visually inspect the inspection path for obstacles and clean any dust or oil from the workpiece surface to avoid affecting image acquisition quality. The external power supply uses a magnetic connector. The operator must confirm that the plug is fully attached to the vehicle connector. At this point, the power indicator will turn from red to green, indicating that the system is receiving stable power. No additional wiring or configuration is required. Refer to the startup process for automotive GPS detectors. Once powered on, the system automatically loads the preset program. If the indicator light is abnormal (such as flashing red), check the power adapter or connector for poor contact.

[0079] After the system starts, it will automatically perform a self-test procedure. The operator needs to monitor the following status through the control panel:

[0080] Infrared probes, cameras, and other devices must return to the "Ready" state. Refer to the probe detection logic of the infrared tracking car to ensure that the sensor sensitivity meets the preset threshold.

[0081] After the OpenCV image processing module is initialized, the screen displays the real-time image. The operator needs to confirm the image clarity and focus to avoid blur or distortion.

[0082] Specifically, for real-time sensor detection and automatic triggering, the inspection cart requires photoelectric sensors (such as infrared or ultrasonic sensors) installed on both sides of the immersion tank to detect the presence of target objects in real time. The sensors should cover the entire immersion tank area, and the detection distance should be adjusted according to the tank depth (recommended accuracy ≤ 1 cm). Detection signals are transmitted to the main control system via a PLC or embedded controller.

[0083] When the sensor detects an object, it sends a high-level signal to the image acquisition program, triggering the system to enter the detection process. Upon receiving the sensor signal, the image acquisition program immediately activates a high-speed industrial camera (such as a global shutter CMOS camera) to capture images at a 1Hz frequency and implement dynamic comparison algorithms.

[0084] The camera resolution must be ≥1080P and equipped with an anti-reflective lens. Each frame of the image is saved with a timestamp.

[0085] Dynamic comparison includes: using OpenCV difference algorithm (background subtraction or inter-frame difference method) to calculate the pixel difference between adjacent images in real time. If the difference value exceeds a threshold (such as 5%), it is marked as "detection abnormality" and an alarm is triggered.

[0086] S2. Continuously capture images and compare and analyze them with the previous image to detect and mark any signs of rupture in the water film. This embodiment continuously captures images through a sensor (such as a camera). Image data is transmitted to a computer system via a data acquisition card, a network, or other interfaces. These images are sequentially stored in a storage medium, and each time a new image is captured, the system needs to compare it with the previous image. The purpose of image comparison is to detect changes in the water film, especially signs of rupture. The specific process includes the following steps:

[0087] S21. Compare the current image collected with the previous image in sequence to determine abnormal images with differences or feature changes among the multiple images; in step S21, the abnormal images are determined by using any one of pixel-level difference comparison, image feature extraction comparison, or structural similarity index, wherein:

[0088] Pixel-level difference comparison: Compare the pixel values of the two images and calculate the difference value D(x,y) between the pixels. If the difference value D(x,y) exceeds the preset threshold, it indicates that the water film has changed or broken, that is:

[0089] D(x,y)=|I1(x,y)-I2(x,y)|

[0090] Where I1(x,y) and I2(x,y) are the pixel values corresponding to the current image and the previous image respectively, where (x,y) is the pixel coordinate;

[0091] Image feature extraction and comparison: By extracting key feature points from the image and comparing their changes between the current and previous images, we can determine whether the water film has broken or deformed. Key feature points include edges and corners. Common feature extraction methods include SIFT (Scale Invariant Feature Transform) and SURF (Speeded Robust Features).

[0092] Structural Similarity (SSIM) Index: This is an indicator used to measure the similarity between two images, which can more effectively evaluate the overall structural changes of the image. The calculation formula of the SSIM index is:

[0093]

[0094] Where μ x ,μ y are the means of the two images, σ x ,σ y are their variances, σ xy is their covariance, C1 and C2 are constants to avoid division by zero errors. The lower the SSIM(x,y) value, the greater the difference between the two images, which indicates that the water film has broken or deformed.

[0095] S22. Threshold processing is performed on the grayscale value or color distribution of the abnormal image to identify possible crack or rupture areas, wherein the threshold processing can be performed using the following formula:

[0096]

[0097] Where T is the set threshold; I(x,y) is the grayscale value of the pixel; I`(x,y) is the processed image.

[0098] In this embodiment, an edge detection algorithm (such as Canny edge detection) is used to detect crack edges on the water film. Canny edge detection detects possible cracks by looking for areas in the image with dramatic grayscale changes. A machine learning or deep learning model is then used to train the crack detector. For example, a convolutional neural network (CNN) is used to train a model to identify the presence of cracks in an image. These models can determine whether the water film is broken by inputting image features.

[0099] S23. Mark the abnormal image, including the bounding box of the crack or rupture area, the time, and the location. Once signs of a water film rupture are identified, they are marked in the image. This marking can be done by marking the bounding box of the rupture area on the image, color-coding the rupture area (e.g., a red frame), and recording information such as the time, location, and severity of the rupture for further analysis and tracking.

[0100] S3. Detect rupture images showing signs of water film rupture, and sequentially store and record the rupture image, the time of rupture, the location, and the rupture area. After each new image is acquired, this embodiment uses instance segmentation technology to determine the location and outline of the water film, and then compares the segmentation results with the previous image. If the segmented outline of the water film changes significantly, the system will determine that it has ruptured.

[0101] Specifically, by detecting missing, cracked, or broken water film contours in instance segmentation results and combining them with morphological algorithms (such as edge detection), the accuracy of rupture detection can be further enhanced. The YOLOv8 model is capable of detecting objects at different scales and can further determine whether a local rupture has occurred by comparing the changes in large and small water films in the image.

[0102] If a water film rupture is detected, the water film is recorded, the detected rupture image is saved, and the rupture detection result is displayed. At this time, the relevant images and segmentation masks of the water film rupture are saved. These images will include the changes before and after the rupture, as well as the specific marked water film rupture area.

[0103] Finally, a report is automatically generated, including the detected rupture image, the time of occurrence, location, and area of the rupture. The user interface displays the rupture detection results in real time, annotating the image and providing detailed information about the rupture. The instance segmentation results of the YOLOv8 model allow for convenient data backtracking, locating the locations of water film ruptures at different time points, and analyzing trends or causes of water film ruptures.

[0104] S4: Continue to capture images and compare them with the rupture image to determine whether the water film rupture phenomenon persists. If so, the authenticity of the rupture phenomenon is confirmed and the process proceeds to step S5. If not, the image is reanalyzed and the process returns to step S2 to determine whether it is a false positive. Monitoring continues for 2 seconds after rupture detection. During this period, if the water film rupture phenomenon persists, the authenticity of the rupture phenomenon is confirmed. If the rupture area disappears or cannot be detected again, the image is reanalyzed to determine whether it is a false positive.

[0105] To avoid misjudgments due to accidental factors, this embodiment employs a method of continuously acquiring and comparing images, continuing monitoring for two seconds after a crack is detected. By comparing and analyzing the latest image with previously acquired images, it ensures that any crack features in the image are not merely accidental but actual signs of a crack. This continuous acquisition and comparative analysis method improves detection accuracy.

[0106] Sometimes, due to changes in the external environment (such as light flickering or accidental obstruction by objects), the sensor may temporarily misidentify the object. To prevent these incidental factors from affecting the detection results, the system will continue to collect and compare images after the initial detection of a crack to ensure that the crack persists.

[0107] In order to confirm the authenticity of the rupture, the detection results of the previous and next images will be combined. If the water film rupture in the previous and next images is consistent and the change is greater than the preset error range, the authenticity of the water film rupture can be confirmed. If the rupture phenomenon is detected to be only temporary or a misjudgment caused by changes in the external environment, it will be marked as an invalid detection to avoid false positives. In several consecutive frames of images, if the ruptured area continues to appear and changes consistently, it means that the rupture is persistent. Otherwise, it may be a misjudgment. By combining YOLOv8's instance segmentation and difference detection algorithms, it is possible to more accurately determine whether the water film rupture is real. Specifically, the authenticity of the water film rupture is comprehensively judged by detecting the shape changes of the water film at multiple time points, the expansion or disappearance of the cracks.

[0108] Different from the traditional multi-frame difference method and optical flow analysis method, the water film boundary perturbation entropy and time series stability factor are introduced as new judgment bases to continuously detect the authenticity of water film rupture. Based on the perturbation information of the water film boundary, it is possible to detect subtle fluctuations of the water film when it ruptures, rather than relying solely on area changes. Combined with time series stability analysis, short-term environmental interference (such as light flicker and wind disturbance) can be filtered out to reduce misjudgment. It is suitable for dynamic environments, and statistical modeling is performed on the morphological changes of water film rupture to predict the rupture trend. The specific process includes:

[0109] S41. Use YOLOv8 network model instance segmentation to obtain water film image I t The water film boundary contour in

[0110]

[0111] S42. Calculate the given water film boundary contour C t The local curvature of the i-th point in is calculated as:

[0112]

[0113] Among them, It is the water film image (input image) of the tth frame; (x′ i ,y′ i ) is the tangent direction of the boundary point; (x″ i ,y″ i ) is the second-order derivative of the boundary point, which represents the curvature of the boundary; κ i Reflects the local curvature change of the water film boundary; C t is the boundary contour of the water film in the tth frame;

[0114] S43, all local curvatures κ i Normalized to get the curvature distribution P(κ i ), the expression is:

[0115]

[0116] Where N is the total number of pixels on the water film boundary; P(κ) represents the distribution of different curvature values;

[0117] S44, using the information entropy formula according to the curvature distribution P(κ i ) Calculate the water film boundary perturbation entropy H that reflects the complexity of the water film boundary morphology t , the calculation formula is:

[0118]

[0119] If H t The continuous increase indicates that the boundary of the water film is changing dynamically, indicating that the rupture phenomenon persists.

[0120] If H t Small or decreasing fluctuations indicate that the water film has stabilized and may be a misjudgment;

[0121] This embodiment calculates the shape change of the water film boundary to extract the disturbance feature, and uses the water film boundary disturbance entropy to measure the stability of the water film boundary as a key indicator for rupture detection.

[0122] S45, using the dynamic time warping algorithm to calculate the time used to represent the previous frame C t With the previous frame C t-1 The dynamic time warping distance D between similarities t , the calculation formula is:

[0123] D t =DTW(C t ,C t-1 )

[0124] If D t is larger, indicating that the water film morphology changes dramatically;

[0125] If Dt is smaller, indicating that the water film boundary is more stable;

[0126] Among them, C t is the water film boundary contour of the tth frame; D t is the dynamic time warping distance between the t-th frame and the previous frame; DTW is used to calculate the similarity of two time series;

[0127] In order to distinguish between transient ruptures and stable ruptures, this embodiment filters out transient interference by calculating the variability of the water film morphology in the time dimension.

[0128] S46, based on dynamic time warping distance D t Define the time series stability factor S t , the expression is:

[0129]

[0130] If S t >T S , the rupture persists;

[0131] If S t <T S , then the rupture area is unstable and may be a misjudgment;

[0132] Among them, T S To set the threshold, the value needs to be determined through experiments based on the actual target, and this embodiment does not impose any limitation; N is the number of dynamic time warping distances;

[0133] S47, according to the water film boundary perturbation entropy H t and the time series stability factor S t Continuously detect the authenticity of water film rupture;

[0134] If H t >T H And S t >T S , it means that the water film rupture phenomenon persists at multiple time points, confirming the rupture;

[0135] If H t <T H or S t <T S , it means that the fluctuation in the water film rupture area is small, which may be environmental noise or false detection, and needs to be re-analyzed.

[0136] In this embodiment, H t is the perturbation entropy of the water film boundary at frame t, which measures the complexity of the water film boundary morphology. The larger the value, the more severe the boundary fluctuation, which may be a real rupture. HIs the threshold for judging water film rupture, if H t >T H , it indicates that the change of the water film boundary is large enough and may be a real rupture. t The stability of the water film morphology is calculated in the time dimension, indicating the degree of change in the water film rupture area at multiple time points. The larger the value, the more unstable the water film morphology is, and it may be a real rupture. S is the set time stability threshold, if S t >T S , it means that the change of water film morphology is stable in time and the rupture phenomenon persists.

[0137] S5. After the test is completed, the test trolley is moved to the box cabin and vacuum extraction is performed to prevent further corrosion of the equipment, such as Figure 2 and Figure 3 When all image acquisition and analysis are complete and the final test results are obtained, the user will be notified via a display screen or voice prompt that the test is complete and ready for subsequent operations. All acquired images, test results, and related data are saved, and reports can be generated. Users can view detailed information such as each image during the inspection, the fracture location, and the timestamp, ensuring data integrity and traceability.

[0138] Since acidic gases may be generated in the immersion tank (such as gases generated during the pickling process), these gases may corrode the electronic equipment and sensors of the cart. Therefore, after the inspection is completed, the inspection cart needs to be safely pushed back to the cabin. The cabin is equipped with an air purification function, which usually includes equipment such as air filtration and negative ion generators, which can effectively remove acidic gases and harmful substances in the air to prevent corrosion from causing long-term damage to the equipment. After pushing the cart back into the cabin, it can not only prevent corrosion, but also provide a suitable storage environment for the equipment, extending the service life of the equipment. In the cabin, the equipment is in a safe, temperature and humidity controlled environment, which can prevent the external environment from affecting it to the greatest extent. And after the inspection is completed, the system will automatically perform a self-check to ensure that all sensors and hardware equipment are fault-free and are ready for the next inspection.

[0139] In the overall concept of this embodiment, a water film online detection system is also proposed. First, the operator needs to manually position the mobile detection cart to the area to be inspected. After connecting the system power supply, the sensor self-test program needs to be executed to verify the core parameters such as the sensitivity of the ultrasonic probe and the resolution of the image acquisition module to ensure that the equipment meets the requirements. Secondly, when the multispectral sensor detects that the target object enters the detection area, the system automatically activates the high-speed imaging unit. The unit uses an industrial-grade camera with a frame rate of 30 frames per second and realizes dynamic monitoring through a real-time image difference algorithm. Subsequently, the system continuously collects 30 seconds of image data (about 30 sets of samples) and uses a convolutional neural network (CNN) to extract features of the water film morphology. When a sudden change of the water film contact angle of >15° or a continuous fracture is detected, the three-level alarm mechanism is immediately triggered. It is worth noting that the system will continue to monitor for 2 seconds after the initial judgment of the rupture (i.e., obtain 60 frames of supplementary data). The expansion speed and morphological evolution of the rupture area are secondary verified by time series analysis. After the detection is completed, the operator needs to immediately move the detection cart back to the closed cabin. The cabin is equipped with a positive pressure nitrogen protection system, which can control the concentration of acidic gas to <5ppm, effectively preventing corrosion of sensors and other equipment.

[0140] This embodiment builds an intelligent closed-loop detection system based on industrial testing standards and water film integrity monitoring technical requirements. First, mechanical guide rails achieve millimeter-level positioning (positioning accuracy ±0.5mm), combined with a sensor self-test program to ensure the ultrasonic probe sensitivity reaches ±0.1dB. Second, multispectral sensing triggers a high-speed imaging unit, using a 30fps industrial camera and a differential algorithm to achieve dynamic monitoring. The core detection process uses CNN three-dimensional reconstruction technology to analyze 30-second image sequences. A sudden change in contact angle greater than 15° triggers a level 3 alarm, which meets the requirements of ASTM F22-2013.

[0141] An innovative 2-second continuous monitoring mechanism eliminates 90% of false positives through time series analysis; ultimately, a nitrogen protection system maintains the acid gas level in the chamber below 5 ppm. An innovative multi-period cross-validation mechanism significantly improves the accuracy of rupture determination. The integration of machine vision and deep learning algorithms overcomes the limitations of traditional contact-based inspection, establishing a full-lifecycle equipment protection system and extending the service life of key components. Significant economic benefits are reflected in three key areas: automated processes multiply inspection efficiency, extend equipment life by over 40%, and increase inspection accuracy to 98.7%, reducing rework losses.

[0142] Those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiment methods can be accomplished by instructing the relevant hardware through a program. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0143] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to in detail. For the above embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For relevant parts, please refer to the partial description of the method embodiments.

[0144] The above embodiments provide a detailed introduction to the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A water film online detection method, characterized in that: The method comprises the following steps: S1. Move the inspection vehicle to the inspection position of the workpiece to be tested. After the sensor detects the workpiece to be tested, start the image acquisition program to collect images at a frequency of 1 Hz and perform dynamic comparison with the previous image. S2, continuously collecting images and comparing and analyzing them with the previous image to detect whether the water film shows signs of rupture and mark them; S3, detecting the rupture image of the water film showing signs of rupture, and sequentially storing and recording the rupture image, the time, location and area of the rupture; S4, continue to collect images and compare them with the rupture image to determine whether the water film rupture phenomenon persists; if so, confirm the authenticity of the rupture phenomenon and proceed to step S5; if not, re-analyze the image and return to step S2 to determine again whether it is a false positive; S5. After the inspection is completed, the inspection trolley is moved into the box cabin and vacuum extraction is performed to prevent further corrosion of the equipment.

2. The water film online detection method according to claim 1, characterized in that: The dynamic comparison includes: using the OpenCV difference algorithm to calculate the pixel differences between adjacent images in real time. If the difference value exceeds a threshold, it is marked as "detection anomaly" and an alarm is triggered.

3. The water film online detection method according to claim 1, characterized in that: In step S2, the specific process includes the following steps: S21, sequentially comparing the current image acquired with the previous image, and determining abnormal images having differences or feature changes among the images; S22, performing threshold processing on the grayscale value or color distribution of the abnormal image to identify possible crack or rupture areas; S23. Marking the abnormal image, including the boundary box, time, and location of the crack or rupture area.

4. The water film online detection method according to claim 3, characterized in that: In step S21, abnormal images are determined by using any one of pixel-level difference comparison, image feature extraction comparison, or structural similarity index, wherein: Pixel-level difference comparison: Compare the pixel values of the two images and calculate the difference value D(x,y) between the pixels. If the difference value D(x,y) exceeds the preset threshold, it indicates that the water film has changed or broken, that is: D(x,y)=|I1(x,y)-I2(x,y)| Where I1(x,y) and I2(x,y) are the pixel values corresponding to the current image and the previous image respectively, where (x,y) is the pixel coordinate; Image feature extraction and comparison: By extracting key feature points from the image and comparing the changes in key feature points between the current image and the previous image, it is possible to determine whether the water film has broken or deformed. Structural Similarity Index: The lower the value of the SSIM, the greater the difference between the two images, which means that the water film has broken or deformed. The calculation formula of the SSIM(x,y) value is: Where μ x ,μ y are the means of the two images respectively; σ x ,σ y are their variances; σ xy is their covariance; C1 and C2 are constants to avoid division by zero errors.

5. The water film online detection method according to claim 4, characterized in that: The key feature points include edge points and corner points in the image.

6. The water film online detection method according to claim 1, characterized in that: In step S4, it includes: Continue to collect and compare images, and continue monitoring within 2 seconds after the rupture is detected. If the rupture area continues to appear and changes consistently in several consecutive frames of images, it means that the rupture is persistent. Otherwise, it may be a misjudgment.

7. The water film online detection method according to claim 1 or 6, characterized in that: In step S4, the specific process includes: S41. Use YOLOv8 network model instance segmentation to obtain water film image I t The water film boundary contour in S42. Calculate the given water film boundary contour C t The local curvature of the i-th point in is calculated as: Among them, I t is the water film image (input image) of the tth frame; (x′ i ,y′ i ) is the tangent direction of the boundary point; (x″ i ,y″ i ) is the second-order derivative of the boundary point, which represents the curvature of the boundary; κ i Reflects the local curvature change of the water film boundary; C t is the boundary contour of the water film in the tth frame; S43, all local curvatures κ i Normalized to get the curvature distribution P(κ i ), the expression is: Where N is the total number of pixels on the water film boundary; P(κ) represents the distribution of different curvature values; S44, using the information entropy formula according to the curvature distribution P(κ i ) Calculate the water film boundary perturbation entropy H that reflects the complexity of the water film boundary morphology t , the calculation formula is: If H t The continuous increase indicates that the water film boundary is changing dynamically, indicating that the rupture phenomenon persists; If H t Small or decreasing fluctuations indicate that the water film has stabilized and may be a misjudgment; S45, using the dynamic time warping algorithm to calculate the time used to represent the previous frame C t With the previous frame C t-1 The dynamic time warping distance D between similarities t , the calculation formula is: D t =DTW(C t ,C t-1 ) If D t is larger, indicating that the water film morphology changes dramatically; If D t is smaller, indicating that the water film boundary is more stable; Among them, C t is the water film boundary contour of the tth frame; D t is the dynamic time warping distance between the t-th frame and the previous frame; DTW is used to calculate the similarity of two time series; S46, based on dynamic time warping distance D t Define the time series stability factor S t , the expression is: If S t >T S , the rupture persists; If S t <T S , then the rupture area is unstable and may be a misjudgment; Among them, T S To set the threshold, the value needs to be determined through experiments based on the actual target, and this embodiment does not impose any limitation; N is the number of dynamic time warping distances; S47, according to the water film boundary perturbation entropy H t and the time series stability factor S t Continuously detect the authenticity of water film rupture; If H t >T H And S t >T S , it means that the water film rupture phenomenon persists at multiple time points and is confirmed to be rupture; If H t <T H or S t <T S , it means that the fluctuation in the water film rupture area is small, which may be environmental noise or false detection, and needs to be re-analyzed.

8. A system for implementing the water film online detection method according to any one of claims 1 to 7, characterized in that: include: The dynamic comparison module moves the inspection vehicle to the inspection position of the workpiece to be tested. After the sensor detects the workpiece to be tested, it starts the image acquisition program to collect images at a frequency of 1Hz and dynamically compares them with the previous image. The rupture sign detection module continuously collects images and compares and analyzes them with the previous image to detect and mark whether the water film shows signs of rupture; The rupture detection module detects the rupture image of the water film showing signs of rupture, and sequentially stores and records the rupture image, the time, location and area of the rupture; The continuous detection module continues to collect images and compares them with the rupture images to determine whether the water film rupture phenomenon persists. If so, the authenticity of the rupture phenomenon is confirmed; if not, the image is re-analyzed to determine whether it is a false positive. Reset the module. After the inspection is completed, move the inspection cart into the box cabin and perform vacuum extraction to prevent further corrosion of the equipment.

9. A device for implementing the water film online detection method according to any one of claims 1 to 7, characterized in that: include: processor; Memory; robotic arm; Box cockpit; Detection control cabinet; and one or more programs, wherein the one or more programs are stored in a memory and configured to be executed by the processor, and the programs are used for a computer to execute the water film online detection method according to any one of claims 1 to 6.

Images