Bubble identification method, application and system based on machine vision

Abstract

The invention relates to a bubble recognition method based on machine vision, and the method comprises the steps: obtaining an image sequence of a test liquid environment, enabling the upstream side of a target object to be provided with a gas cavity, enabling the image sequence to be obtained when gas pressure is applied to the upstream side of the target object, and enabling a focal plane corresponding to the imaging of the image sequence to form an included angle larger than 0 degree with a vertical plane, and performing bubble recognition on the image sequence based on an image recognition algorithm, judging whether the image sequence contains bubbles according to a recognition result so as to obtain a first bubble-containing image, and determining the gas pressure applied to the upstream side of the target object at the collection moment according to the collection moment of the first bubble-containing image. According to the invention, the problem of low first bubble point detection accuracy is solved. The first bubble point is detected by obtaining the image sequence, large manual detection errors are avoided, and the detection accuracy is improved. Meanwhile, an included angle is formed between the focal plane and the vertical plane, so that a single image acquisition device can obtain clear images of the bubbles.

Description

Technical Field

[0001] The present application relates to the field of machine vision, and in particular to a bubble recognition method, application and system based on machine vision. Background Art

[0002] Before using the filter membrane and filter element, the product body or sample needs to be tested for integrity to ensure the safety and effectiveness of the filtration process; before mass production and formal use of joints, closed containers and pipelines, the product also needs to be tested for fluid sealing performance to avoid leakage during use.

[0003] In current technologies, for example, integrity testing of filter membranes and filter elements is mostly completed with integrity testers. During the detection process, the gas flow through the filter is monitored by a flow meter. Since the volume of the first bubble is very small (sometimes below 0.01mm), and the gas is easily compressed, the flow meter may not be able to detect the first bubble. Also, since the gas is greatly affected by temperature, it is easy to cause misdetection by the flow meter and result in an error in recording the first bubble point.

[0004] There is also a method of placing the product in a liquid, applying gas pressure to the upstream side of the product to be measured to drive bubbles to appear, and detecting by directly observing the bubbles. However, this method may result in differences in the recognition and judgment of bubbles by different testers, resulting in affected consistency of test results. In addition, the test results are also affected by the fatigue and attention of the tester. Long-term observation and identification may cause tester fatigue and affect concentration, thereby increasing the risk of misjudgment.

[0005] In view of this, this application is filed. Summary of the invention

[0006] The embodiments of the present application provide a machine vision-based bubble recognition method, application, and system to at least solve the problem of low first bubble point detection accuracy in the related art.

[0007] In a first aspect, an embodiment of the present application provides a bubble recognition method based on machine vision, the method comprising:

[0008] Acquire an image sequence of a test liquid environment, wherein the upstream side of the target object has a gas cavity, the image sequence is acquired when gas pressure is applied to the upstream side of the target object, and the focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane;

[0009] Bubble recognition is performed on the image sequence based on an image recognition algorithm, and whether bubbles are included in the image sequence is determined according to the recognition result to obtain a first bubble-containing image.

[0010] By acquiring an image sequence to detect bubbles, the large error of manual detection is avoided and the accuracy of first bubble point detection is improved. At the same time, the focal plane corresponding to the image sequence imaging forms an angle greater than 0° with the vertical plane, so that the generated bubbles will always pass through the focal plane of the image acquisition device at a certain moment, so that a clear image of the bubble can be obtained through a single image acquisition device.

[0011] In some of the embodiments, before acquiring the image sequence of the test liquid environment, the method further includes: expelling bubbles attached to the surface of the target object and / or dissolved in the test liquid through ultrasonic vibration.

[0012] Ultrasonic waves can vibrate small bubbles attached to the surface of the target object and expel them from the test liquid. This can also help to convert the gas dissolved in the liquid from a dissolved state to a free state, and then expel it from the test liquid, avoiding misjudgment of the first bubble point caused by bubbles attached to the filter surface.

[0013] In some of the embodiments, when acquiring the image sequence of the test liquid environment, the method further includes: driving bubbles existing in the blind spots of the target object to float up through ultrasonic vibration.

[0014] In order to prevent the bubbles generated by pressurization from existing in the blind spots of the target object (corners where bubbles are difficult to expel under the buoyancy of the test liquid), resulting in misjudgment of the first bubble point, ultrasound will continue to be turned on when acquiring the image sequence to drive the bubbles out of the blind spots and float up.

[0015] In some of the embodiments, when acquiring the image sequence of the test liquid environment, the method further includes: driving the target object to rotate in the test liquid so as to make bubbles existing in the blind spots of the target object float up.

[0016] This embodiment can also drive the target object to rotate in the test liquid through the rotating device to make the bubbles escape from the blind spot and float up, thereby avoiding the generated bubbles existing in the blind spot of the target object and being unable to float up, resulting in misjudgment of the first bubble point.

[0017] In some of the embodiments, the method further includes: acquiring an optical flow field of the bubble based on the image sequence; filtering a horizontal component of the optical flow field, and extracting a vertical component of the optical flow field.

[0018] The calculated optical flow field is decomposed into horizontal and vertical directions, and the horizontal component is filtered out to reduce the influence of liquid ripples and Brownian motion of foreign objects on bubble detection, avoid ripples and particles in the water being misjudged as bubbles, and improve the accuracy of bubble detection.

[0019] In some of the embodiments, the method further includes: obtaining the movement rate of the pixel point corresponding to the bubble according to the optical flow field; setting an active bubble movement rate threshold interval based on the gas pressure upstream of the target object, and determining the display state of the pixel point according to the movement rate threshold and the movement rate.

[0020] The movement rate of the pixel point corresponding to the bubble is determined by the vertical component of the optical flow field, and the aperture size of the bubble is known based on the movement rate. The aperture size is intuitively displayed through different display states to facilitate operator understanding and analysis.

[0021] In some embodiments, the method further comprises:

[0022] Based on a preset division rule, performing region division in the vertical direction of the bubble image to obtain at least two image regions;

[0023] Based on the bubble movement rate of each of the image regions, movement rate thresholds corresponding to different image regions are set.

[0024] The image is divided into multiple areas and different motion rate thresholds are set for correction, which avoids the problem of reduced accuracy caused by different bubble rising heights.

[0025] In some embodiments, the display state includes: pixel value, gray value or pixel shape.

[0026] The aperture size is intuitively displayed through significant distinctions such as different pixel values, grayscale values ​​or pixel shapes.

[0027] In some embodiments, the method further comprises:

[0028] In some embodiments, the method further comprises:

[0029] If bubbles are detected without reaching the preset initial pressure threshold, warning information indicating poor airtightness of the target object is generated based on the pressure value when the bubbles are detected and a preset warning classification mechanism.

[0030] An alarm mechanism is set up to provide real-time warning of filter element leakage or poor valve air tightness, preventing potential safety accidents and improving the safety and reliability of the equipment.

[0031] In some of the embodiments, the image acquisition device for acquiring the image sequence is selected from: any one of a tilt-shift camera and an industrial camera.

[0032] The image acquisition device for acquiring the image sequence can be a tilt-shift camera or a tilted industrial camera. The tilt-shift camera adjusts the focal plane by tilting the lens, so that the photographer can selectively control the depth of field and the tilt angle of the focal plane. Both the tilted industrial camera and the tilt-shift camera can adjust the focal plane to a suitable position so that the generated bubbles always pass through the focal plane at a certain moment, so that a clear image of the bubbles can be obtained through a single image acquisition device.

[0033] In a second aspect, the present application also provides an application of a machine vision bubble recognition method in integrity testing, wherein the target object is a filter element or a filter membrane; a gradually increasing gas pressure is applied to the upstream side of the target object, and the gas pressure applied to the upstream side of the target object at the acquisition time is determined based on the acquisition time corresponding to the first bubble-containing image.

[0034] This method can repeatedly perform the inspection task under the same conditions, without being affected by operator fatigue or subjective judgment, thus ensuring the consistency and reliability of the inspection results. In addition, by collecting images in real time through the lens, the first bubble that appears can be accurately captured, thus achieving high-precision inspection.

[0035] In some of the embodiments, the method further comprises: acquiring target images with multiple time stamps within a preset time period from the image sequence, and determining the volume of bubbles in each of the target images;

[0036] If the bubble volume difference is within a preset difference range, it is determined that the target object is in a stable bubble state;

[0037] The pressure data corresponding to the stable bubbling state is acquired, and the average pore diameter of the target object is determined according to the pressure data.

[0038] When the gas pressure is the same, the larger the bubble rate, the larger the hole. According to the timestamp of the target image in the stable bubble state, the gas pressure applied on the upstream side of the target object in the stable bubble state is determined, and the pore size is determined by the gas pressure and the bubble movement rate. The pore size is calculated in the case of stable bubble, which improves the accuracy.

[0039] On the third aspect, the present application also provides an application of a machine vision bubble recognition method in a tightness test, wherein the target object is a closed pipe, a closed joint, a closed valve or a closed container; a constant preset gas pressure is applied to the upstream side of the target object, and based on the acquisition time corresponding to the first bubble-containing image, the time when bubbles appear in the target product under the preset gas pressure is determined.

[0040] By testing air tightness through this method, tiny changes in bubbles can be captured, thereby detecting extremely small leaks, improving the accuracy and sensitivity of detection; in addition, this method is adaptable to a variety of different detection environments and objects, including products of different shapes, sizes and materials; further, by adjusting the range of the focal plane, this method may also accurately identify the leakage point.

[0041] In a fourth aspect, an embodiment of the present application provides a bubble detection system based on machine vision, the system comprising: a test container, a pressurizing device, an image acquisition device and an information processing device, wherein:

[0042] The test container contains a test liquid;

[0043] The pressurizing device is connected to the upstream side of the target object to apply gas pressure to the upstream side of the target object immersed in the test liquid;

[0044] The image acquisition device is used to acquire an image sequence of the target object and the test fluid environment, wherein a focal plane of the image acquisition device forms an angle greater than 0° with a vertical plane;

[0045] The information processing device is used to perform bubble recognition on the image sequence based on an image recognition algorithm, and determine whether the image sequence contains bubbles according to the recognition result to obtain the first image containing bubbles; and determine the pressure applied by the pressurizing device at the corresponding moment according to the timestamp of the first image containing bubbles.

[0046] By acquiring the image sequence to detect the first bubble point, the large error of manual detection is avoided and the accuracy of the first bubble point detection is improved. At the same time, the focal plane corresponding to the image sequence imaging forms an angle greater than 0° with the vertical plane, so that the generated bubbles will always pass through the focal plane of the image acquisition device at a certain moment, so that a clear image of the bubble can be obtained through a single image acquisition device.

[0047] In some of the embodiments, a light strip is provided on the top of the test container;

[0048] The test container has a transparent first side wall, and the image acquisition device is arranged on the side of the first side wall; the test container has a second side wall arranged opposite to the first side wall, and the second side wall has a light shielding layer.

[0049] A light strip is arranged on the top of the test container and a light shielding layer is arranged on the second side wall to provide sufficient light for image acquisition, making the image clearer and improving the accuracy of detection.

[0050] Compared with the related art, the bubble recognition method based on machine vision provided in the embodiment of the present application obtains an image sequence of the test liquid environment. The upstream side of the target object has a gas cavity. The image sequence is obtained when gas pressure is applied to the upstream side of the target object. The focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane. The image sequence is subjected to bubble recognition based on the image recognition algorithm. It is determined whether the image sequence contains bubbles according to the recognition result to obtain the first bubble-containing image. According to the acquisition time of the first bubble-containing image, the gas pressure applied to the upstream side of the target object at the acquisition time is determined, which solves the problem of low accuracy of first bubble point detection. By acquiring the image sequence to detect the first bubble point, large manual detection errors are avoided and the accuracy of first bubble point detection is improved. The focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane, so that the generated bubbles will always pass through the focal plane of the image acquisition device at a certain moment, so that clear imaging of the bubbles can be obtained through a single image acquisition device. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] 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:

[0052] Figure 1 is a flow chart of a bubble recognition method based on machine vision according to an embodiment of the present application;

[0053] Figure 2 is a schematic diagram of an image sequence acquisition environment according to an embodiment of the present application;

[0054] Figure 3 is a schematic diagram of a bubble display state according to an embodiment of the present application;

[0055] Figure 4 It is a structural block diagram of a system according to an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0057] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0058] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0059] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantitative limitation, and may represent the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0060] Embodiment 1:

[0061] Figure 1 is a flow chart of a bubble recognition method based on machine vision according to an embodiment of the present application, such as Figure 1As shown, the process includes the following steps:

[0062] Step S101, obtaining an image sequence of a test liquid environment, wherein the upstream side of the target object has a gas cavity, the image sequence is obtained when gas pressure is applied to the upstream side of the target object, and the focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane.

[0063] The method of this embodiment is applied to integrity testing, and the target object is a filter element or a filter membrane. Based on the capillary principle, the pores in the filter element and the filter membrane are filled with a wetting liquid, and the minimum pressure required to squeeze the liquid out of the membrane pores, i.e., the first bubble point pressure, is measured by gradually increasing the gas pressure.

[0064] An image sequence refers to a series of image frames arranged in time order, which together constitute a video stream. The focal plane corresponding to the image sequence imaging forms an angle greater than 0° with the vertical plane, that is, the focal plane of the image acquisition device is tilted relative to the vertical plane, which makes the generated bubbles always pass through the focal plane at a certain moment, thereby obtaining a clear image of the bubbles.

[0065] It should be noted that the image acquisition device for acquiring the image sequence can be a tilt-shift camera or an industrial camera placed at an angle. The tilt-shift camera adjusts the angle and position of the optical axis of the lens and the imaging sensor by tilting and translating the lens, so that the photographer can selectively control the tilt angle of the focal plane and the shooting range. Figure 2 is a schematic diagram of an image sequence acquisition environment according to an embodiment of the present application, such as Figure 2 As shown, the image acquisition device is a tilt-shift camera, whose focal plane forms an angle greater than 0° with the vertical plane. The focal plane is adjusted forward and backward to a suitable position before image acquisition is performed, so that the generated bubbles will always pass through the focal plane at a certain moment, so that a clear image of the bubbles can be obtained through a single image acquisition device. It should be noted that the specific angle of the focal plane and the vertical plane can be adjusted according to the area where the bubbles may appear. For example, if the area where the bubbles may appear in the test container is large, the angle of the focal plane can be appropriately increased; when the adjusted angle is the largest, the focal plane is parallel to the horizontal plane.

[0066] In some of the embodiments, before acquiring the image sequence of the test liquid environment, the method further includes: expelling bubbles attached to the surface of the target object and / or dissolved in the test liquid through ultrasonic vibration.

[0067] The target object is placed in a container with an ultrasonic generator. Ultrasonic waves can vibrate small bubbles attached to the surface of the target object and expel them from the test liquid. This can also help to convert the gas dissolved in the liquid from a dissolved state to a free state, and then expel it from the test liquid, thus avoiding misjudgment of the first bubble point caused by bubbles attached to the filter surface.

[0068] In some of the embodiments, when acquiring the image sequence of the test liquid environment, the method further includes: driving bubbles existing in the blind spots of the target object to float up through ultrasonic vibration.

[0069] In order to avoid the bubbles generated by pressurization existing in the dead corners of the target object (corners where bubbles are difficult to be discharged under the buoyancy of the test liquid), resulting in misjudgment of the first bubble point, the ultrasonic wave will continue to be turned on when acquiring the image sequence to drive the bubbles out of the dead corners and float up. It should be noted that since the ultrasonic generator will produce cavitation effect in the liquid, technicians in this field should adjust the frequency and intensity of the ultrasonic generator to a reasonable range when using the ultrasonic generator, so as to avoid bubbles due to the cavitation effect.

[0070] In some of the embodiments, when acquiring the image sequence of the test liquid environment, the method further includes: driving the target object to rotate in the test liquid so as to make bubbles existing in the blind spots of the target object float up.

[0071] This embodiment can also drive the target object to rotate in the test liquid through the rotating device to make the bubbles escape from the blind spot and float up, thereby avoiding the generated bubbles existing in the blind spot of the target object and being unable to float up, resulting in misjudgment of the first bubble point.

[0072] Continue to refer to Figure 1 , after acquiring the image sequence, execute step S102.

[0073] Step S102 , performing bubble recognition on the image sequence based on an image recognition algorithm, and judging whether the image sequence contains bubbles according to the recognition result, so as to obtain the first bubble-containing image.

[0074] The first image containing bubbles in the image sequence is obtained by using an image recognition algorithm, including:

[0075] Image preprocessing: Preprocess the image frames in the image sequence, convert the color image into a grayscale image to simplify the processing process, and use a filtering algorithm (e.g., Gaussian filtering, median filtering) to remove image noise.

[0076] Feature extraction: Based on the contrast difference between bubbles and background, a threshold is set to separate bubbles from the background. Features are extracted from possible bubble regions, including but not limited to shape, size, texture, and grayscale value.

[0077] Bubble recognition: Bubbles can be recognized based on features extracted from image frames (e.g., circles, specific size ranges) based on preset recognition rules. Bubbles can also be recognized based on machine learning, using trained classifiers (e.g., support vector machines, deep learning models, etc.).

[0078] Step S103, determining the gas pressure applied to the upstream side of the target object at the acquisition time according to the acquisition time corresponding to the first bubble-containing image.

[0079] The acquisition time of the first bubble-containing image is the first bubble point time of the target object.

[0080] Through the above steps, an image sequence of the test liquid environment is obtained. The upstream side of the target object has a gas cavity. The image sequence is obtained when gas pressure is applied to the upstream side of the target object. The focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane. Bubble recognition is performed on the image sequence based on the image recognition algorithm. Whether the image sequence contains bubbles is determined based on the recognition result to obtain the first bubble-containing image. According to the acquisition time of the first bubble-containing image, the gas pressure applied to the upstream side of the target object at the acquisition time is determined. This method can repeatedly perform the detection task under the same conditions and is not affected by operator fatigue or subjective judgment, thereby ensuring the consistency and reliability of the detection results. On the other hand, by acquiring images in real time through the lens, the first bubble that appears can be accurately captured, thereby achieving high-precision detection, thereby solving the problem of low accuracy of first bubble point detection.

[0081] The focal plane corresponding to the image sequence imaging forms an angle greater than 0° with the vertical plane, so that the generated bubbles will always pass through the focal plane of the image acquisition device at a certain moment, so that a clear image of the bubbles can be obtained through a single image acquisition device.

[0082] In some of the embodiments, the method further includes: acquiring an optical flow field of the bubble based on the image sequence; filtering a horizontal component of the optical flow field and extracting a vertical component of the optical flow field.

[0083] The optical flow field contains the movement speed and direction information of each corresponding pixel point in the image sequence. When the image sequence contains bubbles, the optical flow field of the bubbles is obtained based on the pixel changes between adjacent frames or multiple continuous frames in the image sequence.

[0084] Optical flow is a computer vision technique used to estimate the motion field in an image. It can provide speed and direction information for each pixel. Optionally, the OpenCV image processing algorithm is used to calculate the optical flow field for the motion between adjacent frames of the video stream. The calculated optical flow field is decomposed into horizontal and vertical directions, and the horizontal component is filtered out to reduce the impact of liquid ripples and foreign body Brownian motion on bubble detection, avoid ripples and particles in the water being misjudged as bubbles, and improve the accuracy of bubble detection.

[0085] In some embodiments, the bubble recognition method based on machine vision further includes:

[0086] Step S201, obtaining the movement rate of the pixel point corresponding to the bubble according to the optical flow field.

[0087] Step S202, setting an active bubble movement rate threshold interval based on the gas pressure upstream of the target object, and determining a display state of the pixel point according to the movement rate threshold and the movement rate.

[0088] When the air pressure is the same, if the bubble rate is greater, the corresponding hole on the filter element that produces the bubble will be larger. The movement rate of the pixel corresponding to the bubble is determined by the vertical component of the optical flow field, and the size of the aperture that produces the bubble is known based on the movement rate. The aperture size is intuitively displayed through different display states, which is convenient for operators to understand and analyze.

[0089] Optionally, bubbles of different motion speeds are labeled by different pixel values ​​(RGB values ​​of pixels), for example, the greater the motion speed, the greater the pixel value of the corresponding pixel. In this embodiment, bubbles of different motion speeds can also be labeled by different gray values ​​or pixel shapes.

[0090] In some embodiments, setting the active bubble movement rate threshold interval based on the gas pressure upstream of the target object in step S202 includes:

[0091] Step S2021: Based on a preset division rule, the bubble image is divided into regions in the vertical direction to obtain at least two image regions.

[0092] Step S2022: Based on the bubble movement rate of each image area, a movement rate threshold corresponding to each image area is set.

[0093] It should be noted that the vertical direction of the bubble image is the vertical direction in the corresponding real scene. Since there is a certain angle between the focal plane and the vertical direction, the rising height of the bubble is different in different areas in the vertical direction of the same bubble image, resulting in different movement rates for bubbles of the same size.

[0094] Figure 3 is a schematic diagram of a bubble display state according to an embodiment of the present application, such as Figure 3 As shown, the bubble image is divided into two different areas, a relatively proximal area and a relatively distal area. The motion rate threshold of the relatively proximal area is smaller than the motion rate threshold corresponding to the same bubble volume in the relatively distal area, and bubbles with the same labeled shape have the same volume.

[0095] By dividing the image into multiple areas and setting different motion rate thresholds for correction, the problem of reduced accuracy caused by different bubble rising heights is avoided. The motion rate threshold can be adjusted in real time according to actual conditions.

[0096] In some embodiments, the target object includes a filter element or a filter membrane; the bubble recognition method based on machine vision also includes:

[0097] Step S301 : acquiring target images with multiple time stamps within a preset time period from an image sequence, and determining the volume of bubbles in each target image.

[0098] Step S302: If the bubble volume difference is within a preset difference range, it is determined that the target object is in a stable bubble state.

[0099] Step S303, obtaining pressure data corresponding to the stable bubbling state, and determining the average pore diameter of the target object according to the pressure data.

[0100] If the volume difference of bubbles generated by the same bubbling point (bubbles at the same position in each target image) is within the preset difference range within the preset time period, it is considered to be a stable bubbling state. When the air pressure is the same, if the bubble rate is larger, the hole is larger. According to the timestamp of the target image in the stable bubbling state, the gas pressure applied to the upstream side of the target object in the stable bubbling state is determined, and the pore size is determined by the gas pressure and the bubble movement rate. The pore size is calculated in the case of stable bubbling, which improves the accuracy.

[0101] In some of the embodiments, the machine vision-based bubble recognition method further includes: when bubbles are detected without reaching a preset initial pressure threshold, a warning message indicating poor airtightness of the target object is generated based on the pressure value when the bubbles are detected and a preset warning classification mechanism.

[0102] An alarm mechanism is set up to provide real-time warning of filter element leakage or poor valve air tightness, preventing potential safety accidents and improving the safety and reliability of the equipment.

[0103] For example, for a sterilizing filter element with an average pore size of 0.22μm, its bubble point pressure is usually between 0.35 and 0.40MPa; the initial pressure threshold can be preset to 0.32MP, and if the corresponding pressure of the first bubble point detected is between 0.32 and 0.35MP, a first-level warning is issued, if the corresponding pressure of the first bubble point detected is between 0.3 and 0.32MP, a second-level warning is issued, and if the corresponding pressure of the first bubble point detected is below 0.3MP, a third-level warning is issued. Optionally, different alarm sounds are used for the first-level warning, the second-level warning, and the third-level warning.

[0104] It should be noted that the steps shown in the above process or the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0105] This embodiment also provides a bubble detection system based on machine vision, which is used to implement the above embodiments and preferred implementations, and will not be repeated here. As used below, the terms "module", "unit", "subunit", etc. can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0106] Figure 4 is a structural block diagram of a bubble detection system based on machine vision according to an embodiment of the present application, such as Figure 4 As shown, the system includes: a test container 41, a pressurizing device 42, an image acquisition device 43 and an information processing device 44.

[0107] The test container 41 contains a test liquid.

[0108] The pressurizing device 42 is connected to the upstream side of the target object to apply gas pressure to the upstream side of the target object immersed in the test liquid.

[0109] The image acquisition device 43 is used to acquire an image sequence of the test liquid environment, wherein the focal plane of the image acquisition device 43 forms an angle greater than 0° with the vertical plane.

[0110] The information processing device 44 is used to perform bubble recognition on the image sequence based on the image recognition algorithm, and determine whether the image sequence contains bubbles according to the recognition result to obtain the first image containing bubbles; according to the timestamp of the first image containing bubbles, determine the pressure applied by the pressurizing device at the corresponding moment.

[0111] Through the above system, the test container 41 contains the test liquid inside, the pressure device 42 is connected to the upstream side of the target object to apply gas pressure to the upstream side of the target object immersed in the test liquid, and the image acquisition device 43 is an image sequence of the test liquid environment, wherein the focal plane of the image acquisition device 43 forms an angle greater than 0° with the vertical plane, and the information processing device 44 performs bubble recognition on the image sequence based on the image recognition algorithm, and determines whether the image sequence contains bubbles according to the recognition result to obtain the first bubble-containing image; according to the timestamp of the first bubble-containing image, the pressure applied by the pressure device at the corresponding moment is determined to solve the problem of low accuracy of first bubble point detection. By acquiring the image sequence to detect the first bubble point, large manual detection errors are avoided and the accuracy of first bubble point detection is improved. The focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane, so that the generated bubbles will always pass through the focal plane of the image acquisition device at a certain moment, so that a clear image of the bubble can be obtained through a single image acquisition device.

[0112] In some embodiments, a light strip is arranged on the top of the test container. The test container has a transparent first side wall, and the image acquisition device is arranged on the side of the first side wall; the test container has a second side wall arranged opposite to the first side wall, and the second side wall has a light shielding layer.

[0113] A light strip is provided on the top of the test container, and a light shielding layer is provided on the second side wall. The light shielding layer may be an opaque coating or an opaque sticker. The light shielding layer may be located on the inner wall or outer wall of the test container to provide sufficient light for image acquisition, and at the same time, the light entering from the second side wall may affect the image acquisition, thereby making the image clearer and improving the accuracy of the detection. In addition, light shielding layers may also be provided on the two side walls adjacent to the first side wall and the second side wall.

[0114] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0115] Embodiment 2:

[0116] In the method application of the present embodiment, the target object is selected from: a pipe, a joint, a valve or a container, and the method further includes: closing the target object, applying a constant preset gas pressure to the upstream side of the target object, and determining the time when bubbles appear in the target product under the preset gas pressure according to the acquisition time corresponding to the first bubble-containing image.

[0117] It should be noted that sealing the target object means sealing the target object except for the inflation port connected to the pressurizing device to prevent gas leakage from other outlets.

[0118] It should be noted that the target object applicable to this embodiment is a container used for fluid containment and transportation. It is different from a filter element and a filter membrane in that it is more applicable to gas pressure holding tests. The time when the first bubble appears under the corresponding pressure can be calculated through the acquisition time corresponding to the first bubble-containing image, and then it can be understood whether the target object has the possibility of leakage under the predetermined pressure, which helps to more intuitively understand the gas pressure environment that the target object can meet, so as to prevent potential accidents and injuries. By detecting air tightness through this method, tiny changes in bubbles can be captured, thereby detecting extremely small leaks, improving the accuracy and sensitivity of detection; in addition, this method is suitable for a variety of different detection environments and objects, including products of different shapes, sizes and materials;

[0119] Furthermore, by adjusting the focal plane range by shifting the camera distance and panning or tilting the lens, this method may also accurately identify the leakage point.

[0120] Those skilled in the art should understand that the technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A bubble recognition method based on machine vision, characterized in that: The method comprises: Acquire an image sequence of a test liquid environment, wherein the upstream side of the target object has a gas cavity, the image sequence is acquired when gas pressure is applied to the upstream side of the target object, and the focal plane corresponding to the imaging of the image sequence forms an angle greater than 0° with the vertical plane; Bubble recognition is performed on the image sequence based on an image recognition algorithm, and whether bubbles are included in the image sequence is determined according to the recognition result to obtain the first bubble-containing image.

2. The method according to claim 1, characterized in that Before acquiring the image sequence of the test liquid environment, the method further comprises: discharging bubbles attached to the surface of the target object and / or dissolved in the test liquid by ultrasonic vibration.

3. The method according to claim 1 or 2, characterized in that: When acquiring the image sequence of the test liquid environment, the method further includes: driving bubbles existing in the blind spots of the target object to float up through ultrasonic vibration.

4. The method according to claim 1 or 2, characterized in that: When acquiring the image sequence of the test liquid environment, the method further includes: driving the target object to rotate in the test liquid so as to make bubbles existing in the blind spots of the target object float up.

5. The method according to claim 1, characterized in that The method further includes: acquiring an optical flow field of the bubble based on the image sequence; filtering a horizontal component of the optical flow field and extracting a vertical component of the optical flow field.

6. The method according to claim 5, characterized in that The method also includes: acquiring the movement rate of the pixel point corresponding to the bubble according to the optical flow field; setting an active bubble movement rate threshold based on the gas pressure upstream of the target object, and determining the display state of the pixel point according to the movement rate threshold and the movement rate.

7. The method according to claim 6, characterized in that The method further comprises: Based on a preset division rule, performing region division in the vertical direction of the bubble image to obtain at least two image regions; Based on the bubble movement rate of each of the image regions, movement rate thresholds corresponding to different image regions are set.

8. The method according to claim 6, characterized in that The display status includes: pixel value, gray value or pixel shape.

9. The method according to claim 1, characterized in that: The method further comprises: If bubbles are detected without reaching the preset initial pressure threshold, warning information indicating poor airtightness of the target object is generated based on the pressure value when the bubbles are detected and a preset warning classification mechanism.

10. The method according to claim 1, characterized in that The image acquisition device for acquiring the image sequence is selected from: any one of a tilt-shift camera and an industrial camera.

11. An application of the method according to any one of claims 1 to 10 in integrity testing, characterized in that: The target object is a filter element or a filter membrane.

12. The use according to claim 11, characterized in that: A gradually increasing gas pressure is applied to the upstream side of the target object, and the gas pressure applied to the upstream side of the target object at the acquisition time is determined according to the acquisition time corresponding to the first bubble-containing image.

13. The use according to claim 12, characterized in that: Acquire target images at multiple acquisition moments within a preset time period from the image sequence, and determine the volume of bubbles in each of the target images; If the bubble volume difference is within a preset difference range, it is determined that the target object is in a stable bubble state; The pressure data corresponding to the stable bubbling state is acquired, and the average pore diameter of the target object is determined according to the pressure data.

14. An application of the method according to any one of claims 1 to 10 in a tightness test, characterized in that: It is characterized in that The target object is a pipe, a joint, a valve or a container.

15. The use according to claim 14, characterized in that: The target object is sealed, and a constant preset gas pressure is applied to the upstream side of the target object, and the time when bubbles appear in the target product under the preset gas pressure is determined according to the acquisition time corresponding to the first bubble-containing image.

16. A bubble recognition system based on machine vision, characterized in that: The system includes: a test container, a pressurizing device, an image acquisition device and an information processing device, wherein: The test container contains a test liquid; The pressurizing device is connected to the upstream side of the target object to apply gas pressure to the upstream side of the target object immersed in the test liquid; The image acquisition device is used to acquire an image sequence of the target object and the test fluid environment, wherein a focal plane of the image acquisition device forms an angle greater than 0° with a vertical plane; The information processing device is used to perform bubble recognition on the image sequence based on an image recognition algorithm, and determine whether the image sequence contains bubbles according to the recognition result to obtain the first image containing bubbles; and determine the pressure applied by the pressurizing device at the corresponding moment according to the timestamp of the first image containing bubbles.

17. The bubble detection system according to claim 14, characterized in that: A light strip is provided on the top of the test container; The test container has a transparent first side wall, and the image acquisition device is arranged on the side of the first side wall; the test container has a second side wall arranged opposite to the first side wall, and the second side wall has a light shielding layer.

Images