Method for determining state of grease containing particles through image recognition

Abstract

The invention relates to a method for determining the state of a grease containing particles by means of image recognition. The method comprises the following steps: sampling the grease and smearing the grease on a filtering film; carrying out image recording on the sampled grease through a microscope, wherein the image recording comprises the steps of setting an amplification factor of the microscope, setting a distance from a lens of the microscope to the filtering film, and recording an image of the sampled grease through the microscope; performing image recognition on the image through artificial intelligence and determining the particle content in the grease; and determining the state of the grease according to the particle content.

Description

Technical Field

[0001] The present disclosure relates to a method for determining the state of grease containing particles through image recognition. Background Art

[0002] During the operation of mechanical equipment, lubricating grease is needed to lubricate the running equipment to reduce the friction between the parts in the mechanical equipment, thereby improving the performance of the mechanical equipment and protecting the mechanical equipment. Therefore, in order to achieve a better lubrication effect, the lubricating grease needs to maintain a good grease condition. On the contrary, a poor grease condition will lead to early abnormal operation of the mechanical equipment and later failure or damage.

[0003] Therefore, it is very important to understand the grease condition of lubricating grease. Traditional detection methods require special detection equipment or take a long time to detect. In order to determine the grease condition of lubricating grease more simply and quickly, a new detection method needs to be developed. Summary of the invention

[0004] The present disclosure provides a method for determining the state of grease containing particles by image recognition. When the grease state of lubricating grease is detected by image recognition at the equipment site, the method according to the present disclosure can improve the image recording quality of the grease image, thereby improving the speed and accuracy of image recognition. As a result, the grease state or the degree of grease contamination can be determined quickly and accurately, so that corresponding maintenance measures can be taken as early as possible.

[0005] The present disclosure provides a method for determining the state of grease containing particles through image recognition, the method comprising: sampling the grease and applying it on a filter film; recording an image of the sampled grease through a microscope, which comprises: setting the magnification of the microscope, setting the distance from the lens of the microscope to the filter film, and recording an image of the sampled grease through the microscope; performing image recognition on the image through artificial intelligence and determining the particle content in the grease; and determining the state of the grease based on the particle content.

[0006] In an embodiment according to the present disclosure, determining the state of grease according to the particle content includes: continuously setting a plurality of particle content intervals, wherein the particle content intervals classify the particle content, and a particle content interval with a larger value corresponds to a greater degree of grease contamination; and determining the degree of grease contamination according to the particle content interval in which the particle content is located.

[0007] In an embodiment of the present disclosure, when multiple image recordings of the sampled oil are performed through a microscope, if two of the particle contents determined based on the multiple image recordings are located in non-adjacent particle content intervals, it is assumed that the determination of the particle content is invalid; if the particle contents determined based on the multiple image recordings are located in adjacent particle content intervals, the degree of oil contamination is determined based on the particle content interval where the majority of the particle contents are located.

[0008] In an embodiment according to the present disclosure, a grease contamination indication is provided according to the grease contamination degree, wherein the grease contamination indication includes: normal grease, contamination warning and severe contamination.

[0009] In an embodiment of the present disclosure, recording an image of the sampled grease through a microscope further includes: using polarized light to make particles in the grease appear black.

[0010] In an embodiment according to the present disclosure, recording an image of the sampled oil and fat by means of a microscope further comprises: fixing a position of the microscope relative to the sampled oil and fat by means of a fixing bracket.

[0011] In an embodiment according to the present disclosure, the magnification of the microscope is 100 to 500 times.

[0012] In an embodiment according to the present disclosure, the filter membrane remains flat and is placed horizontally.

[0013] In an embodiment according to the present disclosure, the filter membrane is a filter membrane with a thickness of 5 μm to 40 μm.

[0014] In an embodiment according to the present disclosure, the microscope is a portable microscope.

[0015] In an embodiment of the present disclosure, the sampled oil and fat is image-recorded through a microscope, and the process also includes: determining whether the image quality of the image can be used for the image recognition, wherein, only when the image quality meets the requirements, the image is recognized through artificial intelligence and the particle content in the oil and fat is determined; and the state of the oil and fat is determined based on the particle content.

[0016] In an embodiment according to the present disclosure, performing image recognition on the image and determining the particle content in the oil through artificial intelligence includes: determining characteristic values ​​about the particles in the image through artificial intelligence, and determining the particle content based on the characteristic values ​​about the particles, wherein the characteristic values ​​about the particles include: statistical characteristic values ​​of the values ​​of pixels in the image, the number, density and total area of ​​the particles.

[0017] By using the method according to the present disclosure, the microscope and the image recording method can be reasonably adjusted and set to record the image or photo of the sampled grease with high quality. Under the guidance of the method according to the present disclosure, the grease state of the lubricating grease can be quickly and accurately determined at the equipment site. Therefore, the maintenance personnel of the mechanical equipment can conveniently and quickly perform the test on site, and can immediately determine whether the lubricating grease needs to be replaced, so that the sampled grease does not need to be brought back to the laboratory for testing, nor does it need to be tested using special testing equipment in the laboratory. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for describing the embodiments. The drawings described below are only exemplary embodiments of the present disclosure.

[0019] Figure 1 A flow chart of a method for determining the state of grease containing particles by image recognition according to an embodiment of the present disclosure is shown.

[0020] Figure 2 A schematic diagram showing a grease image that can be used for detection according to an embodiment of the present disclosure, and

[0021] Figure 3 A schematic diagram showing a grease image that cannot be used for detection according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present disclosure more obvious, the exemplary embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure, and it should be understood that the present disclosure is not limited to the exemplary embodiments described here.

[0023] In this specification and the accompanying drawings, substantially the same or similar steps and elements are represented by the same or similar reference numerals, and repeated descriptions of these steps and elements will be omitted. At the same time, in the description of the present disclosure, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance or ranking.

[0024] In this specification and the accompanying drawings, elements are described in singular or plural form, depending on the embodiment. However, the singular and plural forms are appropriately selected for the proposed situation only for the convenience of explanation and are not intended to limit the present disclosure thereto. Therefore, the singular form may include the plural form, and the plural form may also include the singular form, unless the context clearly indicates otherwise.

[0025] Figure 1A flow chart of a method 100 for determining the state of grease containing particles by image recognition according to an embodiment of the present disclosure is shown. The method 100 comprises: sampling the grease and applying it on a filter film (step S110); recording an image of the sampled grease by a microscope (step S120); performing image recognition on the image by artificial intelligence and determining the particle content in the grease (step S130); and determining the state of the grease according to the particle content (step S140).

[0026] In this embodiment, step S120 may, for example, include: setting the magnification of the microscope (step S121), setting the distance from the lens of the microscope to the filter membrane (step S122), and recording an image of the sampled oil through the microscope (step S123).

[0027] In an embodiment according to the present disclosure, the grease is especially the lubricating grease in the equipment, and the equipment is, for example, equipment with bearings, especially a wind turbine with huge bearings. In order to ensure the normal operation of the equipment, reduce wear and extend the life, it is necessary to add lubricating grease to the bearings. Despite the addition of lubricating grease, the operation of the bearings will also cause wear, so that metal particles, especially iron particles, will appear in the lubricating grease. Therefore, the analysis of the content of iron particles in the lubricating grease is an important indicator for detecting whether the lubricating grease needs to be maintained or replaced and whether the bearing is faulty.

[0028] The traditional grease detection method requires taking the sample to the laboratory for testing. This detection method requires a long detection time due to the long-distance transportation and experimental waiting time. This results in the inability to understand the grease status and the status of the corresponding equipment in a timely manner, which may increase the risk of equipment damage and further increase maintenance costs.

[0029] According to the method of the embodiment of the present disclosure, the grease state can be directly determined by image recognition at the equipment site without taking the grease to the laboratory. In method step S110, for example, the grease in the equipment can be sampled and applied to a filter film. In the embodiment according to the present disclosure, the filter film can be, for example, a filter film of 5 μm to 40 μm, so that particles in the grease, such as iron particles, can penetrate the filter film and be easy to observe and record images subsequently. The grease is then applied to the filter film, preferably evenly, and the application thickness can be, for example, within 3 mm, preferably not more than 2 mm. In addition, the filter film is preferably kept flat and placed horizontally, otherwise the images observed and recorded subsequently will appear blurred, thereby affecting the recognition effect of the iron particles.

[0030] In method step S120, the sampled grease is imaged by a microscope. The microscope may be, for example, a portable microscope to facilitate on-site observation and recording of grease images. In an embodiment according to the present disclosure, for example, the magnification of the magnifying glass may be adjusted first to clearly observe the particles in the grease. Too high a magnification may cause the outline of the filter film to be magnified, which may affect the subsequent identification of iron particles, while too low a magnification may result in the inability to distinguish between the background and the iron particles. The magnification is preferably 100 to 500 times. At the same time, for example, the distance from the lens of the microscope to the filter film may be set to clearly observe the particles in the grease. In an embodiment according to the present disclosure, a fixed bracket may be used to fix the position of the microscope relative to the sampled grease to ensure that no shaking occurs when the image is recorded by the microscope. In an embodiment according to the present disclosure, polarized light is also used to make the particles in the grease appear black, so as to facilitate the subsequent observation and identification of the iron particles.

[0031] Through some or all of the above method steps, a clear image can be obtained when recording the image of the sampled oil through a microscope, and then the image can be recognized by artificial intelligence and the particle content in the oil can be determined. Figure 2 A schematic diagram of a grease image that can be used for detection according to an embodiment of the present disclosure is shown. Figure 2 Figure a) in Figure 1 shows an image recording of grease containing particles recorded by a microscope. Iron particles can be clearly seen in the image, and the iron particles are evenly distributed in the image. The image can be used for subsequent artificial intelligence recognition to determine the particle content in the grease. Figure 2 Figure b) shows the iron particles identified by artificial intelligence, which are very clear and uniform.

[0032] Figure 3 A schematic diagram showing a grease image that cannot be used for detection according to an embodiment of the present disclosure is shown. Figure 3 The upper left corner of the figure a) is compared to Figure 2 Figure a) is rather blurry. This may be caused by an inappropriate distance between the microscope lens and the filter membrane, an uneven filter membrane, or uneven grease application. Figure 3 Figure b) in FIG. 1 shows iron particles identified by artificial intelligence. It can be seen from this image that no iron particles can be identified in the upper left corner of the image. Figure 3Figure a) in the figure cannot be used for subsequent image recognition. In this case, it is necessary to re-record the sampled grease or re-sample the grease. In an embodiment according to the present disclosure, recording the sampled grease by a microscope also includes: determining whether the image quality of the image can be used for the image recognition, wherein only when the image quality meets the requirements, performing image recognition on the image and determining the particle content in the grease by artificial intelligence; and determining the state of the grease according to the particle content.

[0033] In the embodiment according to the present disclosure, when there are multiple grease images, for example, it is necessary to name these grease images, and the naming may include, for example, equipment model, equipment location, equipment unit number, equipment part location, sampling point and sampling time. For example, the grease image may be named in the following manner, i.e., "equipment model-equipment location-equipment unit number-equipment part location-sampling point-sampling time". In the case where the equipment is a wind turbine, for example, the naming may be in the following manner, "wind turbine model-wind farm name-wind turbine unit number-bearing location-sampling point-sampling time".

[0034] In an embodiment according to the present disclosure, the artificial intelligence can be implemented, for example, by artificial intelligence algorithms such as deep learning algorithms, convolutional neural networks, etc., which determine the particle content in the grease by identifying the particles themselves in the grease, or the characteristics of the particles. In an embodiment according to the present disclosure, the image recognition function of the artificial intelligence can be implemented in a device, especially a portable device, for example, the image recognition function of the artificial intelligence can be integrated into a portable microscope. Alternatively, the image recognition function of the artificial intelligence can also be implemented on a server in the cloud. In this case, the image of the grease recorded by the microscope can be directly uploaded to the server in the cloud in the form of data, or a camera can be used to capture the image of the grease magnified by the microscope, and then the image can be uploaded to the server in the cloud, for example, through an application.

[0035] In an embodiment according to the present disclosure, performing image recognition on the image and determining the particle content in the oil by artificial intelligence includes: determining feature values ​​about the particles in the image by artificial intelligence, and determining the particle content based on the feature values ​​about the particles.

[0036] In an embodiment according to the present disclosure, the characteristic values ​​about the particles include: statistical characteristic values ​​of the values ​​of pixels in the image, the number, density and total area of ​​the particles.

[0037] An image can be considered as a collection of pixels, each of which has a corresponding value. For example, for a grayscale image, the higher the value of a pixel, the whiter / brighter the color of the pixel, and conversely, the smaller the value of a pixel, the blacker / darker the color of the pixel. In addition, for example, in an RGB (red, green, blue) color image, the value of a pixel can be, for example, the value of a pixel in one of the color channels. Therefore, if the value of the pixel is larger, the color is stronger, and conversely, the color is weaker. Alternatively, for example, in an HSV (hue, saturation, brightness) color image, the value of a pixel can be, for example, the value of a pixel in one of the channels. Therefore, for hue, the value of a pixel represents the corresponding color; for saturation, the value of a pixel represents the height of color purity; for brightness, the value of a pixel represents the brightness of a color. Since the image contains particles, the size of the value of a pixel in the image will be affected by the particles, for example, the pixels constituting the particles will have a larger value. Therefore, for example, a statistical analysis can be performed on the value of a pixel and a statistical characteristic value of the value of a pixel can be determined.

[0038] In addition, after obtaining an image containing particles, the particles can be identified from the image. For example, the number of particles can be directly used as a characteristic value about the particles, or the density of the particles in the image can be calculated based on the number of particles and the size of the image and used as a characteristic value about the particles, or the areas of all particles can be added together to obtain the total area of ​​the particles and use it as a characteristic value about the particles.

[0039] The characteristic value of particles is related to the particle content in the grease, so the particle content can be determined based on the characteristic value of particles. For example, the relationship between the characteristic value of particles and the particle content in the grease can be established by artificial intelligence, such as an artificial neural network.

[0040] In an embodiment according to the present disclosure, determining the state of grease according to the particle content (step S140) may, for example, include: continuously setting a plurality of particle content intervals, wherein the particle content intervals classify the particle content, and a particle content interval with a larger value corresponds to a greater degree of grease contamination; and determining the degree of grease contamination according to the particle content interval in which the particle content is located.

[0041] In an embodiment according to the present disclosure, two or more particle content intervals may be set, for example, five particle content intervals may be set, such as [0-first threshold ppm], [first threshold ppm-second threshold ppm], [second threshold-third threshold ppm], [third threshold ppm-fourth threshold ppm], [>fourth threshold ppm], wherein ppm (parts per million) is the unit of iron particle content in commonly used lubricating grease. These particle content intervals represent different levels of particle content in the grease, and the above-mentioned particle content intervals correspond to level one, level two, level three, level four and level five, respectively. The higher the level, the higher the particle content in the grease, that is, the more serious the contamination of the grease.

[0042] In the implementation according to the present disclosure, for example, a grease contamination indication can be provided according to the grease contamination degree, wherein the grease contamination indication includes: grease normal, contamination warning, and severe contamination. As shown in the table below, for the first-level particle content interval, the lubricating grease can be considered to be in a normal state; for the second-level and third-level particle content intervals, the lubricating oil can be considered to be slightly contaminated and therefore in a contamination warning state; for the fourth-level and fifth-level particle content intervals, the lubricating oil can be considered to be in a severe contamination state.

[0043]

[0044] In an embodiment according to the present disclosure, in order to determine the particle content in grease as accurately as possible, multiple image recordings and image recognitions may be performed, preferably 3, 5, 7 or more image recordings and image recognitions. If the same result is obtained through multiple image recordings and corresponding image recognitions, for example, the determined particle content is in the same particle content interval, the particle content level or pollution degree or grease state may be determined according to the corresponding particle content interval. If the various particle contents determined according to multiple image recordings are in different particle content intervals, it is necessary to consider each case separately. In the first case, if two of the determined particle contents are in non-adjacent particle content intervals, it is assumed that the determination of the particle content is invalid. For example, if one of the particle contents is in [0-first threshold ppm], and the other particle content is in [second threshold-third threshold ppm], in this case, it is considered that the determined particle content deviation is too large, so the determination of the particle content is invalid. In the second case, if the various particle contents determined according to multiple image recordings are in adjacent particle content intervals, the grease pollution degree is determined according to the particle content interval where the majority of the particle contents are located.

[0045] In an embodiment according to the present disclosure, when the particle content is divided into 5 particle content intervals and thus divided into 5 levels, the degree of grease contamination may be determined, for example, according to the rules in the following table.

[0046] According to the table below, when multiple images of grease are recorded, and these grease images are image-recognized by artificial intelligence and the particle content in the grease is determined respectively, if all the determined particle contents are level one, the grease state can be directly determined to be normal, and the grease does not need to be processed. If the determined particle contents include level one and level two, the grease state is determined according to the majority of cases or the cases with a high probability of occurrence. For example, when most of them are level one, the grease is determined to be normal and does not need further processing; and when most of them are level two, the grease state is determined to be a contamination warning and other methods can be used to further determine the particle content in the grease, such as taking the grease back to the laboratory for more accurate testing. If the determined particle content includes level two and / or level three and level four, the grease state is determined according to the majority of cases or the cases with a high probability of occurrence. For example, when most of them are level two and / or level three, the grease state is determined to be a contamination warning and other methods can be used to further determine the particle content in the grease; and when most of them are level four, the grease state can be determined to be severely contaminated and the grease needs to be cleaned or replaced. If the determined particle content includes level 2 and / or level 3 and level 5 or level 4 and / or level 5, it can be directly determined that the grease is severely contaminated and needs to be cleaned and replaced. In addition, if the determined particle content includes both level 1 and level 4 and / or level 5, it is considered that the result of the determined particle content has a large deviation, that is, the result is invalid. In this case, it is necessary to re-determine the particle content or re-collect the grease.

[0047]

[0048]

[0049] By using the method according to the present disclosure, the microscope and the image recording method can be reasonably adjusted and set to record the image or photo of the sampled grease with high quality. Under the guidance of the method according to the present disclosure, the grease state of the lubricating grease can be quickly and accurately determined at the equipment site. Therefore, the maintenance personnel of the mechanical equipment can conveniently and quickly perform the test on site, and can immediately determine whether the lubricating grease needs to be replaced, so that the sampled grease does not need to be brought back to the laboratory for testing, nor does it need to be tested using special testing equipment in the laboratory.

[0050] According to the method disclosed in the present invention, it is possible to determine the content of particles, such as iron particles, in the fluid to be detected, such as lubricating grease, only by using an image of the fluid to be detected, such as lubricating grease. Therefore, maintenance personnel of mechanical equipment can conveniently and quickly perform detection on site, and can immediately determine whether the lubricating grease needs to be replaced, so that there is no need to bring the sampled grease back to the laboratory for detection, nor to use special detection equipment in the laboratory for detection.

[0051] The block diagrams of the circuits, units, devices, apparatuses, equipment, and systems involved in the present disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these circuits, units, devices, apparatuses, equipment, and systems may be connected, arranged, and configured in any manner as long as the desired purpose can be achieved. The circuits, units, devices, and apparatuses involved in the present disclosure may be implemented in any suitable manner, such as using a dedicated integrated circuit, a field programmable gate array (FPGA), etc., or a general processing unit may be used in combination with a known program.

[0052] Those skilled in the art should understand that the above-mentioned specific embodiments are merely examples and not limitations, and various modifications, combinations, partial combinations and replacements may be made to the embodiments of the present disclosure according to design requirements and other factors. As long as they are within the scope of the attached claims or their equivalents, they belong to the scope of rights to be protected by the present disclosure.

Claims

1. A method for determining the state of grease containing particles by image recognition, comprising: sampling the grease and applying it to a filter membrane; The sampled oil is imaged and recorded by microscope, including: - setting the magnification of the microscope, - setting the distance from the lens of the microscope to the filter membrane, and - recording images of the sampled oil through the microscope; Performing image recognition on the image and determining the particle content in the grease by artificial intelligence; and The state of the grease is determined based on the particle content.

2. The method according to claim 1, wherein: Determining the state of the grease based on the particle content includes: A plurality of particle content intervals are continuously set, wherein the particle content intervals classify the particle content, and a particle content interval with a larger value corresponds to a greater degree of grease contamination; and The degree of grease contamination is determined according to the particle content interval in which the particle content falls.

3. The method according to claim 2, wherein: When multiple images of the sampled oil are recorded through a microscope, If two of the individual particle contents determined from the plurality of image recordings are located in non-adjacent particle content intervals, it is assumed that the determination of the particle content is invalid. If the particle contents determined based on multiple image records are located in adjacent particle content intervals, the degree of grease contamination is determined based on the particle content interval in which the majority of particle contents are located.

4. The method according to claim 2 or 3, wherein: A grease contamination indication is provided according to the grease contamination degree, wherein the grease contamination indication includes: normal grease, contamination warning and severe contamination.

5. The method according to claim 1, wherein: Recording the image of the sampled oil and fat through a microscope also includes: using polarized light to make the particles in the oil and fat appear black.

6. The method according to claim 1, wherein: Recording the image of the sampled oil and fat by means of a microscope also includes: fixing the position of the microscope relative to the sampled oil and fat by means of a fixing bracket.

7. The method according to claim 1, wherein: The microscope has a magnification of 100 to 500 times.

8. The method according to claim 1, wherein: The filter membrane remains flat and is placed horizontally.

9. The method according to claim 1, wherein: The sampled oil is imaged and recorded by microscope, including: determining whether the image quality of the image is capable of being used for the image recognition, Among them, only when the image quality meets the requirements, the image is recognized by artificial intelligence and the particle content in the oil is determined; and the state of the oil is determined according to the particle content.

10. The method according to claim 1, wherein: Performing image recognition on the image and determining the particle content in the grease by artificial intelligence includes: determining characteristic values ​​for the particles in the image by artificial intelligence, and Determining the particle content according to the characteristic value related to the particles, The characteristic values ​​about the particles include: statistical characteristic values ​​of the values ​​of the pixels in the image, the number, density and total area of ​​the particles.

Images