Intelligent ferrographic analysis device and method

Abstract

The invention relates to the field of mechanical equipment abrasion and oil abrasive particle detection, in particular to an intelligent ferrographic analysis device and method.The device comprises an imaging system, an oil inlet mechanism and an oil pump; the oil inlet mechanism is connected with the imaging system through an oil duct; the oil pump is arranged on the oil duct; the X-axis movement mechanism is arranged on the base, and the abrasive particle collection unit is installed on the X-axis movement mechanism; the Z-axis movement mechanism is arranged on a fixing support perpendicular to the base, and the camera assembly is installed on the Z-axis movement mechanism; and the oil duct is connected with the abrasive particle collecting unit. According to the method, the abrasive particle images at different horizontal positions and different heights are extracted and used for depth-of-field fusion and mapsheet splicing, clear imaging of all large and small abrasive particles in a deposition area in one image is achieved, local details are reserved, a global view angle is provided, and the image quality is improved. And in combination with a deep learning abrasive particle recognition analysis algorithm, the quantitative evaluation level of the wear state is greatly improved, the process is fully automatically completed, operation is simple and rapid, and the influence of manual operation on detection is avoided.

Description

Technical Field

[0001] The present invention relates to the field of mechanical equipment wear and oil particle detection, and specifically to an intelligent ferrograph analysis device and method. Background Art

[0002] The detection and monitoring of oil are common means to evaluate the conditions of various machines and predict various faults. The detection of abrasive particles in oil has become an essential method in modern industrial maintenance activities, and it is a conventional detection method to ensure the long-term effective and stable operation of various mechanical equipment, which can reduce the accident rate and improve work efficiency.

[0003] Generally, optical or electromagnetic detection methods are used, and both need to separately determine the grade quantity and number of large and small abrasive particles for analyzing and judging the abrasive particle state in the lubricating oil of mechanical equipment and judging the wear state of mechanical equipment.

[0004] For example, in the existing analytical ferrograph, its principle is to use a high-gradient magnetic field to adsorb the abrasive particles in the oil flow onto a plane to make a ferrograph, and then observe the wear particles under a microscope. However, the microscopic imaging field of view is limited, and only local observation and imaging can be carried out in regions, and the overall appearance of all deposited abrasive particles cannot be seen, which is not conducive to quantitative analysis, is easily interfered by human operation factors, and is also prone to omission of key and important abrasive particles. On the other hand, the ferrograph microscope cannot balance between large depth of field and high magnification, and mainly realizes the separate imaging of large and small abrasive particles by switching lenses with different magnifications, and it is very difficult to clearly image abrasive particles with sizes ranging from a few micrometers to hundreds of micrometers at the same time. In addition, the existing analytical ferrograph has a low degree of automation, and the production of ferrographs, the observation and imaging under the microscope are all manual operations, which are time-consuming and laborious and have high requirements for the skill level of users, and it is difficult to promote and apply.

[0005] Therefore, an intelligent ferrograph analysis method and device are needed, which can automatically and conveniently collect images of ferromagnetic abrasive particles in oil, extract multi-dimensional information and calculate indexes, so as to accurately and quickly judge the wear condition of mechanical equipment. Summary of the Invention

[0006] To overcome the problems existing in the prior art, the purpose of the present invention is to provide an intelligent ferrograph analysis device and method.

[0007] To achieve the above purpose, the present invention provides the following technical solution: An intelligent ferrograph analysis device includes an imaging system, an oil inlet mechanism and an oil pump; the oil inlet mechanism is connected to the imaging system through an oil channel, and the oil pump is arranged on the oil channel;

[0008] The imaging system includes a two-axis servo slide table mechanism, an abrasive particle collection unit, and a camera assembly; the two-axis servo slide table mechanism includes a base, an X-axis motion mechanism, a Z-axis motion mechanism, a servo driver, a servo motor, and a fixing bracket; the fixing bracket is vertically installed on the base; the X-axis motion mechanism is arranged on the base, and the abrasive particle collection unit is installed on the X-axis motion mechanism; the Z-axis motion mechanism is arranged on the fixing bracket, and the camera assembly is installed on the Z-axis motion mechanism; the X-axis motion mechanism and the Z-axis motion mechanism are respectively driven by the servo motor under the control of the servo driver to achieve motion; the oil passage is connected to the abrasive particle collection unit.

[0009] The present invention is further configured as: the intelligent ferrography analysis device further includes:

[0010] A servo driver, used to control the servo motor to achieve transmission positioning;

[0011] A cooling fan, used for the heat dissipation of the intelligent ferrography analysis device;

[0012] A filter, used to ensure the purity and stability of the input power supply;

[0013] A switching power supply, used to provide power for the intelligent ferrography analysis device;

[0014] A graphics processing card, installed on the main control board, used for the accelerated analysis, calculation, and display output of the abrasive particle image algorithm;

[0015] A main control board, used to achieve the overall control, information management, and data communication of the intelligent ferrography analysis device; the servo driver, the graphics processing card, and the oil pump are communicatively connected to the main control board.

[0016] An intelligent ferrography analysis method, used in cooperation with the above intelligent ferrography analysis device, includes the following steps:

[0017] S1: The main control board controls the oil pump to introduce the oil sample into the oil passage. When the oil sample passes through the abrasive particle collection unit, ferromagnetic wear particles in the oil are deposited and adsorbed on the glass surface of the deposition observation window of the abrasive particle collection unit by a high-gradient strong magnetic field.

[0018] S2: Adjust the horizontal position and height of the camera assembly in the imaging system, and take a number of original abrasive particle images at different horizontal positions and different heights from one end edge to the other end edge of the abrasive particle field of view.

[0019] S3: The graphics processing card performs preprocessing on the abrasive particle images, performs depth-of-field fusion and image stitching on the original abrasive particle images to obtain the abrasive particle target image.

[0020] S4: Send the wear particle target image to the Mask R-CNN network, extract the wear particle area through the segmentation network, obtain the accurate wear particle instance image, and obtain the shape contour of each wear particle;

[0021] S5: Performing graphic calculation analysis on the abrasive particle example diagram to obtain calculation indicators of each abrasive particle;

[0022] S6: input each abrasive particle instance image into the convolutional neural network to perform preliminary classification of characteristic abrasive particles, revise the classification status of some abrasive particles in combination with the calculation index of the abrasive particles obtained in step S4, and determine the characteristic abrasive particle classification information;

[0023] S7: According to the calculation index of each abrasive particle, the overall abrasive particle state index is obtained to evaluate the wear state of the mechanical equipment of the collected oil sample.

[0024] The present invention is further configured as follows: Step S1 specifically comprises the following steps:

[0025] S21: The imaging system is controlled to move by a two-axis servo slide mechanism, the X-axis motion mechanism drives the wear particle collection unit to move horizontally, with the edge of the deposition observation window as the horizontal starting position, and the Z-axis motion mechanism drives the camera assembly to the height of the glass surface where the abrasive particles of the deposition observation window are located, which can be clearly imaged, as the vertical starting position, to take the first wear particle image;

[0026] S22: Keeping the horizontal position of the imaging system unchanged, the Z-axis motion mechanism adjusts the vertical height upward, and the height adjustment is controlled within the imaging depth of field range, and the next wear particle image is taken. After completion, the aforementioned upward adjustment steps are repeated until the total height adjustment covers the maximum detection range of the wear particles, thereby obtaining multiple wear particle images with different heights at the same horizontal position;

[0027] S23: After completing the photography of abrasive grain images at different heights at a certain plane position, the Z-axis motion mechanism lowers the height to the vertical starting position, and the X-axis motion mechanism drives the abrasive grain collection unit to horizontally translate from the horizontal starting position toward the other edge of the deposition observation window. After moving into position, the first image at the current horizontal position is photographed, and then the horizontal position is kept unchanged, and step S22 is repeated to realize the layered image photography of the abrasive grains in the area;

[0028] S24: Repeat step S23 until the abrasive particle area photographed by the imaging system reaches the other end edge of the deposition observation window, completing the original collection of all abrasive particle images.

[0029] The present invention is further configured as follows: in step S23, the distance of the horizontal translation does not exceed the length of the single imaging field of view in the horizontal direction, and there is a certain overlap between the two imagings.

[0030] The present invention is further configured such that: specifically, in step S3, for all the layer images at the same horizontal position region with different heights, the clear parts in each image are extracted and synthesized into a depth-of-field target image by applying an image depth-of-field fusion processing algorithm; then, starting from the depth-of-field target image corresponding to the horizontal starting position, all the depth-of-field target images are sequentially stitched one by one by applying an image stitching algorithm until all the depth-of-field target images are stitched into a final abrasive grain target image with a full field-of-view range.

[0031] The present invention is further configured such that: specifically, in step S4, the Mask R-CNN network analyzes and identifies the abrasive grain target region and the background region based on an artificial intelligence semantic segmentation model, extracts the abrasive grain target region through a segmentation network to obtain an accurate abrasive grain instance image, and applies an edge detection algorithm to the abrasive grain target region to extract the outer contour of each abrasive grain.

[0032] The present invention is further configured such that: the calculation indexes of the abrasive grains in step S5 include the major axis size, the major / minor axis size ratio, the perimeter, the circularity, and the projected area of the abrasive grains.

[0033] The present invention is further configured such that: the overall abrasive grain state indexes include the percentage of the covered area of the abrasive grains in the abrasive grain target image, the quantity distribution of the abrasive grains in each size segment, the quantity of the abrasive grains with various category features, the proportion of large abrasive grains and the proportion of small abrasive grains, and the wear severity index.

[0034] The present invention is further configured such that: the wear severity index is calculated by the following formula

[0035] I S =D L ×(D L -D S )

[0036] wherein, D L is the proportion of large abrasive grains in the abrasive grain target image, and D S is the proportion of small abrasive grains in the abrasive grain target image.

[0037] It should be noted that the boundary between large abrasive grains and small abrasive grains varies according to different friction pairs of the monitored object. Abrasive grains with sizes exceeding the friction pair clearance are defined as large abrasive grains, and abrasive grains with sizes smaller than the friction pair clearance are defined as small abrasive grains.

[0038] In summary, the beneficial effects of the above technical solutions of the present invention are as follows:

[0039] The present invention provides an intelligent ferrographic analysis device, which can be used in conjunction with an intelligent ferrographic analysis method to extract abrasive images at different horizontal positions and different heights for depth-of-field fusion and image stitching. Moreover, the intelligent ferrographic analysis method of the present invention realizes clear imaging of all large and small abrasives in the deposition area in one image, retaining both local details and providing a global perspective, providing a good basis for abrasive images for qualitative observation and quantitative analysis. The operation is simple and fast, and the process is fully automated, avoiding the influence of manual operation on detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0041] Figure 1 It is a schematic diagram of the intelligent ferrographic analysis device in Embodiment 1.

[0042] Figure 2 It is a schematic diagram of the imaging system structure in Embodiment 1.

[0043] In the drawings, the meanings of the respective marks are as follows:

[0044] 1, main control board; 2, graphics processing card; 3, imaging system; 301, base; 302, X-axis movement mechanism; 303, servo motor; 304, camera assembly; 305, Z-axis movement mechanism; 306, fixed bracket; 307, abrasive collection unit; 4, cooling fan; 5, filter; 6, switching power supply; 7, servo driver; 8, oil pump; 9, oil inlet mechanism. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the following will clearly and completely describe the technical solutions of the present invention in conjunction with the drawings of the present invention. Based on the embodiments of the present invention, other similar embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0046] The present invention will be further described below in conjunction with the drawings and preferred embodiments.

[0047] Embodiment 1:

[0048] As Figure 1 - Figure 2 shown, it is a preferred embodiment of the present invention. An intelligent ferrographic analysis device includes an imaging system 3, an oil inlet mechanism 9, and an oil pump 8; the oil inlet mechanism 9 is connected to the imaging system 3 through an oil passage, and the oil pump 8 is arranged on the oil passage;

[0049] The imaging system 3 includes a two-axis servo slide table mechanism, an abrasive particle collection unit 307, and a camera assembly 304; the two-axis servo slide table mechanism includes a base 301, an X-axis movement mechanism 302, a Z-axis movement mechanism 305, a servo motor 303, and a fixed bracket 306; the fixed bracket 306 is installed on the base 301, and the fixed bracket 306 is perpendicular to the base 301; the X-axis movement mechanism 302 is arranged on the base 301, and the abrasive particle collection unit 307 is installed on the X-axis movement mechanism 302; the Z-axis movement mechanism 305 is arranged on the fixed bracket 306, and the camera assembly 304 is installed on the Z-axis movement mechanism 305; the X-axis movement mechanism 302 and the Z-axis movement mechanism 305 are respectively controlled by the servo motor 303; the oil passage is connected to the abrasive particle collection unit 307.

[0050] The intelligent ferrography analysis device further includes: a servo driver 7, which is used to control the servo motor 303 to realize the transmission and positioning of the two-axis servo slide table mechanism;

[0051] A cooling fan 4, which is used for the heat dissipation of the intelligent ferrography analysis device;

[0052] A filter 5, which is used to ensure the purity and stability of the input power supply;

[0053] A switching power supply 6, which is used to provide power for the intelligent ferrography analysis device;

[0054] A graphics processing card 2, which is installed on the main control board 1 and is used for the accelerated analysis and calculation of abrasive particle images;

[0055] The main control board 1 is used to receive or output signals to realize the control of the intelligent ferrography analysis device; the servo driver 7, the graphics processing card 2, and the oil pump 8 are communicatively connected to the main control board 1.

[0056] The oil pump 8 introduces an oil sample into the oil passage of the detection system. When passing through the abrasive particle collection unit 307, ferromagnetic wear particles in the oil are deposited and adsorbed on the deposition observation window by a high-gradient strong magnetic field. Through the optical imaging system, that is, the camera assembly 304, a microscopic image of the abrasive particles is obtained, and image analysis algorithms are applied based on the graphics processing card 2 to realize the analysis and calculation of the abrasive particles. Finally, based on the calculation results, the evaluation of the mechanical equipment wear state of the collected oil sample is realized.

[0057] Embodiment 2:

[0058] An intelligent ferrography analysis method, used in conjunction with the intelligent ferrography analysis device in Embodiment 1, includes the following steps:

[0059] S1: The main control board controls the oil pump to introduce the oil sample into the oil passage. When the oil sample passes through the abrasive particle collection unit, ferromagnetic wear particles in the oil are deposited and adsorbed on the glass surface of the deposition observation window of the abrasive particle collection unit by a high-gradient strong magnetic field.

[0060] S2: Adjust the horizontal position and height of the camera component in the imaging system, and take a number of original abrasive particle images at different horizontal positions and heights from one edge to the other edge of the abrasive particle field of view.

[0061] S21: The imaging system is controlled by the two-axis servo slide mechanism to move. The X-axis movement mechanism drives the abrasive particle collection unit to move horizontally, with the edge of the deposition observation window as the horizontal starting position, and the Z-axis movement mechanism drives the camera component to take the first abrasive particle image at the vertical starting position where the glass surface where the abrasive particles in the deposition observation window are located can be clearly imaged.

[0062] S22: Keep the horizontal position of the imaging system unchanged, and the Z-axis movement mechanism raises the vertical height within the imaging depth of field range, and takes the next abrasive particle image. After completion, repeat the above raising steps until the total elevation covered by the raising reaches the maximum detection range of the abrasive particles, thus obtaining multiple abrasive particle images at different elevations at the same horizontal position.

[0063] S23: After taking the abrasive particle images at different heights at a certain plane position, the Z-axis movement mechanism lowers the height to the vertical starting position, and the X-axis movement mechanism drives the abrasive particle collection unit to perform a horizontal translation from the horizontal starting position in the direction of the other edge of the deposition observation window range. After moving in place, take the first image at this horizontal position, and then keep the horizontal position unchanged, and repeat step S22 to realize the layered image shooting of the abrasive particles in this area; where the horizontal translation distance does not exceed the length of the single imaging field of view in the horizontal direction, and preferably the translation distance each time does not exceed 80% of the length of the imaging field of view in the moving direction.

[0064] S24: Repeat step S23 until the abrasive particle area photographed by the imaging system reaches the other edge of the deposition observation window, and complete the original acquisition of all abrasive particle images.

[0065] S3: The graphics processing card performs preprocessing on the abrasive particle images, performs depth of field fusion and image stitching on the original abrasive particle images, and obtains the abrasive particle target image.

[0066] The depth-of-field fusion specifically refers to: for the stratified images at all heights obtained in the same horizontal position area, feature extraction is performed on each captured image respectively, and methods such as gradient and Laplace transform are used to analyze the sharpness of each area in the image. According to the sharpness characteristics, the weight of each pixel in each image is calculated, and pixels with higher sharpness will be assigned higher weights. Compare the sharpness weights of the same area in each image, select the one with the highest weight and copy it to the newly created target image with the same size as the captured image. Repeat the above operations until the copying of all areas is completed, and finally save it as the depth-of-field fusion target image.

[0067] The specific process of the map splicing is as follows: Read in the depth-of-field fusion target image that was captured and processed first as the process image, and then read in the second depth-of-field fusion target image in chronological order as the image to be spliced; Use the feature point detection algorithm on the process image and the image to be spliced respectively to find the feature points in the images; Use the feature descriptor to compare the feature points in the process image and the image to be spliced to find the matching feature points; Based on the matching feature points in the process image and the image to be spliced, calculate the size of the area outside the overlapping part of the image to be spliced and the process image; Expand the size of the process image and paste the extra area in the image to be spliced onto the expanded process image to form a new spliced process image; Continue to read in the next depth-of-field fusion target image and repeat the above splicing steps until all depth-of-field fusion images are spliced and processed. Finally, the formed process image is trimmed at the field of view edge and saved as the spliced abrasive particle target image.

[0068] S4: Send the abrasive particle target image into the Mask R-CNN network, extract the abrasive particle area through the segmentation network to obtain an accurate abrasive particle instance image, and obtain the outer contour of each abrasive particle;

[0069] The Mask R-CNN network analyzes and identifies the abrasive particle target area and the background area based on the artificial intelligence semantic segmentation model, extracts the abrasive particle target area through the segmentation network to obtain an accurate abrasive particle instance image, applies the edge detection algorithm to the abrasive particle target area, and extracts the outer contour of each abrasive particle.

[0070] S5: Perform graphic calculation and analysis on the abrasive particle instance image to obtain the calculation indexes of each abrasive particle, including the major axis size, major / minor axis size ratio, perimeter, circularity, and projected area of the abrasive particle.

[0071] Among them, the major axis size is calculated according to the length of the minimum circumscribed rectangle of a single abrasive particle graph; the perimeter is the total length of the outer contour of a single abrasive particle identified in the image; the projected area is the planar area enclosed within the contour of a single abrasive particle; the circularity is e = (4π×projected area) / (contour perimeter×contour perimeter); the major / minor axis size ratio is the ratio of the major axis size to the minor axis size of a single abrasive particle identified in the image.

[0072] S6: Input the diagrams of each abrasive grain instance into a convolutional neural network for initial classification of characteristic abrasive grains. Revise the classification status of the abrasive grains by combining the calculated indexes of the abrasive grains obtained in step S4 to determine the classification information of the characteristic abrasive grains.

[0073] The convolutional neural network mainly identifies and classifies abrasive grains from their morphology. However, the definitions of different types of abrasive grains often also include some statistical-based empirical features. For example, the length of severe sliding abrasive grains is greater than 20 μm. It is necessary to use this empirical data to check and revise the results identified by the algorithm.

[0074] S7: Obtain the overall abrasive grain status index according to the calculated indexes of each abrasive grain.

[0075] The overall abrasive grain status index includes the percentage of the covered area of the abrasive grains in the abrasive grain target diagram, the number distribution of abrasive grains in each size segment, the number of characteristic abrasive grains of each category, the proportion of large abrasive grains and the proportion of small abrasive grains, and the wear severity index.

[0076] The percentage of the covered area of the abrasive grains is the percentage of the total area of the identified abrasive grains in the image area, reflecting the amount of wear. The number distribution of abrasive grains in each size segment is distinguished by the major axis size, defining multiple size intervals, and the numerical value of the number of abrasive grains counted according to each size interval segment reflects the amount of wear and the severity of wear. Abrasive grains are generally classified into 8 categories according to their characteristics, including flake abrasive grains, massive adhesive abrasive grains, sliding abrasive grains, etc., corresponding to different wear conditions, and the wear conditions can be judged according to the characteristic categories.

[0077] The wear severity index is calculated by the following formula:

[0078] I S = D L × (D L - D S )

[0079] where D L is the proportion of large abrasive grains in the abrasive grain target diagram, and D S is the proportion of small abrasive grains in the abrasive grain target diagram.

[0080] Based on the abrasive grain status index, the wear status of the mechanical equipment of the collected oil sample is evaluated. The larger the percentage of the covered area of the abrasive grains, the larger the proportion of large abrasive grains or the higher the wear severity index, the more serious the wear of the mechanical equipment.

[0081] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than a limitation on the protection scope of the present invention. Any simple modification or equivalent replacement of the technical solution of the present invention by those of ordinary skill in the art shall not depart from the essence and scope of the technical solution of the present invention.

Claims

1. An intelligent ferrography analysis device, characterized in that: It includes an imaging system, an oil inlet mechanism and an oil pump; the oil inlet mechanism is connected to the imaging system through an oil channel, and the oil pump is arranged on the oil channel; The imaging system includes a two-axis servo slide mechanism, an abrasive particle collection unit and a camera assembly; the two-axis servo slide mechanism includes a base, an X-axis motion mechanism, a Z-axis motion mechanism, a servo driver, a servo motor and a fixed bracket; the fixed bracket is vertically installed on the base; the X-axis motion mechanism is arranged on the base, and the abrasive particle collection unit is installed on the X-axis motion mechanism; the Z-axis motion mechanism is arranged on the fixed bracket, and the camera assembly is installed on the Z-axis motion mechanism; the X-axis motion mechanism and the Z-axis motion mechanism are respectively driven by the servo motor to complete the action under the control of the servo driver; the oil channel is connected to the abrasive particle collection unit.

2. The intelligent ferrography analysis device according to claim 1, characterized in that: The intelligent ferrography analysis device also includes: A servo driver, used to control the servo motor to achieve transmission positioning; Cooling fan, used for cooling the intelligent ferrography analysis device; Filter, used to ensure the purity and stability of input power; A switching power supply, used to provide power to the intelligent ferrography analysis device; Graphics processing card, installed on the main control board, used for accelerated analysis, calculation and display output of wear particle image algorithm; The main control board is used to realize the overall control, information management and data communication of the intelligent ferrography analysis device; the servo driver, the graphic processing card and the oil pump are communicatively connected with the main control board.

3. An intelligent ferrographic analysis method, used in conjunction with the intelligent ferrographic analysis device according to claim 2, characterized in that: The following steps are involved: S1: The main control board controls the oil pump to introduce the oil sample into the oil channel. When the oil sample passes through the wear particle collection unit, the high-gradient strong magnetic field deposits and adsorbs the ferromagnetic wear particles in the oil onto the glass surface of the deposition observation window of the wear particle collection unit. S2: Adjust the horizontal position and height of the camera assembly in the imaging system, and take a number of original wear particle images at different horizontal positions and heights from one edge of the wear particle field of view to the other edge; S3: The graphics processing card performs wear particle image preprocessing, performs depth of field fusion and image splicing on the original wear particle image, and obtains the wear particle target image; S4: Send the wear particle target image to the Mask R-CNN network, extract the wear particle area through the segmentation network, obtain the accurate wear particle instance image, and obtain the shape contour of each wear particle; S5: Performing graphic calculation analysis on the abrasive particle example diagram to obtain calculation indicators of each abrasive particle; S6: input each abrasive particle instance image into the convolutional neural network to perform preliminary classification of characteristic abrasive particles, revise the abrasive particle classification state in combination with the calculation index of the abrasive particles obtained in step S4, and determine the characteristic abrasive particle classification information; S7: According to the calculation index of each wear particle, the overall wear particle state index is obtained to evaluate the wear state of the mechanical equipment of the collected oil sample.

4. The intelligent ferrographic analysis method according to claim 3, characterized in that: Step S1 specifically includes the following steps: S21: The imaging position is controlled by the two-axis servo slide mechanism, the X-axis motion mechanism drives the abrasive particle collection unit to move horizontally, with the edge of the deposition observation window as the horizontal starting position, and the Z-axis motion mechanism drives the camera assembly to take the height of the clear imaging of the glass surface where the abrasive particles of the deposition observation window are located as the vertical starting position, and the first abrasive particle image is taken; S22: Keeping the horizontal position of the wear particle collection unit unchanged, the Z-axis motion mechanism drives the camera assembly to adjust the vertical height upward, and the height adjustment is controlled within the imaging depth of field range, and the next wear particle image is taken. After completion, the above-mentioned upward adjustment steps are repeated until the total height adjustment covers the maximum detection range of the wear particles, thereby obtaining multiple wear particle images with different heights at the same horizontal position; S23: After completing the photography of abrasive grain images at different heights at a certain plane position, the Z-axis motion mechanism lowers the height to the vertical starting position, and the X-axis motion mechanism drives the abrasive grain collection unit to horizontally translate from the horizontal starting position toward the other edge of the deposition observation window. After moving into position, the first image at the current horizontal position is photographed, and then the horizontal position is kept unchanged, and step S22 is repeated to realize the layered image photography of the abrasive grains in the area; S24: Repeat step S23 until the abrasive particle area photographed by the imaging system reaches the other end edge of the deposition observation window, completing the original collection of all abrasive particle images.

5. The intelligent ferrographic analysis method according to claim 4, characterized in that: In step S23, the distance of the horizontal translation does not exceed the length of the single imaging field of view in the horizontal direction, and there is a certain overlap between the two imagings.

6. The intelligent ferrographic analysis method according to claim 4, characterized in that: Step S3 specifically comprises: for all layered images of all heights obtained in the same horizontal position area, an image depth of field fusion processing algorithm is applied to extract the clear part of each image and synthesize it into a depth of field target image; then starting from the depth of field target image corresponding to the horizontal starting position, all depth of field target images are stitched one by one in the order of shooting using an image stitching algorithm until all depth of field target images are stitched into a final abrasive target image with a full field of view.

7. The intelligent ferrographic analysis method according to claim 3, characterized in that: Specifically, step S4 includes: the Mask R-CNN network analyzes and identifies the abrasive target area and background area based on the artificial intelligence semantic segmentation model, extracts the abrasive target area through the segmentation network to obtain an accurate abrasive instance map, applies the edge detection algorithm to the abrasive target area, and extracts the shape contour of each abrasive particle.

8. The intelligent ferrographic analysis method according to claim 3, characterized in that: The calculation indexes of the abrasive particles in step S5 include the major axis size, major / minor axis size ratio, circumference, circularity and projected area of ​​the abrasive particles.

9. The intelligent ferrography analysis method according to claim 3, characterized in that: The overall abrasive particle state index includes the coverage area percentage of the abrasive particles in the abrasive particle target map, the quantity distribution of abrasive particles in each size segment, the quantity of characteristic abrasive particles in each category, the proportion of large abrasive particles and the proportion of small abrasive particles, and the wear severity index; The wear severity index is calculated by the following formula: I S =D L ×(D L -D S ); Among them, D L is the proportion of large abrasive particles in the abrasive target map, D S is the proportion of small abrasive particles in the abrasive target map.

Images