Bearing bush gap film thickness measuring device based on machine vision and working method

Abstract

The invention discloses a bearing bush gap film thickness measuring device based on machine vision and a working method. The device comprises a bearing bush part, a camera part and a multi-shaft dual-drive mechanical arm part. Wherein the base part is fixed to the ground of a workshop, the multi-shaft dual-drive mechanical arm part, the bearing bush part and the camera part are all fixed to the upper portion of the base part, and the multi-shaft dual-drive mechanical arm part is located between the camera part and the bearing bush part; the multi-shaft dual-drive mechanical arm part is used for carrying out light source irradiation on the bearing bush part; a first multi-axis mechanical arm and a second multi-axis mechanical arm of the multi-axis dual-drive mechanical arm part are fixed to the base part, and colored strip-shaped light sources are clamped on finger joints of the first multi-axis mechanical arm and finger joints of the second multi-axis mechanical arm correspondingly. According to the invention, pictures collected by a plurality of angle light sources are compared, so that the phenomena of black spots, wiredrawing and impurities possibly existing on the surfaces of the alcohol film and the bearing bush can be accurately identified; and a large amount of manpower and material resources are prevented from being consumed for repeatedly loading and unloading the bearing bush and scraping.

Description

Technical Field

[0001] The invention belongs to the technical field of pumped storage rotor processing and manufacturing, and in particular relates to a device for measuring the thickness of a bearing bush gap film based on machine vision. Background Art

[0002] The main shaft and bearing shell, which weigh about 55 tons in total, are the parts where the rotor sliding bearing and the shaft neck contact. They are in the shape of a semi-cylindrical surface in the shape of a shell and are very smooth. When the sliding bearing is working, a very thin oil film is required between the bearing shell and the rotating shaft for lubrication. If the lubrication is poor, there will be direct friction between the bearing shell and the rotating shaft, and the friction will generate very high temperatures. The high temperature generated by direct friction is still enough to burn it. The bearing shell may also burn due to factors such as excessive load, high temperature, impurities in the lubricating oil or abnormal viscosity.

[0003] As the requirements for stator bearings in the machinery manufacturing industry increase, the thickness and quality of the oil film inside the bearings have become issues worthy of attention. At present, the bearings used in the market or factories are all processed in the workshop and then used on site. During processing, it is necessary to constantly adjust the corresponding positions of the upper and lower bearings, and repeatedly load and unload the bearings up and down and scrape them. In particular, the total weight of a set of bearings is about 55 tons, resulting in low work efficiency. Therefore, further improving the efficiency and quality of stator bearing processing and manufacturing is of great significance to the development of the rotor processing and manufacturing industry.

[0004] During the scraping process, alcohol is applied between the bearing and the shaft to keep them clean and tidy, but no lubricating oil is applied, thus forming an alcohol film. When the power station is constructed on site, lubricating oil is applied. The single-sided gap of this very thin alcohol film is about 0.1mm, which is difficult to identify with the naked eye; there is a curvature between the bearing and the shaft, and there is reflection at least at one point when illuminated, which is also difficult to identify by photography. Summary of the invention

[0005] In view of the problems existing in the background technology, the present invention provides a device for measuring the thickness of a bearing gap film based on machine vision, and the technical solution includes: a bearing part, a camera part, a multi-axis dual-drive mechanical arm part and a base part; wherein the base part is fixed to the workshop floor, the multi-axis dual-drive mechanical arm part, the bearing part and the camera part are all fixed above the base part, and the multi-axis dual-drive mechanical arm part is located between the camera part and the bearing part; the multi-axis dual-drive mechanical arm part irradiates the bearing part with a light source;

[0006] The camera part includes: an industrial camera lens, an upper top plate, a camera pitch rotation control mechanism, a column, an industrial computer, a lower support plate and a bottom plate, wherein the industrial camera lens is fixed on the upper top plate, the industrial camera lens faces the end surface of the bearing part, and the upper top plate is fixed to the top of the column; the industrial computer is fixed to the other side of the column through the lower support plate, and the bottom plate is fixed to the base part;

[0007] The multi-axis dual-drive robotic arm part includes: a light source controller, a color strip light source, a first multi-axis robotic arm and a second multi-axis robotic arm, wherein the first multi-axis robotic arm and the second multi-axis robotic arm are fixed to the base part, and a color strip light source is clamped on the finger joints of the first multi-axis robotic arm and the second multi-axis robotic arm, and the two color strip light sources are connected to the light source controller. The color strip light source is moved by the first multi-axis robotic arm and the second multi-axis robotic arm to hit the required position on the end face of the bearing, and the light source controller is placed on the upper surface of the base plate.

[0008] The first multi-axis robot arm and the second multi-axis robot arm have the same structure and both include: a spindle, a spindle motor, an elbow joint, an upper arm part, a rib plate, a first wrist joint, a dual motor box, a rotating platform, a second wrist joint and finger joints; the outer shell of the spindle motor is fixed to the base part, and the spindle is connected to the power output of the spindle motor to drive the spindle to rotate; the elbow joint is fixed to the upper end surface of the spindle, the first ends of the two parallel upper arm parts are hinged to the elbow joint, the two ends of the rib plate are respectively fixed to the middle parts of the two upper arm parts, the two ends of the first wrist joint are respectively hinged to the second ends of the two upper arm parts, and a first joint rotating gear is provided in the first wrist joint, and the first joint rotating gear is fixed to the upper arm part; the dual motor box The machine box is fixed to the first wrist joint, the shell of the first wrist driving motor and the shell of the second wrist rotating motor are both fixedly installed in the double motor box, the power drive of the first wrist driving motor is connected to the first joint rotating gear through gear meshing, the power drive of the second wrist rotating motor is connected to the second joint rotating gear through gear meshing, the second joint rotating gear is fixed in the second wrist joint, the second wrist joint is rotationally connected to the rotating platform, and the axes of the first wrist joint and the second wrist joint are perpendicular; the rotating platform is fixed to the second wrist joint, the root of the finger joint is fixed to the second wrist joint, a finger joint driving motor is installed in the second wrist joint, and a finger in the finger joint is connected to the finger joint driving motor through a gear set.

[0009] The two spindle motors, two first wrist drive motors, two second wrist rotation motors and two finger joint drive motors in the first multi-axis robotic arm and the second multi-axis robotic arm are all connected to the robotic arm control module in the industrial computer. The robotic arm control module controls the movement of the first multi-axis robotic arm and the second multi-axis robotic arm. The camera module in the industrial computer is connected to the industrial camera lens, and the direction control motor in the camera pitch and rotation control mechanism is connected to the camera posture adjustment module in the industrial computer.

[0010] The column is also provided with a grid laser emitter facing the end surface of the bearing bush.

[0011] The bearing part includes a hollow main shaft, a half-bearing to be tested, two support frames and a pad. The top ends of the two support frames with the same structure are arc-shaped support structures that match the shape of the half-bearing to be tested, and the lower ends of the support frames are fixed to the base part through the pad. An alcohol film is provided between the hollow main shaft and the half-bearing to be tested.

[0012] A working method of a bearing gap film thickness measuring device based on machine vision is also provided, and the technical solution includes:

[0013] Step A: taking multiple photos of the support frame at different light source positions, including:

[0014] Step A1: Take a photo of the end face of the bearing shell:

[0015] The light source controller turns on two colored strip light sources, and then the camera module calls the industrial camera lens to take the first photo of the end faces of the half-bearing and hollow spindle to be tested;

[0016] The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the first direction, and the camera module calls the industrial camera lens to take a second photo;

[0017] The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the second direction, and the photographing module calls the industrial camera lens to take a third photograph; the first direction and the second direction form an angle of 90° on the vertical plane;

[0018] Step A2, taking a photo of the inner cylindrical surface of the half bearing to be tested, including:

[0019] The hollow spindle is removed, the camera posture adjustment module is called, and the industrial camera lens is directed toward the inner cylindrical surface of the tile;

[0020] The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move and rotate the colored strip light source along the third direction, and the photographing module calls the industrial camera lens to take a fourth photograph to photograph the inner cylindrical surface of the bearing bush; the third direction is perpendicular to the plane formed by the first and second directions;

[0021] The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the fourth direction, and the camera module calls the industrial camera lens to take a fifth photo to take a photo of the inner cylindrical surface of the bearing bush again; the fourth direction rotates 45° around the axis of the third direction;

[0022] Step B: identify defects in each image;

[0023] Step C, scraping and grinding according to the identified image defects;

[0024] Step D: The technicians shall inspect no less than 1-3 contact points per square centimeter; if the technical requirement of 1-3 contact points per square centimeter is not met, return to step A.

[0025] The process of defect recognition for each image is as follows: first, the photos collected in step A are sorted, and the extracted feature maps are converted into final outputs through deep learning convolutional neural network technology, and black and white grayscale photos are output; the system recognizes the black and white grayscale photos and records the defect coordinates;

[0026] The defects include: black spots, wire drawing, impurities and missing alcohol film; when small black dot-like shadows appear in the picture, it is judged as a black spot; when white long and thin strips appear, it is judged as wire drawing; when irregularly shaped heterogeneous substances appear, it is judged as impurities; when a large area of ​​circular white edges appears, it is judged as missing alcohol film caused by uneven liquid flow.

[0027] Form an alcohol film before step A:

[0028] Place the hollow main shaft and the half-bearing to be tested on the support frame: place the hollow main shaft flat on the support, adjust it to a horizontal level, clean the working surface of the hollow main shaft with high-purity alcohol, industrial cotton yarn, etc., and fix the hollow main shaft on the second horizontal lathe after cleaning; clean the surface of the half-bearing to be tested and the matching half-bearing in the same way, spread a layer of red lead powder on the inner side of the half-bearing to be tested and the matching half-bearing after cleaning, hoist the half-bearing to be tested on the first horizontal lathe and fix it, hoist the hollow main shaft onto the half-bearing to be tested and stick it tightly, hoist the matching half-bearing and stick it tightly to the hollow main shaft and the half-bearing to be tested, and then synthesize the matching half-bearing and the half-bearing to be tested into an integral bearing; rotate the tightly stuck hollow main shaft, and a layer of alcohol film will be formed during the matching rotation of the hollow main shaft and the integral bearing, and then remove the matching half-bearing; place the half-bearing to be tested and the hollow main shaft on the support frame one after another.

[0029] After each step A is completed, the camera posture adjustment module is called to return the industrial camera lens to a position facing the end face of the bearing shell part.

[0030] The beneficial effects of the present invention are:

[0031] 1. The present invention provides a bearing gap film thickness measuring device based on machine vision, which adopts industrial camera data analysis to take pictures and accurately analyze the thickness of the alcohol film and different lighting angles; by comparing the pictures collected by light sources at multiple angles, it can accurately identify the possible defects of the alcohol film and the bearing surface, such as black spots, wire drawing, and impurities; avoiding the consumption of a large amount of manpower and material resources for repeated loading and unloading of bearings and scraping.

[0032] 2. Through multiple image processing, the alcohol film between the upper and lower bearings and the inner cylindrical surface of the half bearing to be tested are detected, and the thickness between the upper and lower bearings after installation is determined to analyze the defects of scraping. Since the single-sided gap between the upper and lower bearings is less than 0.1mm, the accidental errors and negligent errors that occurred when using 0.05mm and 0.08mm thick feeler gauges for dimensional detection are avoided, reducing the number of repeated loading and unloading of bearings and repeated scraping.

[0033] 3. After scraping, multiple vertical cross light lines are marked on the inner cylindrical surface of the bearing to facilitate the inspection of the technical requirement of no less than 1-3 contact points per square centimeter and reduce the workload of the marking workers. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the overall structure of an embodiment of a device for measuring the thickness of a bearing gap film based on machine vision according to the present invention;

[0035] Figure 2 A schematic diagram of the structure of a camera part in an embodiment of the present invention;

[0036] Figure 3 This is a schematic diagram of the structure of a multi-axis dual-drive robotic arm in an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram of the structure of a multi-axis robotic arm in an embodiment of the present invention;

[0038] Figure 5 This is a schematic diagram of the structure of a strip light source of a multi-axis mechanical arm in an embodiment of the present invention;

[0039] Figure 6 It is a schematic structural diagram of the bearing shell part in an embodiment of the present invention.

[0040] Among them, 1-camera part; 11-industrial camera lens; 12-upper top plate; 13-camera pitch and rotation control mechanism; 14-column; 15-industrial computer; 16-lower support plate; 17-light source controller; 18-bottom plate; 2-multi-axis dual-drive robotic arm part; 21-first multi-axis robotic arm; 22-second multi-axis robotic arm; 211-spindle; 212-elbow joint; 213-upper arm part; 214-rib 1; 215-first wrist joint; 216-dual motor box; 217-rotating platform; 218-second wrist joint; 219-finger joint; 22-right multi-axis robotic arm; 23-steel plate base; 24-color strip light source; 3-bearing part; 31-hollow spindle; 32-semi-bearing to be tested; 33-support frame; 34-pad. DETAILED DESCRIPTION

[0041] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0042] like Figure 1 to Figure 6 The embodiment of the present invention shown in the figure comprises: a bearing part 3, a camera part 1, a multi-axis dual-drive mechanical arm part 2 and a base part 4; wherein the base part 4 is fixed to the workshop floor, the multi-axis dual-drive mechanical arm part 2, the bearing part 3 and the camera part 1 are all fixed above the base part 4, and the multi-axis dual-drive mechanical arm part 2 is located between the camera part 1 and the bearing part 3; the camera part 1 takes a picture of the bearing part 3, and the multi-axis dual-drive mechanical arm part 2 performs a light source irradiation action on the bearing part 3;

[0043] like Figure 2 The camera part 1 shown includes: an industrial camera lens 11, an upper top plate 12, a camera pitch rotation control mechanism 13, a column 14, an industrial computer 15, a lower support plate 16 and a bottom plate 18, wherein the industrial camera lens 11 is a lens for taking photos at industrial sites, the industrial camera lens 11 is fixed by the camera pitch rotation control mechanism 13 and the upper top plate 12, and the upper top plate 12 is fixed to the top of the column 14; the industrial computer 15 is fixed to the other side of the column 14 through the lower support plate 16, and the bottom plate 18 is fixed to the base part 4;

[0044] In this embodiment, the direction control motor (not shown in the figure) in the camera pitch and rotation control mechanism 13 is connected to the camera posture adjustment module in the industrial computer 15, and the direction control motor in the camera pitch and rotation control mechanism 13 is used to control the industrial camera lens 11 to adjust the pitch angle (and a matching hinge structure is provided); the camera module in the industrial computer 15 is connected to the industrial camera lens 11.

[0045] like Figure 3 The multi-axis dual-drive mechanical arm part 2 shown includes: a light source controller 17, a steel plate base 23, a color strip light source 24, a first multi-axis mechanical arm 21 and a second multi-axis mechanical arm 22, wherein the first multi-axis mechanical arm 21 and the second multi-axis mechanical arm 22 are fixed to the base part 4 through the steel plate base 12, and a color strip light source 24 is clamped on the finger joints of the first multi-axis mechanical arm 21 and the second multi-axis mechanical arm 22, and the two color strip light sources 24 are connected to the light source controller 17. The color strip light source 24 is moved by the first multi-axis mechanical arm 21 and the second multi-axis mechanical arm 22 to hit the required position of the bearing part 3, and the light source controller 17 is placed on the upper surface of the bottom plate 18;

[0046] like Figure 3 and Figure 4As shown, the first multi-axis robot 21 and the second multi-axis robot 22 with the same structure both include: a main shaft 211, a main shaft motor (not shown in the figure), an elbow joint 212, a large arm part 213, a rib plate 214, a first wrist joint 215, a dual motor box 216, a rotating platform 217, a second wrist joint 218 and a finger joint 219; the shell of the main shaft motor is fixed to the base part 23, and the main shaft 211 is connected to the power output of the main shaft motor to drive the main shaft to rotate; the elbow joint 212 is fixed to the upper end surface of the main shaft 211, the first ends of the two parallel large arm parts 213 are hinged to the elbow joint 212, the two ends of the rib plate 214 are respectively fixed to the middle parts of the two large arm parts 213, the two ends of the first wrist joint 215 are respectively hinged to the second ends of the two large arm parts 213, and the first wrist joint 215 is provided with a first joint rotating gear (not shown in the figure), and the first joint rotating gear is fixed to the large arm part 213; the dual motor box 216, the rotating platform 217, the rotating platform 217, the rotating platform 218 and the finger joint 219; ... The machine box 216 is fixed to the first wrist joint 215, and the shell of the first wrist drive motor (not shown in the figure) and the shell of the second wrist rotation motor (not shown in the figure) are both fixedly installed in the dual motor box 216. The power drive of the first wrist drive motor is connected to the first joint rotation gear through gear meshing, and the power drive of the second wrist rotation motor is connected to the second joint rotation gear through gear meshing. The second joint rotation gear is fixed in the second wrist joint 218, and the second wrist joint 218 is rotationally connected to the rotating platform 217. The axes of the first wrist joint 215 and the second wrist joint 218 are perpendicular; the rotating platform 217 is fixed to the second wrist joint 218, and the root of the finger joint 219 is fixed to the second wrist joint 218. A finger joint drive motor is installed in the second wrist joint 218, and one finger in the finger joint 219 is connected to the finger joint drive motor through a gear set, thereby realizing the clamping and loosening of two fingers of the finger joint 219.

[0047] In this embodiment, the two spindle motors, two first wrist drive motors, two second wrist rotation motors and two finger joint drive motors in the first multi-axis robot arm 21 and the second multi-axis robot arm 22 are all connected to the robot arm control module in the industrial computer 15. The robot arm control module controls the movement of the two multi-axis robot arms, including the rotation of the spindle 211 and the elbow joint 212, the rotation of the upper arm part 213, the rotation of the first wrist joint 215 and the dual motor box 216, the rotation of the second wrist joint 218 and the rotating platform 217, the rotation of the second wrist joint 218 and the finger joint 219, and the opening and closing of the finger joint 219;

[0048] In this embodiment, a grid laser (not shown in the figure) is also provided directly above the bearing portion 3 and is directed toward the inner cylindrical surface of the half bearing 32 to be tested, so as to form a 1 cm 2 The square frame laser makes it easy for technicians to inspect the technical requirements of no less than 1-3 contact points per square centimeter;

[0049] like Figure 6 As shown, the bearing part 3 includes a hollow main shaft 31, a half-bearing 32 to be tested, two support frames 33 and two pads 34. An alcohol film (not shown in the figure) is provided between the hollow main shaft 31 and the half-bearing 32 to be tested, and the alcohol film is the part to be tested; the top ends of the two support frames 33 with the same structure are both arc-shaped support structures matching the shape of the half-bearing 32 to be tested, and the lower ends of the support frames 33 are fixed to the base part 4 through the pads 34, and the support frames 33 play the role of fixing the bearing part.

[0050] The working method of this embodiment is as follows:

[0051] The present invention uses an industrial camera lens 11 and a color strip light source 24 to detect the thickness of the alcohol film on the end face of the bearing, eliminating the need for repeated loading and unloading of upper and lower bearings and repeated scraping of the bearing surface, avoiding waste of manpower, material and financial resources, and saving time and labor costs. During operation, the position and angle of each joint in the multi-axis dual-drive mechanical arm part 2 are flexibly adjusted through a computer control device to achieve the optimal position for the color strip light source 24 to illuminate the end face of the bearing, and the posture of the industrial camera lens 11 is adjusted through the computer control part of the camera part 1 to achieve the optimal image of the industrial camera lens 11 recording the end face of the bearing part and the inner cylindrical surface of the half bearing to be tested; when the scraping is completed and inspected, multiple vertical cross light source lines are drawn on the inner cylindrical surface of the bearing, which is convenient for the inspection of the technical requirements of no less than 1-3 contact points per square centimeter, reducing the workload of the marking workers; this device is used for bearing processing, which improves the repeated and cumbersome problems in previous processing and further improves production efficiency.

[0052] In actual use, it includes:

[0053] Step 1, first form an alcohol film, and set the hollow main shaft 31 and the half-bearing 32 to be tested on the support frame 33: place the hollow main shaft 31 flat on the support, adjust it to a horizontal level, clean the working surface of the hollow main shaft 31 with high-purity alcohol, industrial cotton yarn, etc., and fix the hollow main shaft 31 on the second horizontal lathe after cleaning; clean the surface of the half-bearing 32 to be tested and the matching half-bearing (not shown in the figure, but a conventional component in this field) in the same way, and spread a layer of red lead powder on the inner side surfaces of the half-bearing 32 to be tested and the matching half-bearing after cleaning, and hoist the half-bearing 32 to be tested on the first horizontal lathe. The hollow main shaft 31 is hoisted on the half-bearing 32 to be tested and pressed tightly against it. The matching half-bearing is hoisted and pressed tightly against the hollow main shaft 31 and the half-bearing 32 to be tested. Then, the matching half-bearing and the half-bearing 32 to be tested are combined into an integral bearing. The tightly pressed hollow main shaft 31 is rotated. During the matching rotation of the hollow main shaft 31 and the integral bearing, a layer of alcohol film is formed. After that, the matching half-bearing is removed, and the half-bearing 32 to be tested and the hollow main shaft 31 are successively placed on the support frame 33. At this time, the alcohol film still exists between the half-bearing 32 to be tested and the hollow main shaft 31 for subsequent testing.

[0054] Step 2: taking multiple photos of the support frame 33 at different light source positions, including:

[0055] Step 21: Take a photo of the alcohol film on the end surface of the bearing:

[0056] The light source controller 17 turns on the two colored strip light sources 24, and then the photographing module calls the industrial camera lens 11 to take the first photograph of the end faces of the half bearing 32 and the hollow main shaft 31 to be tested;

[0057] The robot control module simultaneously calls the first multi-axis robot 21 and the second multi-axis robot 22 to move the color strip light source 24 along the first direction, and the photographing module calls the industrial camera lens 11 to take a second photograph;

[0058] The robot control module simultaneously calls the first multi-axis robot 21 and the second multi-axis robot 22 to move the colored strip light source 24 along the second direction, and the photographing module calls the industrial camera lens 11 to take a third photograph; the first direction and the second direction form an angle of 90° on the vertical plane;

[0059] Step 22, taking a photo of the alcohol film on the inner cylindrical surface of the half-bearing to be tested, including:

[0060] Move the hollow spindle 31 away, call the camera posture adjustment module, and tilt the industrial camera lens 11 upward by 30-60 degrees, that is, point the industrial camera lens toward the inner cylindrical surface of the half-bearing to be tested;

[0061] The robot control module simultaneously calls the first multi-axis robot 21 and the second multi-axis robot 22 to move and rotate the colored strip light source 24 along the third direction, and the photographing module calls the industrial camera lens 11 to take a fourth photograph to photograph the inner cylindrical surface of the bearing bush; the third direction is perpendicular to the plane formed by the first and second directions;

[0062] The robot control module simultaneously calls the first multi-axis robot 21 and the second multi-axis robot 22 to move the colored strip light source 24 along the fourth direction, and the photographing module calls the industrial camera lens 11 to take a fifth photograph, again photographing the inner cylindrical surface of the bearing bush; the fourth direction rotates 45° around the third direction axis; the industrial camera lens 11 remains in place;

[0063] Step 3, identify defects in the collected images; organize the photos collected in step 2, and use deep learning convolutional neural network technology (CNN) to transform the extracted feature map into the final output, and output black and white grayscale photos; the system identifies the black and white grayscale photos and records the defect coordinates; defects include: "black spots", "drawing", "impurities" and "lack of alcohol film"; specifically, if small black dot-shaped shadows appear in the picture, it is judged as "black spots"; if white elongated strips appear, it is judged as "drawing"; if irregular shaped heterogeneous substances appear, it is judged as "impurities"; if a large area of ​​circular white edges appears, it is judged as "lack of alcohol film" caused by uneven liquid flow;

[0064] Step 4: Scrape and grind according to the identified image defects;

[0065] Step 5: Then, draw multiple red vertical cross light lines on the inner cylindrical surface of the tile to form a 1cm 2 The technicians will inspect no less than 1-3 contact points per square centimeter. If the technical requirement of 1-3 contact points per square centimeter is not met, return to step 2.

[0066] In this embodiment, after each step 2 is completed, the camera posture adjustment module is called to return the industrial camera lens 11 to a position facing the end face of the bearing shell portion.

Claims

1. A device for measuring the thickness of a bearing gap film based on machine vision, characterized in that: include: A bearing part (3), a camera part (1), a multi-axis dual-drive mechanical arm part (2) and a base part (4); wherein the base part (4) is fixed to the workshop floor, the multi-axis dual-drive mechanical arm part (2), the bearing part (3) and the camera part (1) are all fixed above the base part (4), and the multi-axis dual-drive mechanical arm part (2) is located between the camera part (1) and the bearing part (3); The multi-axis dual-drive mechanical arm part (2) irradiates the bearing bush part (3) with a light source; The camera part (1) comprises: an industrial camera lens (11), an upper top plate (12), a camera pitch rotation control mechanism (13), a column (14), an industrial computer (15), a lower support plate (16) and a bottom plate (18), wherein the industrial camera lens (11) is fixed on the upper top plate (12), the industrial camera lens (11) faces the end surface of the bearing part (3), and the upper top plate (12) is fixed on the top of the column (14); the industrial computer (15) is fixed on the other side of the column (14) through the lower support plate (16), and the bottom plate (18) is fixed to the base part (4); The multi-axis dual-drive mechanical arm part (2) comprises: a light source controller (17), a color bar light source (24), a first multi-axis mechanical arm (21) and a second multi-axis mechanical arm (22), wherein the first multi-axis mechanical arm (21) and the second multi-axis mechanical arm (22) are fixed to the base part (4), a color bar light source (24) is clamped on the finger joints of the first multi-axis mechanical arm (21) and the second multi-axis mechanical arm (22), and the two color bar light sources (24) are connected to the light source controller (17), and the color bar light source (24) is illuminated at a required position on the end face of the bearing shell through the movement of the first multi-axis mechanical arm (21) and the second multi-axis mechanical arm (22), and the light source controller (17) is placed on the upper surface of the bottom plate (18).

2. The device for measuring the thickness of a bearing gap film based on machine vision according to claim 1, characterized in that: The first multi-axis mechanical arm (21) and the second multi-axis mechanical arm (22) have the same structure, and both include: a main shaft (211), a main shaft motor, an elbow joint (212), a large arm part (213), a rib plate (214), a first wrist joint (215), a dual motor box (216), a rotating platform (217), a second wrist joint (218) and a finger joint (219); the shell of the main shaft motor is fixed to the base part (23), and the main shaft (211) is connected to the main shaft The power output of the motor is connected to drive the main shaft to rotate; the elbow joint (212) is fixedly connected to the upper end surface of the main shaft (211), the first ends of the two parallel large arm parts (213) are hinged to the elbow joint (212), the two ends of the rib plate (214) are respectively fixed to the middle parts of the two large arm parts (213), the two ends of the first wrist joint (215) are respectively hinged to the second ends of the two large arm parts (213), the first wrist joint (215) is provided with a first joint rotation gear, and the first The joint rotation gear is fixed to the upper arm part (213); the double motor box (216) is fixed to the first wrist joint (215); the housing of the first wrist driving motor and the housing of the second wrist rotating motor are both fixedly installed in the double motor box (216); the power drive of the first wrist driving motor is connected to the first joint rotation gear through gear meshing; the power drive of the second wrist rotating motor is connected to the second joint rotation gear through gear meshing; the second joint rotating gear is fixed in the second wrist joint (218); the second wrist joint (218) is rotationally connected to the rotating platform (217); the axes of the first wrist joint (215) and the second wrist joint (218) are perpendicular; the rotating platform (217) is fixed to the second wrist joint (218); the root of the finger joint (219) is fixed to the second wrist joint (218); a finger joint driving motor is installed in the second wrist joint (218); a finger in the finger joint (219) is connected to the finger joint driving motor through a gear set.

3. The device for measuring the thickness of a bearing gap film based on machine vision according to claim 2, characterized in that: The two spindle motors, two first wrist drive motors, two second wrist rotation motors and two finger joint drive motors in the first multi-axis robotic arm (21) and the second multi-axis robotic arm (22) are all connected to a robotic arm control module in an industrial computer (15); the robotic arm control module controls the movement of the first multi-axis robotic arm (21) and the second multi-axis robotic arm (22); the camera module in the industrial computer (15) is connected to an industrial camera lens (11); and the direction control motor in the camera pitch rotation control mechanism (13) is connected to a camera posture adjustment module in the industrial computer (15).

4. The device for measuring the thickness of a bearing gap film based on machine vision according to claim 1, characterized in that: The column (14) is also provided with a grid laser emitter facing the end surface of the bearing bush.

5. The device for measuring the thickness of a bearing gap film based on machine vision according to claim 1, characterized in that: The bearing shell part (3) comprises a hollow main shaft (31), a half bearing shell (32) to be tested, two support frames (33) and a pad (34); the top ends of the two support frames (33) having the same structure are arc-shaped support structures matching the outer shape of the half bearing shell (32) to be tested, and the lower ends of the support frames (33) are fixed to the base part (4) via the pad (34); and an alcohol film is provided between the hollow main shaft (31) and the half bearing shell (32) to be tested.

6. A working method of the device for measuring the thickness of a bearing gap film based on machine vision as claimed in claim 1, characterized in that: include: Step A: taking multiple photos of the support frame at different light source positions, including: Step A1: Take a photo of the end face of the bearing shell: The light source controller turns on two colored strip light sources, and then the camera module calls the industrial camera lens to take the first photo of the end faces of the half-bearing and hollow spindle to be tested; The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the first direction, and the camera module calls the industrial camera lens to take a second photo; The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the second direction, and the photographing module calls the industrial camera lens to take a third photograph; the first direction and the second direction form an angle of 90° on the vertical plane; Step A2, taking a photo of the inner cylindrical surface of the half bearing to be tested, including: The hollow spindle is removed, the camera posture adjustment module is called, and the industrial camera lens is directed toward the inner cylindrical surface of the tile; The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move and rotate the colored strip light source along the third direction, and the photographing module calls the industrial camera lens to take a fourth photograph to photograph the inner cylindrical surface of the bearing bush; the third direction is perpendicular to the plane formed by the first and second directions; The robot control module simultaneously calls the first multi-axis robot and the second multi-axis robot to move the colored strip light source along the fourth direction, and the camera module calls the industrial camera lens to take a fifth photo to take a photo of the inner cylindrical surface of the bearing bush again; the fourth direction rotates 45° around the axis of the third direction; Step B: Defect identification for each image; Step C, scraping and grinding according to the identified image defects; Step D: The technicians shall inspect no less than 1-3 contact points per square centimeter; if the technical requirement of 1-3 contact points per square centimeter is not met, return to step A.

7. The working method of the device for measuring the thickness of the bearing gap film based on machine vision according to claim 6 is characterized in that: The process of defect recognition for each image is as follows: first, the photos collected in step A are sorted, and the extracted feature maps are converted into final outputs through deep learning convolutional neural network technology, and black and white grayscale photos are output; the system recognizes the black and white grayscale photos and records the defect coordinates; The defects include: black spots, wire drawing, impurities and missing alcohol film; when small black dot-like shadows appear in the picture, it is judged as a black spot; when white long and thin strips appear, it is judged as wire drawing; when irregularly shaped heterogeneous substances appear, it is judged as impurities; when a large area of ​​circular white edges appears, it is judged as missing alcohol film caused by uneven liquid flow.

8. The working method of the device for measuring the thickness of the bearing gap film based on machine vision according to claim 6 or 7, characterized in that: Form an alcohol film before step A: Place the hollow main shaft and the half-bearing to be tested on the support frame: place the hollow main shaft flat on the support, adjust it to a horizontal level, clean the working surface of the hollow main shaft with high-purity alcohol, industrial cotton yarn, etc., and fix the hollow main shaft on the second horizontal lathe after cleaning; clean the surface of the half-bearing to be tested and the matching half-bearing in the same way, spread a layer of red lead powder on the inner side of the half-bearing to be tested and the matching half-bearing after cleaning, hoist the half-bearing to be tested on the first horizontal lathe and fix it, hoist the hollow main shaft onto the half-bearing to be tested and stick it tightly, hoist the matching half-bearing and stick it tightly to the hollow main shaft and the half-bearing to be tested, and then synthesize the matching half-bearing and the half-bearing to be tested into an integral bearing; rotate the tightly stuck hollow main shaft, and a layer of alcohol film will be formed during the matching rotation of the hollow main shaft and the integral bearing, and then remove the matching half-bearing; place the half-bearing to be tested and the hollow main shaft on the support frame one after another.

9. The working method of the device for measuring the thickness of the bearing gap film based on machine vision according to claim 6 or 7, characterized in that: After each step A is completed, the camera posture adjustment module is called to return the industrial camera lens to a position facing the end face of the bearing shell part.

Images