Cylindrical roller sliding-rolling ratio measuring method based on three-dimensional digital image correlation

Abstract

The invention discloses a cylindrical roller sliding-rolling ratio measuring method based on three-dimensional digital image correlation. The method is based on a binocular camera composed of a visual platform capable of bearing a to-be-measured bearing and driving an inner ring of the to-be-measured bearing to rotate and a left camera and a right camera with the same frame rate. Comprising the following steps: S1, placing a to-be-detected bearing on a visual platform, respectively manufacturing different speckle patterns on the end face of each roller at the same side end of the to-be-detected bearing, and driving the inner ring of the to-be-detected bearing to rotate at a preset inner ring driving rotating speed so as to enable the to-be-detected bearing to enter a running state, the binocular camera performs synchronous frame-by-frame acquisition on the end face image of the bearing to be measured in the running state; s2, selecting a reference subarea in the speckle pattern of one roller shot by the left camera, and taking the initial frame of the speckle pattern of the same roller shot by the left camera and the right camera as the current frame for stereo matching; according to the method, the sliding-rolling ratio and the change of each cylindrical roller in the bearing operation process can be measured, so that the real-time dynamic fluctuation of the sliding-rolling ratio generated along with the change of working conditions is reflected.

Description

Technical Field

[0001] The invention relates to the technical field of cylindrical roller bearing detection, and in particular to a cylindrical roller sliding-rolling ratio measurement method based on three-dimensional digital image correlation. Background Art

[0002] When the bearing is working, the roller not only bears the load, but also moves relative to the inner and outer rings. The sliding-rolling ratio directly reflects the friction characteristics between the rollers and raceways in the bearing, and is an important indicator for evaluating the friction performance and working status of the bearing. Maintaining a reasonable sliding-rolling ratio is crucial for the normal operation of the bearing. A sliding-rolling ratio that is too large or too small will cause bearing wear, noise and temperature rise, affecting the life and reliability of the bearing. Therefore, accurate measurement of the sliding-rolling ratio of the bearing is one of the keys to ensuring bearing performance and life.

[0003] The invention patent with publication number CN110118656B provides a device and method for measuring the friction and sliding-rolling ratio between the bearing steel ball and the inner and outer rings, which uses an encoder and a laser sensor to simultaneously measure the friction and sliding-rolling ratio between the roller and the inner and outer rings. This measurement method has the following problems: First, it cannot accurately measure the sliding-rolling ratio of each roller, but can only provide the average sliding-rolling ratio of the entire bearing, and cannot reflect the specific working status of individual rollers. Secondly, this measurement method cannot reflect the dynamic fluctuations of the sliding-rolling ratio caused by changes in working conditions in real time.

[0004] With the development of computer vision and digital image processing technology, the measurement method of bearing sliding-rolling ratio based on computer vision has gradually become a research hotspot. Since the bearing itself lacks obvious texture features, it poses a considerable challenge to the current image processing technology. Therefore, a measurement method of cylindrical roller bearing rotating components based on three-dimensional digital image correlation is carried out, and a cylindrical roller sliding-rolling ratio measurement method based on three-dimensional digital image correlation is proposed. Summary of the invention

[0005] The purpose of the present invention is to propose a cylindrical roller slip-roll ratio measurement method based on three-dimensional digital image correlation, which can measure the slip-roll ratio of each cylindrical roller and its changes during the operation of the bearing, thereby reflecting the real-time dynamic fluctuations of the slip-roll ratio caused by changes in working conditions.

[0006] The technical solution adopted by the present invention is: a cylindrical roller sliding-rolling ratio measurement method based on three-dimensional digital image correlation, the method is based on a visual platform capable of carrying the bearing to be measured and driving the inner ring of the bearing to be measured to rotate, and a binocular camera composed of a left camera and a right camera with the same frame rate;

[0007] The following steps are involved:

[0008] S1. Place the bearing to be tested on a visual platform, and make different speckle patterns on the end face of each roller on the same side of the bearing to be tested. Drive the inner ring of the bearing to be tested to rotate at a preset inner ring driving speed to put it into operation. Use a binocular camera to synchronously capture the end face image of the bearing to be tested frame by frame in the operation state.

[0009] S2, selecting a reference sub-region in the speckle pattern of a roller photographed by the left camera, taking the initial frame of the same roller speckle pattern photographed by the left camera and the right camera as the current frame for stereo matching, and determining a matching sub-region on the speckle pattern photographed by the right camera that matches the reference sub-region;

[0010] Based on the reference sub-area and matching sub-area of ​​the initial frame of the same roller speckle pattern photographed by the left camera and the right camera, the three-dimensional coordinate point cloud P of the reference sub-area of ​​the initial frame speckle pattern is obtained. 0 ;

[0011] S3, performing time sequence matching between the current frame of the same roller speckle pattern photographed by the left camera and the next frame thereof, and determining a target sub-region on the next frame of speckle pattern photographed by the left camera that matches the reference sub-region;

[0012] S4, taking the target sub-region of the speckle pattern of the same roller in the next frame photographed by the left camera as a new reference sub-region, performing stereo matching on the speckle patterns of the same roller in the next frame photographed by the left camera and the right camera, and determining a new matching sub-region on the speckle pattern of the next frame photographed by the right camera that matches the new reference sub-region;

[0013] Based on the new reference sub-region and the new matching sub-region of the next frame of the same roller speckle pattern photographed by the left camera and the right camera, the three-dimensional coordinate point cloud P of the new reference sub-region of the next frame of speckle pattern is obtained. 1 ;

[0014] S5, repeat S3-S4 to obtain the three-dimensional coordinate point cloud P of the reference sub-area of ​​each frame of the speckle pattern of the same roller 2 , P 3 , P 4 ,…P i ;

[0015] S6. Obtain the pose matrix between the three-dimensional coordinate point clouds of adjacent frames through the point cloud registration algorithm, and obtain the rotation matrix R of the pose matrix through singular value decomposition to obtain the inter-frame rotation angle α of the corresponding roller adjacent frames i ;

[0016] S7, according to the inner ring driving speed, the frame rotation angle α i The instantaneous slide-to-roll ratio of the corresponding roller is calculated using the camera frame rate.

[0017] As a preferred solution, the search range is determined by epipolar constraints during stereo matching, each coordinate point along the epipolar line in the speckle pattern photographed by the right camera is taken as the center, and areas with the same size as the reference sub-area of ​​the left camera are selected in sequence around the center as target sub-areas for stereo matching, the correlation between each target sub-area and the reference sub-area is calculated by the zero-mean normalized mutual correlation coefficient, and the target sub-area with the highest correlation is selected as the matching sub-area of ​​the reference sub-area.

[0018] As a preferred solution, during stereo matching, the reverse Gauss-Newton iteration method is used to iterate and obtain the sub-pixel coordinates of each pixel point in the matching sub-area.

[0019] As a preferred solution, during stereo matching, three-dimensional reconstruction is performed based on the coordinates of each pixel point in the reference sub-area and the coordinates of the pixel points in the corresponding matching sub-area to obtain the three-dimensional coordinates of each pair of pixel points in the reference sub-area of ​​the roller speckle pattern.

[0020] As a preferred solution, the rotation matrix R is expressed as:

[0021] q m =Rp m +t

[0022] Where: p is the feature point coordinate of the current frame, p = (p 1 ,p 2 ,…,p m );q is the feature point coordinate of the next frame, q=(q 1 ,q 2 ,…,q m ), m is the feature point number, m = 1, 2, ..., n; R contains three rotation displacements; t is the translation vector describing the movement before and after, including three translation displacements.

[0023] As a preferred solution, an error function is constructed, and the optimal transformation of the rotation matrix R is solved by the least squares equation so that the mean square error between the point sets corresponding to the feature points is minimized.

[0024] As a preferred solution, the frame rotation angle α i It is expressed as:

[0025]

[0026] Where: The rotation matrix R is regarded as a composite matrix of three basic rotation matrices Rx, Ry, and Rz that rotate around the rotation axis.

[0027] As a preferred solution, the instantaneous slide-to-roll ratio ξ i The calculation formula is as follows:

[0028]

[0029] Where: n is the inner ring driving speed; R 1 is the inner circle radius; R 2 is the roller radius; f is the camera frame rate.

[0030] As a preferred solution, steps S2-S7 are repeated to obtain the instantaneous slide-to-roll ratio of each roller.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. The present invention provides a method for measuring the sliding-rolling ratio of a cylindrical roller bearing based on three-dimensional digital image correlation. The method based on three-dimensional digital image correlation is used to achieve accurate measurement of the sliding-rolling ratio of each cylindrical roller bearing;

[0033] 2. The high frame rate and fine image capture capability of high-speed cameras can accurately record the relative motion between each roller and the inner and outer rings in the bearing, and then dynamically analyze the change process of the sliding-rolling ratio;

[0034] 3. The image data collected by the high-speed camera can be traced back at any time, and the data of different time periods during the operation of the bearing can be freely selected for detailed analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 Schematic diagram of the matching process of the speckle patterns of adjacent frames in the present invention;

[0037] Figure 2 This is a flow chart of the point cloud timing matching of the present invention.

[0038] Figure numerals: 1. outer ring, 2. inner ring, 3. roller, 4. speckle pattern, 5. search range, 6. reference sub-area, 7. new reference sub-area, 8. matching sub-area, 9. new matching sub-area, 10. new search range. DETAILED DESCRIPTION

[0039] The present invention is described in detail below by way of exemplary embodiments. However, it should be understood that, without further description, elements, structures and features in one embodiment may also be beneficially combined in other embodiments.

[0040] It should be noted that: unless otherwise defined, the technical terms or scientific terms used in this document shall have the usual meanings understood by people with ordinary skills in the field to which the invention belongs. The words "one", "an" or "the" and the like used in the patent application specification and claims of the present invention do not express quantity restrictions, but indicate the existence of at least one; the words "first", "second" and "third" used in this document shall not be regarded as restrictions on the order of components, but are only used to distinguish different components; words such as "include" or "comprise" and the like indicate that the elements or objects appearing before "include" or "comprise" include the elements or objects listed after "include" or "comprise" and their equivalents, but do not exclude other elements or objects with the same function.

[0041] In order to more clearly describe the cylindrical roller sliding-rolling ratio measurement method based on three-dimensional digital image correlation, combined with the attached Figure 1-2 Describe this embodiment:

[0042] like Figure 1 and Figure 2 As shown, a cylindrical roller sliding-rolling ratio measurement method based on three-dimensional digital image correlation is based on a visual platform capable of carrying a bearing to be measured and driving the inner ring of the bearing to be measured to rotate, and a binocular camera consisting of a left camera and a right camera with the same frame rate; the cylindrical roller bearing to be measured includes an outer ring 1, an inner ring 2 and a roller 3;

[0043] The following steps are involved:

[0044] S1. Place the bearing to be tested on a visual platform, and make different speckle patterns 4 on the end faces of each roller 3 on the same side of the bearing to be tested. Drive the inner ring 2 of the bearing to be tested to rotate at a preset inner ring driving speed to put it into operation. Use a binocular camera to synchronously capture the end face image of the bearing to be tested in the operation state frame by frame. The view obtained by the left camera is called the left view, and the view obtained by the right camera is called the right view.

[0045] S2, select a reference sub-area 6 in the speckle pattern 4 of a roller photographed by the left camera (select the current frame image photographed by the left camera as the reference image, and select a rectangular area of ​​20×15 pixels in the center of the speckle pattern of a roller as the reference sub-area), take the initial frame of the same roller speckle pattern photographed by the left camera and the right camera as the current frame for stereo matching, and determine a matching sub-area 8 on the speckle pattern photographed by the right camera that matches the reference sub-area 6;

[0046] Based on the reference sub-area 6 and matching sub-area 8 of the initial frame of the same roller speckle pattern photographed by the left camera and the right camera, a three-dimensional coordinate point cloud P of the reference sub-area 6 of the initial frame speckle pattern is obtained. 0 ;

[0047] Stereo matching specifically includes the following steps:

[0048] Step 1: during stereo matching, the search range 5 is determined by epipolar constraints, and each coordinate point along the epipolar line in the speckle pattern captured by the right camera is taken as the center. Around the center, the area with the same size as the reference sub-area 6 of the left camera is selected in turn as the target sub-area for stereo matching (a rectangular area with a size of 20×15 pixels), and the correlation between each target sub-area and the reference sub-area 6 is calculated by the zero-mean normalized cross-correlation coefficient, and the target sub-area with the highest correlation is selected as the matching sub-area 8 of the reference sub-area;

[0049] After the epipolar constraint, each pair of matching pixels is on the same horizontal line with the same row coordinates and only a deviation in the column coordinates, which is used for subsequent calculation of the disparity. The disparity d refers to the pixel column coordinates of the same point of the object in the left view minus the column coordinates of the corresponding pixel in the right view, combined with the two-dimensional coordinates of the pixel itself to form a three-dimensional coordinate.

[0050] For each pixel to be matched in the reference sub-region 6, the matching cost between it and the possible matching pixels in the target sub-region is preliminarily calculated through the grayscale and position information of the neighborhood to form a cost matrix, which is used to characterize the matching degree of two pixels. The smaller the cost, the greater the correlation between the two pixels, and the higher the possibility that they represent the same physical point in the actual space;

[0051] Step 2: Iterate through the inverse Gauss-Newton iteration method (IC-GN) to obtain the sub-pixel coordinates of each pixel in the matching sub-area;

[0052] Step 3: Perform three-dimensional reconstruction based on the coordinates of each pixel point in the reference sub-area 6 and the coordinates of the pixel points in the corresponding matching sub-area to obtain the three-dimensional coordinates of each pair of pixel points in the reference sub-area 6 of the roller speckle pattern, which are recorded as a three-dimensional coordinate point cloud;

[0053] S3, performing time sequence matching between the current frame of the same roller speckle pattern photographed by the left camera and its next frame, and determining a target sub-region that matches the reference sub-region 6 on the next frame of speckle pattern photographed by the left camera;

[0054] During timing matching, the speckle pattern of the next frame is searched at alternate points within the search range 5 of the reference sub-area of ​​the current frame, and the target sub-area with the highest correlation with the reference sub-area is selected as the new reference sub-area 7. The search range is updated according to the position of the new reference sub-area 7, which is recorded as the new search range 10;

[0055] S4, after taking the target sub-region of the speckle pattern of the same roller in the next frame photographed by the left camera as the new reference sub-region 7, stereo matching is performed on the speckle patterns of the same roller in the next frame photographed by the left camera and the right camera (the same as the stereo matching step in S2), and a new matching sub-region 9 on the speckle pattern of the next frame photographed by the right camera that matches the new reference sub-region is determined;

[0056] Based on the next frame new reference sub-region and the new matching sub-region 9 of the same roller speckle pattern photographed by the left camera and the right camera, the three-dimensional coordinate point cloud P of the next frame new reference sub-region 7 of the speckle pattern is obtained. 1 ;

[0057] S5, repeat S3-S4 to obtain the three-dimensional coordinate point cloud P of the reference sub-area of ​​each frame of the speckle pattern of the same roller 2 , P 3 , P 4 ,…P i ;

[0058] S6. Obtain the pose matrix between the three-dimensional coordinate point clouds of adjacent frames through the point cloud registration algorithm, and obtain the rotation matrix R of the pose matrix through singular value decomposition to obtain the inter-frame rotation angle α of the corresponding roller adjacent frames i ;

[0059] The rotation matrix R is expressed as:

[0060] q m =Rp m +t

[0061] Where: p is the coordinate of the feature point (pixel in the sub-region) of the current frame, p = (p 1 ,p 2 ,…,p m );q is the feature point coordinate of the next frame, q=(q 1 ,q 2 ,…,q m ), m is the feature point number, m = 1, 2, ..., n; R contains three rotation displacements; t is the translation vector describing the movement before and after, including three translation displacements.

[0062] Construct an error function and solve the optimal transformation of the rotation matrix R through the least squares equation so that the mean square error between the corresponding point sets (the set of all pixel points in the sub-area) is minimized;

[0063]

[0064] The rotation matrix R is regarded as a composite matrix of three basic rotation matrices Rx, Ry, and Rz that rotate around the rotation axis. The rotation angle α is solved by SVD matrix decomposition. The inter-frame rotation angle α i It is expressed as:

[0065]

[0066] S7, according to the inner ring driving speed, the frame rotation angle α i The instantaneous sliding-rolling ratio of the corresponding roller is calculated by the camera frame rate;

[0067] S7.1. Calculate the inner circle linear velocity v1 :

[0068]

[0069] S7.2. Calculate the instantaneous linear velocity v of the roller 2 :

[0070]

[0071] S7.3. Calculate the instantaneous slide-roll ratio ξ i :

[0072]

[0073] Where: n is the inner ring driving speed; R 1 is the inner circle radius; R 2 is the roller radius; f is the camera frame rate.

[0074] S8. Repeat steps S2-S7 to obtain the instantaneous slide-to-roll ratio of each roller.

[0075] The parts not described in detail in the above embodiments are prior art.

[0076] It should be noted that although the present invention is described by the above embodiments, the present invention may also have other various embodiments. Without departing from the spirit and scope of the present invention, it is obvious that those skilled in the art may make various corresponding changes and deformations to the present invention, but these changes and deformations should all fall within the scope of protection of the appended claims of the present invention and their equivalents.

Claims

1. A method for measuring the sliding-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation, characterized in that: The method is based on a visual platform that can carry the bearing to be tested and drive the inner ring of the bearing to be tested to rotate, and a binocular camera consisting of a left camera and a right camera with the same frame rate; The following steps are involved: S1. Place the bearing to be tested on a visual platform, and make different speckle patterns on the end face of each roller on the same side of the bearing to be tested. Drive the inner ring of the bearing to be tested to rotate at a preset inner ring driving speed to put it into operation. Use a binocular camera to synchronously capture the end face image of the bearing to be tested frame by frame in the operation state. S2, selecting a reference sub-region in the speckle pattern of a roller photographed by the left camera, taking the initial frame of the same roller speckle pattern photographed by the left camera and the right camera as the current frame for stereo matching, and determining a matching sub-region on the speckle pattern photographed by the right camera that matches the reference sub-region; Based on the reference sub-region and the matching sub-region of the initial frame of the same roller speckle pattern photographed by the left camera and the right camera, a three-dimensional coordinate point cloud P0 of the reference sub-region of the initial frame speckle pattern is obtained; S3, performing time sequence matching between the current frame of the same roller speckle pattern photographed by the left camera and the next frame thereof, and determining a target sub-region on the next frame of speckle pattern photographed by the left camera that matches the reference sub-region; S4, taking the target sub-region of the speckle pattern of the same roller in the next frame photographed by the left camera as a new reference sub-region, performing stereo matching on the speckle patterns of the same roller in the next frame photographed by the left camera and the right camera, and determining a new matching sub-region on the speckle pattern of the next frame photographed by the right camera that matches the new reference sub-region; Based on the new reference sub-region and the new matching sub-region of the next frame of the same roller speckle pattern photographed by the left camera and the right camera, a three-dimensional coordinate point cloud P1 of the new reference sub-region of the next frame of the speckle pattern is obtained; S5, repeat S3-S4 to obtain the three-dimensional coordinate point cloud P2, P3, P4, ... P of the reference sub-area of ​​each frame of the speckle pattern of the same roller i ; S6. Obtain the pose matrix between the three-dimensional coordinate point clouds of adjacent frames through the point cloud registration algorithm, and obtain the rotation matrix R of the pose matrix through singular value decomposition to obtain the inter-frame rotation angle α of the corresponding roller adjacent frames i ; S7, according to the inner ring driving speed, the frame rotation angle α i The instantaneous slide-to-roll ratio of the corresponding roller is calculated using the camera frame rate.

2. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 1, characterized in that: During stereo matching, the search range is determined by epipolar constraints. Each coordinate point along the epipolar line in the speckle pattern captured by the right camera is taken as the center, and areas with the same size as the reference sub-area of ​​the left camera are selected around the center in sequence as target sub-areas for stereo matching. The correlation between each target sub-area and the reference sub-area is calculated by the zero-mean normalized mutual correlation coefficient, and the target sub-area with the highest correlation is selected as the matching sub-area of ​​the reference sub-area.

3. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 2, characterized in that: During stereo matching, the reverse Gauss-Newton iteration method is used to iterate and obtain the sub-pixel coordinates of each pixel in the matching sub-area.

4. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 3, characterized in that: During stereo matching, three-dimensional reconstruction is performed based on the coordinates of each pixel point in the reference sub-area and the coordinates of the pixel points in the corresponding matching sub-area to obtain the three-dimensional coordinates of each pair of pixel points in the reference sub-area of ​​the roller speckle pattern.

5. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 1, characterized in that: The rotation matrix R is expressed as: q m =Rp m +t Where: p is the feature point coordinate of the current frame, p=(p1,p2,…,p m );q is the feature point coordinate of the next frame, q=(q1,q2,…,q m ), m is the feature point number, m = 1, 2, ..., n; R contains three rotation displacements; t is the translation vector describing the movement before and after, including three translation displacements.

6. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 5, characterized in that: Construct an error function and solve the optimal transformation of the rotation matrix R through the least squares equation so that the mean square error between corresponding point sets is minimized.

7. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 5, characterized in that: Inter-frame rotation angle α i It is expressed as: Where: The rotation matrix R is regarded as a composite matrix of three basic rotation matrices Rx, Ry, and Rz that rotate around the rotation axis.

8. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 7, characterized in that: Instantaneous slide-roll ratio ξ i The calculation formula is as follows: Where: n is the inner ring driving speed; R1 is the inner ring radius; R2 is the roller radius; f is the camera frame rate.

9. The method for measuring the sliding-to-rolling ratio of cylindrical rollers based on three-dimensional digital image correlation according to claim 1, characterized in that: Repeat steps S2-S7 to obtain the instantaneous slide-to-roll ratio of each roller.

Images