A Quantitative Judgment Method for the Distortion Variable of Circular Wear Scar Images

Abstract

A quantitative determination method for the distortion amount of a circular wear scar image provided by the present invention includes the following steps: Step 1, segment the obtained four-ball friction wear scar image to obtain a wear scar segmentation map; Step 2, calculate the position size of the wear scar segmentation map; Step 3, construct a wear scar circle based on the position size obtained in Step 1; Step 4, calculate the minimum abnormality rate of the wear scar circle; Step 5, obtain a boundary pixel map according to the wear scar segmentation map; Step 6, calculate the wear scar roundness ratio of the boundary pixel map; Step 7, calculate the wear scar distortion amount of the four-ball friction wear scar image based on the minimum abnormality rate of the wear scar circle and the wear scar roundness ratio of the boundary pixel map; The present invention establishes an evaluation index for the distortion degree of the wear scar based on multiple image feature parameters, realizes the determination of the shape distortion amount of the wear scar, and changes the subjectivity and unscientific nature of the qualitative determination by the tester.

Description

Technical Field

[0001] The present invention belongs to the technical field of image data analysis and processing, and particularly relates to a method for quantitatively determining the distortion amount of a circular wear scar image. Background Art

[0002] The research results of the consulting project of the Chinese Academy of Engineering, "Research on the Current Situation and Development Strategy of Tribology Science and Engineering Applications", show that in 2006, the losses caused by friction and wear in China reached about 950 billion yuan, accounting for 4.5% of the gross domestic product (GDP) of that year. In addition, if wear intensifies, it will cause component failure and machine breakdown, and even lead to catastrophic consequences. A lubricant with good performance is a lubricating medium used to reduce the frictional resistance of the friction pair and slow down its wear, and can play roles such as cooling, cleaning, and preventing pollution to the friction pair. Therefore, it is particularly important to timely and accurately measure the performance of the lubricant for protecting machinery and reducing energy consumption. The four-ball friction and wear tester is widely used in the friction coefficient measurement test of lubricating oil due to its characteristics such as convenient operation, simple structure, short test cycle, small oil consumption, and low cost.

[0003] The petroleum and chemical industry standards of China (GB-T 12583-1998 and H-T 0762-2005) stipulate the test process of the friction coefficient of lubricating oil and clarify the observation method of the wear scar morphology characteristics: the wear scar shape is generally circular or oval. When the wear scar deformation is severe, it cannot be used in the test process of the friction coefficient of lubricating oil, and a new running-in test needs to be carried out. Therefore, quickly and accurately determining the shape characteristics of the wear scar is very important for the determination of the wear scar morphology and test effectiveness. However, at present, there is no research on the automatic detection of the wear scar deformation in the four-ball friction test, and it is still qualitatively judged by the tester based on experience, which inevitably produces subjective judgment errors, is neither scientific nor objective, and has little practical guiding significance. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for quantitatively determining the distortion amount of a circular wear scar image, which solves the above-mentioned deficiencies in the prior art.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for quantitatively determining the distortion amount of a circular wear scar image provided by the present invention includes the following steps:

[0007] Step 1, segment the obtained four-ball friction wear scar image to obtain a wear scar segmentation map;

[0008] Step 2, calculate the position dimensions of the wear scar segmentation map;

[0009] Step 3, construct a wear scar circle based on the position dimensions obtained in Step 1;

[0010] Step 4, calculate the minimum abnormality rate of the wear scar circle;

[0011] Step 5, obtain the boundary pixel map according to the wear scar segmentation map;

[0012] Step 6, calculate the wear scar roundness ratio of the boundary pixel map;

[0013] Step 7, calculate the wear scar distortion amount of the four-ball friction wear scar image according to the minimum abnormality rate of the wear scar circle and the wear scar roundness ratio of the boundary pixel map.

[0014] Preferably, in step 2, the position dimensions of the wear scar segmentation map include the centroid and the area. Among them, the centroid and the area of the wear scar segmentation map are calculated respectively by the following formulas:

[0015]

[0016]

[0017] S = num(R) (3)

[0018] where (x, y) is the coordinate of any pixel in the wear scar segmentation map, x and y are the row number and the column number respectively, 1 ≤ x ≤ M and 1 ≤ y ≤ N; M and N are the maximum row number and the maximum column number of the wear scar segmentation map respectively; R(x, y) represents the value of the pixel (x, y) of the wear scar segmentation map R; num(R) is the pixel statistical operator of the wear scar area in the wear scar segmentation map R.

[0019] Preferably, in step 3, the wear scar circle is constructed according to the position dimensions obtained in step 1. The specific method is:

[0020] Take the centroid of the wear scar segmentation map as the center of the circle;

[0021] Calculate the radius according to the area of the wear scar segmentation map;

[0022] Obtain the wear scar circle according to the center of the circle and the radius;

[0023] Determine the pixels of the wear scar circle according to the following formula:

[0024]

[0025] where Q is the wear scar circle, which is a binary image. Q(x, y) = 1 indicates that the pixel is marked as a wear scar in the wear scar circle; Q(x, y) = 0 indicates that the pixel is marked as a non-wear scar in the wear scar circle; r is the radius of the wear scar circle, which is: x C is the row number of the center O of the wear scar circle; y CThe column number of the center O of the wear scar circle; (x, y) is the coordinate of any pixel in the wear scar segmentation map, where x and y are the row number and column number respectively, and 1 ≤ x ≤ M and 1 ≤ y ≤ N.

[0026] Preferably, in step 4, calculate the minimum abnormality rate of the wear scar circle. The specific method is as follows:

[0027] Detect the abnormal pixels of the wear scar segmentation map according to the obtained wear scar circle to obtain a marked map;

[0028] Calculate the wear scar abnormality rate of the wear scar segmentation map according to the obtained marked map;

[0029] Obtain the minimum abnormality rate according to the obtained wear scar abnormality rate.

[0030] Preferably, obtain the marked map according to the following formula. The specific method is as follows:

[0031]

[0032] where E O is the marked map, and the pixel with E O (x, y) = 1 is a normal pixel; the pixel with E O (x, y) = 0 is an abnormal pixel or background; Q(x, y) = 1 indicates that the pixel is marked as a wear scar in the wear scar circle; R(x, y) = 1 indicates that the pixel is marked as a wear scar in the wear scar segmentation map.

[0033] Preferably, in step 5, calculate the abnormality rate of the wear scar circle. The specific method is as follows:

[0034] Construct a virtual circle according to the marked map;

[0035] Calculate the abnormality rate of the virtual circle according to the following formula:

[0036]

[0037] Preferably, in step 5, obtain the boundary pixel map according to the following formula in combination with the wear scar segmentation map:

[0038]

[0039] where A is the boundary pixel map, a binary map; A(x, y) = 1 indicates that the pixel is a boundary pixel, A(x, y) = 0 indicates that the pixel is a non-boundary pixel, 1 ≤ x ≤ M, 1 ≤ y ≤ N; k and l respectively represent the index variables of the row and column.

[0040] Preferably, in step 6, calculate the circularity ratio of the wear scar of the boundary pixel map. The specific method is as follows:

[0041] Calculate the minimum radius and the maximum radius of the boundary pixel map;

[0042] Calculate the roundness ratio of the wear scar according to the obtained minimum radius and maximum radius.

[0043] Preferably, in step 7, according to the minimum abnormality rate of the wear scar circle and the roundness ratio of the wear scar in the boundary pixel map, the distortion amount of the four-ball friction wear scar image is calculated by combining the following formula:

[0044] H = α·y + β·G *

[0045] where α and β are the roundness weight and the distortion weight respectively; H is the distortion amount of the wear scar; y is the roundness ratio of the wear scar; G * is the minimum abnormality rate.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] The present invention provides a method for quantitatively determining the distortion amount of a circular wear scar image. The wear scar image is collected by using an image acquisition device or a scanning electron microscope (both having a magnification function), and the automatic segmentation of the wear scar area is realized by using image analysis and processing technology, which improves the speed and accuracy of segmentation, and at the same time greatly reduces the workload of manual segmentation; on the basis of constructing a standard circle, multiple image feature parameters describing the shape characteristics of the wear scar are established, realizing the quantitative expression of the wear scar shape; an evaluation index for the distortion degree of the wear scar is established based on multiple image feature parameters, realizing the determination of the distortion amount of the wear scar shape, and changing the subjectivity and unscientific nature of the qualitative determination by the tester. This method has the advantages of simplicity, high efficiency, and no need for manual assistance throughout the process, and is very suitable for developing into an automatic analysis and processing software for four-ball friction test data. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is the wear scar image f;

[0049] Figure 2 is the wear scar segmentation map R;

[0050] Figure 3 is the marked map E where (x O , y O ) is (365.2125, 465.3137); O ;

[0051] Figure 4 is the marked map E where (x O , y O ) is (465.2125, 485.3137); O ;

[0052] Figure 5 is the illustration of the boundary pixel set A. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The present invention will be further described in detail below with reference to the accompanying drawings.

[0054] As Figures 1 to 5 shown, a quantitative determination method for the distortion amount of a circular wear scar image provided by the present invention specifically comprises the following implementation steps:

[0055] Step S0: Acquisition of the wear scar image. After a four-ball friction test, a wear scar will be formed on the surface of the steel ball. The surface wear scar (≤1 mm) is not visible to the naked eye, and a special acquisition device such as a high-power microscope or a scanning electron microscope is required to acquire the wear scar image, as Figure 1 shown.

[0056] Step S1: Segmentation of the wear scar image. The segmentation of the wear scar image includes manual segmentation and automatic segmentation, and its purpose is to distinguish the wear scar from the background.

[0057] Compared with the manual segmentation method, the automatic segmentation method has the advantages of high segmentation accuracy and high speed; and the automatic segmentation method is to perform an initial segmentation of the wear scar by establishing a double-extreme value filtering difference metric operator, and adopt double constraints of gray level and distance to subdivide the boundary of the wear scar region precisely, including steps such as graying, denoising, initial segmentation of the wear scar region, and double-constrained boundary fine segmentation. The algorithm has a good segmentation effect on the boundary and a fast running speed.

[0058] In this embodiment, the automatic segmentation method is adopted, and the obtained wear scar segmentation map R is as Figure 2 shown. The white and black regions in the figure are the wear scar (pixel value is 1) and the background (pixel value is 0), respectively.

[0059] Step S2: Extraction of the position and size parameters of the wear scar. The position and size parameters of the wear scar mainly include: the position of the centroid C and the area S. Among them, the position of the centroid C of the wear scar is represented by the two-dimensional coordinates of the pixel coordinate system as (x C , y C ), where x C and y C are the row number and column number respectively; the area refers to the number of pixels in the wear scar region in the wear scar segmentation map, and is represented by S:

[0060]

[0061]

[0062] S = num(R) (3)

[0063] Where, (x, y) are the coordinates of any pixel in the wear spot segmentation map, x and y are the row number and column number respectively, 1 ≤ x ≤ M and 1 ≤ y ≤ N; M and N are the maximum row number and maximum column number of the wear spot segmentation map respectively; R(x, y) represents the value of the pixel (x, y) in the wear spot segmentation map R. R is a binary image, R(x, y) = 1 indicates that the pixel is marked as a wear spot in the wear spot segmentation map, and R(x, y) = 0 indicates that the pixel is marked as the background in the wear spot segmentation map; num(R) is the pixel statistical operator for the wear spot area in the wear spot segmentation map R.

[0064] In this embodiment, the position dimension parameters of the wear spot are: x C = 365.2125, y C = 465.3137, S = 396979 and r = 355.48.

[0065] Step S3: Construct a wear spot circle based on the wear spot.

[0066] Ideally, the shape of the wear spot should be circular. A circle is constructed based on the position dimension parameters of the wear spot for subsequent identification of abnormal pixels. The pixel values inside the circle are all set to 1, and the remaining pixels are 0, obtaining the wear spot circle, as shown in Equation (4):

[0067]

[0068] Where, Q is the wear spot circle, which is a binary image. Q(x, y) = 1 indicates that the pixel is marked as a wear spot in the wear spot circle; Q(x, y) = 0 indicates that the pixel is marked as a non-wear spot in the wear spot circle; r is the radius of the wear spot circle, which is: x O is the row number of the center O of the wear spot circle; y O is the column number of the center O of the wear spot circle; The center O varies within a square area with the centroid C of the wear spot as the center and side length r, that is, it satisfies: and [] is the rounding operation.

[0069] Step S4: Detection and marking of abnormal pixels. To further determine the degree of variation in the shape of the wear spot, it is necessary to compare whether the markings of the pixels in the wear spot circle and the wear spot segmentation map are consistent. When a certain pixel is marked as a wear spot in both the wear spot circle and the wear spot segmentation map, the pixel is considered a normal pixel, otherwise it is considered an abnormal pixel, and finally the marking map E is obtained O .

[0070] When the position of the center O changes, the position of the wear spot circle changes, and the marked normal pixels also change; The calculation formula of the marking map E O is shown as (5):

[0071]

[0072] Among them, E O is a marked graph, and for E O the pixel with E(x,y)=1 is a normal pixel; E O the pixel with E(x,y)=0 is an abnormal pixel or background, E O which will change with the change of the center O of the circle; R(x,y)=1 indicates that the pixel is marked as a wear spot in the wear spot segmentation graph.

[0073] In this embodiment, when (x O , y O ) takes the value of (365.2125, 465.3137), the marked graph E O is as Figure 3 shown; when (x O , y O ) takes the value of (465.2125, 485.3137), the marked graph E O is as Figure 4 shown.

[0074] Step S5: Abnormality rate of the wear spot. The more abnormal pixels there are in the wear spot area of the wear spot segmentation graph, the higher the abnormality rate of the wear spot, and vice versa. Since the area of the wear spot and the wear spot circle in the wear spot segmentation graph is the same, the abnormality rate of the wear spot is related to the number of abnormal pixels and can be expressed by Equation (6):

[0075]

[0076] Among them, G O represents the abnormality rate of the wear spot circle, and its value changes with the change of the center O of the wear spot circle.

[0077] In this embodiment, the change of the abnormality rate G O of the wear spot with the change of the center O is shown in Table 1.

[0078] Table 1 (Partial) change values of the abnormality rate of the wear spot with the center O of the wear spot

[0079]

[0080]

[0081] Step S6: Calculation of the minimum abnormality rate. Assume that when the change of the center O is , the value of G O is the smallest, and at this time the abnormality rate of the wear spot is the smallest, and the obtained wear spot circle is called the optimal wear spot circle.

[0082] Based on the optimal wear spot circle, the smallest abnormality rate is an important index for quantitatively evaluating whether the shape of the wear spot is distorted:

[0083] G* = min(G O ) (7)

[0084]

[0085] where G * is the minimum anomaly rate; is the center of the best wear scar circle.

[0086] In this embodiment, G * = 3.58%, and

[0087] Step S7: Extract the wear scar boundary pixel set. In addition to the anomaly rate, the distortion of the actual wear scar is also reflected in the distance between the abnormal pixels and the center of the circle, which is also an index reflecting the wear scar distortion. Extracting the wear scar boundary pixels is the basis for calculating the roundness of the wear scar. By calculating the distance from the wear scar boundary pixels to the center of the circle, the roundness of the wear scar can be quickly determined.

[0088] The wear scar boundary pixels are defined as at least one pixel in the eight-neighborhood of the pixel on the wear scar segmentation map is non-wear scar, and the boundary pixels are represented by A:

[0089]

[0090] where A is the boundary pixel map, a binary map; A(x, y) = 1 indicates that the pixel is a boundary pixel, A(x, y) = 0 indicates that the pixel is a non-boundary pixel, 1 ≤ x ≤ M, 1 ≤ y ≤ N; k and l respectively represent the index variables of the row and column, both are integers, and the values are -1, 0, 1 respectively.

[0091] In this embodiment, the boundary pixel map A is Figure 5 the white pixels in

[0092] Step S8: Calculate the minimum radius and the maximum radius of the boundary pixel map.

[0093] The minimum radius and the maximum radius of the wear scar can be used to calculate the roundness of the wear scar. The smaller the difference between the two, the closer the shape of the wear scar is to a circle, and vice versa. Therefore, the minimum radius and the maximum radius of the boundary pixel map are calculated by the following formula:

[0094]

[0095]

[0096] where R1 is the minimum radius; R2 is the maximum radius; is the center of the best wear scar circle.

[0097] In this embodiment, R1 = 330.4526 and R2 = 409.3486.

[0098] Step S9: Roundness ratio of the wear scar. Taking the radius r of the wear scar as the comparison benchmark, the roundness ratio of the wear scar is used to measure the change rate of the outer boundary of the wear scar. The smaller the value, the closer the shape of the wear scar is to a circle. The calculation formula for the roundness ratio y of the wear scar is shown in (12):

[0099]

[0100] In this embodiment, the roundness ratio y of the wear scar = 22.19%.

[0101] Step S10: Distortion amount of the wear scar. A quantitative index - the distortion amount of the wear scar is established to measure the degree of distortion of the wear scar. When the abnormality rate or roundness ratio of the wear scar takes a large value, the distortion amount of the wear scar is also large, and vice versa.

[0102] H = α·y + β·G * (13)

[0103] Wherein, α and β are the roundness weight and distortion weight respectively, satisfying: 0 ≤ α ≤ 1, 0 ≤ β ≤ 1, α + β = 1; H is the distortion amount of the wear scar.

[0104] In this embodiment, the roundness weight α and distortion weight β are taken as 0.2 and 0.8, and the distortion amount of the wear scar obtained is: H = 7.30%.

Claims

1. A quantitative determination method for the distortion amount of a circular wear scar image, characterized in that, It includes the following steps: Step 1: Segment the obtained four-ball friction wear scar image to obtain a wear scar segmentation map; Step 2: Calculate the position dimensions of the wear scar segmentation map; Step 3: Construct a wear scar circle based on the position dimensions obtained in Step 2; Step 4: Calculate the minimum abnormality rate of the wear scar circle; Step 5: Obtain a boundary pixel map based on the wear scar segmentation map; Step 6: Calculate the wear scar roundness ratio of the boundary pixel map; Step 7: Calculate the wear scar distortion amount of the four-ball friction wear scar image based on the minimum abnormality rate of the wear scar circle and the wear scar roundness ratio of the boundary pixel map; Step 4: Calculate the minimum abnormality rate of the wear scar circle. The specific method is as follows: Detect the abnormal pixels of the wear scar segmentation map based on the obtained wear scar circle to obtain a marked map; Calculate the wear scar abnormality rate of the wear scar segmentation map based on the obtained marked map; Obtain the minimum abnormality rate based on the obtained wear scar abnormality rate; Obtain the marked map according to the following formula. The specific method is as follows: (5) Among them, is a marker map, the pixels of which are normal pixels; the pixels of which are abnormal pixels or background; indicates that the pixel is marked as a wear spot in the wear spot circle; indicates that the pixel is marked as a wear spot in the wear spot segmentation map; Calculate the wear scar abnormality rate according to the following formula: ； where S is the area of the wear scar segmentation map; In Step 6, calculate the wear scar roundness ratio of the boundary pixel map. The specific method is as follows: Calculate the minimum radius and the maximum radius of the boundary pixel map; Calculate the wear scar roundness ratio based on the obtained minimum radius and maximum radius; In Step 7, based on the minimum abnormality rate of the wear scar circle and the wear scar roundness ratio of the boundary pixel map, calculate the wear scar distortion amount of the four-ball friction wear scar image by combining the following formula: Among them, and are the roundness weight and the distortion weight respectively; is the distortion amount of the wear scar; is the roundness ratio of the wear scar; is the minimum abnormality rate.

2. The quantitative determination method for the distortion amount of a circular wear scar image according to claim 1, characterized in that In Step 2, the position dimensions of the wear scar segmentation map include the centroid and the area. Among them, the centroid and the area of the wear scar segmentation map are calculated respectively by the following formulas: (1) (2) (3) Wherein, is the coordinate of any pixel in the wear scar segmentation map, and are the row number and column number respectively, and ; and are the maximum row number and maximum column number of the wear scar segmentation map respectively; represents the pixel of the wear scar segmentation map value; is the pixel statistical operator of the wear scar area in the wear scar segmentation map ; and are the row number and column number of the two-dimensional coordinates of the centroid of the wear scar segmentation map respectively; is the area of the wear scar segmentation map.

3. A quantitative determination method for the distortion amount of a circular wear scar image according to claim 1, characterized in that, In Step 3, construct a wear scar circle based on the position dimensions obtained in Step 2. The specific method is as follows: Take the centroid of the wear scar segmentation map as the center of the circle; Calculate the radius according to the area of the wear scar segmentation map; Obtain the wear scar circle based on the center of the circle and the radius; Determine the pixels of the wear scar circle according to the following formula: Among them, is the wear scar circle, which is a binary image, indicating that the pixel is marked as a wear scar in the wear scar circle; indicating that the pixel is marked as non-wear scar in the wear scar circle; is the radius of the wear scar circle, which is: ; is the center of the wear scar circle row number; is the center of the wear scar circle column number; is the coordinate of any pixel in the wear scar segmentation map, and are the row number and column number respectively, and ; is the area of the wear scar segmentation map.

4. A quantitative determination method for the distortion amount of a circular wear scar image according to claim 1, characterized in that In Step 5, obtain the boundary pixel map by combining the wear scar segmentation map according to the following formula: Among them, is a boundary pixel map, a binary map; indicates that the pixel is a boundary pixel, indicates that the pixel is a non-boundary pixel, ; and respectively represent the index variables of rows and columns.

Images