CN107358604B - Four-ball friction test speckle image anomaly detection method - Google Patents

Four-ball friction test speckle image anomaly detection method Download PDF

Info

Publication number
CN107358604B
CN107358604B CN201710597206.0A CN201710597206A CN107358604B CN 107358604 B CN107358604 B CN 107358604B CN 201710597206 A CN201710597206 A CN 201710597206A CN 107358604 B CN107358604 B CN 107358604B
Authority
CN
China
Prior art keywords
grinding
area
image
spot
radius
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN201710597206.0A
Other languages
Chinese (zh)
Other versions
CN107358604A (en
Inventor
肖梅
张雷
赵国玉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changan University
Original Assignee
Changan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Changan University filed Critical Changan University
Priority to CN201710597206.0A priority Critical patent/CN107358604B/en
Publication of CN107358604A publication Critical patent/CN107358604A/en
Application granted granted Critical
Publication of CN107358604B publication Critical patent/CN107358604B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/155—Segmentation; Edge detection involving morphological operators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30164—Workpiece; Machine component

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Geometry (AREA)
  • Finish Polishing, Edge Sharpening, And Grinding By Specific Grinding Devices (AREA)
  • Constituent Portions Of Griding Lathes, Driving, Sensing And Control (AREA)

Abstract

The invention provides a four-ball friction test grinding mark image abnormity detection method, which comprises the steps of firstly, taking an area block as a processing object, taking a central area as a reference area, finding out a comparison area with the same size and shape as the reference area in the central horizontal direction, then calculating the grinding mark difference degree of the comparison area and the reference area at the same direction angle, calculating the direction angle of a grinding mark shaft through the grinding mark difference degree, obtaining a grinding mark alignment image according to the direction angle of the grinding mark, and further obtaining a grinding mark seed image and a grinding mark area image; calculating the equivalent radius and the direction radius of the grinding spot area, and calculating the variance and the range of the direction radius according to the direction radius; and calculating the axial difference and the centrifugal degree of the grinding spot area image, and judging the abnormality of the grinding spot image according to any one parameter of the variance of the direction radius, the extreme difference of the direction radius, the axial difference or the centrifugal degree. The invention automatically analyzes the shape and appearance of the wear scar image, finds the abnormal wear scar image and is convenient for the testers to analyze the reason of the abnormality.

Description

Four-ball friction test speckle image anomaly detection method
Technical Field
The invention belongs to the field of automobile running materials, and particularly relates to an anomaly detection method for a four-ball friction test speckle image.
Background
A method for measuring the antiwear performance of a lubricant (SH/T0189-92) is provided, when a tester operates in an irregular mode, the antiwear performance of the lubricant is poor or the lubricant does not completely invade a steel ball and other factors, abnormal abrasion of the steel ball is caused, the collected abrasion mark image needs to be screened out, the abnormal abrasion mark image is detected in time, the tester is reminded to test again or replace the lubricant in time, so that normal abrasion mark image data are ensured to be obtained, and the method has important practical significance for accurately measuring the antiwear performance of the lubricant and reducing the abrasion wear of parts. Based on this, a new anomaly detection method for the four-ball friction test grinding spot image is urgently needed to be provided. The national standard SH/T0189-92 gives a method for judging the misalignment of the shaft center, and provides that if the deviation of the average value of two measurements of a bottom ball and the average value of six measurements is more than 0.04mm, the shaft center alignment of the upper ball and the oil cup is checked. In addition, no literature or data related to the abnormal detection of the wear-mark image is available.
Disclosure of Invention
The invention aims to provide a four-ball friction test grinding spot image anomaly detection method, which solves the problem that the existing steel ball grinding spot image anomaly is visually observed by a tester, and the tester perception brings errors, so that the detection is inaccurate.
In order to achieve the purpose, the invention adopts the technical scheme that:
the invention provides a four-ball friction test grinding spot image anomaly detection method, which comprises the following steps:
firstly, collecting a steel ball to be measured and a white scale by using a scanning electron microscope to obtain a wear scar image F; carrying out gray processing on the grinding spot image F by a weighted average method to obtain a gray grinding spot image F;
secondly, calculating the actual length h of a unit pixel in the grinding spot image;
thirdly, clockwise rotating the gray scale abrasion pattern f by theta degrees around the central point O of the image to obtain a rotary abrasion pattern fθ；
The fourth step, rotating the abrasion pattern fθA certain area is selected as a reference area A by taking the central point O as the centerθ(ii) a In the central horizontal direction, with reference to the area AθAs the center, symmetrically selecting and referencing areas A to both sidesθS area blocks having the same size and shape as the comparison area DθIn total, 2s contrast regions D are obtainedθ；
Fifthly, comparing the areas D according to the same direction angle thetaθAnd a reference area AθDegree of difference C of grinding marksθCalculating the axial angle of the grinding mark
Figure GDA0001425146420000021
Sixthly, rotating the gray scale abrasion pattern f obtained in the first step counterclockwise
Figure GDA0001425146420000022
And (4) obtaining a grinding spot correction diagram
Figure GDA0001425146420000023
Seventh, on the above-mentioned grinding spot correcting map
Figure GDA0001425146420000024
Extracting seed pixels of the grinding crack area to obtain a grinding crack seed image Sa;
eighthly, calculating the gray mean value of all the grinding mark seed point pixels in the grinding mark seed map Sa, and taking the gray mean value as an adaptive threshold value to correct the grinding marks
Figure GDA0001425146420000025
Dividing to obtain an initial pattern H of the wear scar area;
ninth, performing morphological expansion operation on the grinding crack seed graph Sa to obtain a grinding crack seed expansion graph Mg；
The tenth step is to carry out the initial graph H of the grinding spot area and the expansion graph M of the grinding trace seedsgPerforming difference operation to obtain an abnormal speckle pattern Ml；
The tenth step is that an initial pattern H and an abnormal pattern M of the grinding spots are comparedlPerforming difference processing to obtain a speckle difference image Mp；
The twelfth step, the speckle difference map M is comparedpPerforming morphological close operation to obtain a mottled area map Mr；
The tenth step, according to the abrasion spot area diagram MrThe area of the grinding spot area, and calculating a grinding spot area diagram MrEquivalent radius r ofdDirection radius rrAnd centroid OrCoordinate (x) ofc,yc) And converting the equivalent radius r according to the unit pixel length hdThe true distance of (d);
fourteenth, according to the pattern M of the wear-scar regionrRadius in the direction of (r)rCalculating the variance S of the radius of the directionrAnd extreme difference J of direction radiusrThen, the variance S of the direction radius is calculated according to the unit pixel length hrRadial deviation from the sum direction JrThe true distance of (d);
the fifteenth step, according to the abrasion spot area diagram MrLong axis l ofaMinor axis lbAnd axis OzCalculating the pattern M of the mottled arearAxial difference lcAnd degree of centrifugation ldThen, the axial difference l is converted according to the unit pixel length hcAnd degree of centrifugation ldThe true distance of (c).
Sixteenth, according to the variance S of the direction radiusrDirection radius extremely different JrAxial difference lcOr degree of centrifugation ldAnd judging the abnormality of the grinding spot image.
Preferably, in the fourth step, the reference region aθIs a square of 2 omega +1 pixels on a side, the area of which
Figure GDA0001425146420000026
Is of a size of
Figure GDA00014251464200000311
Wherein the value range of omega is 3-20; in the selected contrast area DθWhen the positions of two adjacent area blocks are partially overlapped or not overlapped.
Preferably, in the fifth step, the axial angle of the grinding mark
Figure GDA0001425146420000031
The calculation process of (2) is shown in formula (1):
Figure GDA0001425146420000032
wherein i and j are integers, which are the row number and the column number of the pixel coordinates of the comparison area and the reference area respectively,
Figure GDA0001425146420000033
alpha and beta are respectively k-th contrast area
Figure GDA0001425146420000034
The amount of fine shift in the row direction and the column direction of (a) is taken as α ═ 1,0,1 and β ═ 1,0, 1; a. theθ(i, j) is a reference region A of the direction angle thetaθThe gray value of the pixel point (i, j) in (1);
Figure GDA0001425146420000035
the k-th contrast region at the direction angle theta
Figure GDA00014251464200000310
The gray value of the pixel point (i + α, j + β + k (2 ω +1)) in (1).
Preferably, in the seventh step, the scrub rectification chart is extracted according to the formula (2)
Figure GDA0001425146420000036
And obtaining a grinding crack seed graph Sa by using the seed pixels of the middle grinding crack area, specifically:
Figure GDA0001425146420000037
when Sa (x, y) is 1, the pixel point is a grinding crack seed point; when Sa (x, y) ═ 0, the pixel point is represented as a non-grinding-mark seed point, wherein mu and ν respectively represent the pixel point
Figure GDA0001425146420000038
Offset of rows and columns.
Preferably, in the eighth step, the initial pattern H of the wear pattern area is obtained according to formula (3):
Figure GDA0001425146420000039
wherein, T1The self-adaptive threshold value is obtained, and the value is the gray level mean value of all the pixels of the grinding mark seed points in the grinding mark seed graph Sa; when H (x, y) is 1, the pixel point is a seed point of the wear-scar area; and when H (x, y) is 0, the pixel point is a seed point of the non-grinding area.
Preferably, in the tenth step, an abnormal pattern M of the whetting is obtained according to the formula (4)l：
Figure GDA0001425146420000041
When M islWhen (x, y) is 1, the abnormal graph M representing the pixel point as the grinding spotlPixel points; when M islWhen (x, y) is 0, the abnormal graph M shows that the pixel point is non-grinding spotlAnd (6) pixel points.
Preferably, in the tenth step, the speckle difference map M is obtained according to the formula (5)p：
Figure GDA0001425146420000042
Wherein M isl(x, y) and Mp(x, y) respectively represent an abnormality map MlAnd the speckle differential map MpThe pixel value of the middle pixel point (x, y).
Preferably, in the sixteenth step, when lc≥0.1mm、Sr≥0.01mm、Jr≥0.05mm、rdNot less than 0.35mm or ldAnd when the size is more than or equal to 0.1mm, judging that the grinding spot image is abnormal.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides a four-ball friction test grinding mark image abnormity detection method, which comprises the steps of firstly, taking an area block as a processing object, taking a central area as a reference area, finding out a comparison area with the same size and shape as the reference area in the central horizontal direction, then calculating the grinding mark difference degree of the comparison area and the reference area at the same direction angle, calculating the direction angle of a grinding mark shaft through the grinding mark difference degree, obtaining a grinding mark alignment diagram according to the direction angle of the grinding mark, further obtaining a grinding mark seed diagram, and finally obtaining a grinding mark area diagram; then, according to the area of the grinding area image, calculating the equivalent radius and the direction radius of the grinding area, and according to the equivalent radius and the direction radius, calculating the variance and the range of the direction radius; and calculating the axial difference and the centrifugal degree of the grinding spot area map according to the grinding spot area of the grinding spot area map, and judging the abnormity of the grinding spot image according to any one parameter of the variance of the direction radius, the extreme difference of the direction radius, the axial difference or the centrifugal degree. On the basis of following the existing standard, the method also increases indexes such as equivalent radius, radius range, shaft difference, centrifugation degree and the like to judge the abnormity of the grinding spot image, so that the judgment result is more accurate and effective; meanwhile, the invention automatically analyzes the shape and appearance of the wear scar image, finds abnormal wear scar images and is convenient for testers to analyze the reasons of the abnormal wear scar images.
Drawings
FIG. 1 is an image of an abrasion patch F;
FIG. 2 is a gray scale mura pattern f;
FIG. 3 is a direction angle θ;
FIG. 4 is a diagram of a rotating scrub pattern fθ；
FIG. 5 is a schematic view showing the correction of the grinding spots
Figure GDA0001425146420000051
FIG. 6 is an initial view H of the scrub spot area;
FIG. 7 is a diagram M of the swelling of the seeds with grind marksg；
FIG. 8 is a speckle differential map Mp；
FIG. 9 is a pattern M of the wear patternr。
Detailed Description
The present invention will be described in further detail with reference to the following drawings and examples.
The invention relates to a four-ball friction test grinding spot image abnormity detection method, which specifically comprises the following steps:
step S0, acquiring the wear scar image and carrying out gray processing on the wear scar image:
the steel ball and the scale are placed in a scanning electron microscope, the scanning electron microscope is used for collecting the grinding spot image F (shown in figure 1) of the steel ball and the white scale to be detected, and the grinding spot image F is subjected to gray processing to obtain a gray grinding spot image F (shown in figure 2) so as to eliminate color information on the surface of the grinding spot. Assuming that the size of the collected gray-scale speckle pattern f is mxn, the coordinate of a certain pixel point is (x, y), and x and y represent the row number and the column number of the pixel point, the requirements are as follows: x and y are integers, x is more than or equal to 1 and less than or equal to m, and y is more than or equal to 1 and less than or equal to n. In the embodiment, m is 768, n is 1024, and the graying processing adopts a weighted average method.
Step S1, calculating unit pixel length h:
and detecting a scale area in the grinding spot image, and calculating the actual length h of the unit pixel in the grinding spot image by combining the actual size of the scale. In this embodiment, h is 0.00067 mm/pixel.
Step S2, image rotation:
the grinding spots obtained by the four-ball friction test are mainly in the appearance formed by mutual friction among the steel balls and are expressed as grinding marks and abnormal abrasion (such as ablation, adhesion and the like). The grinding marks are distributed in a strip shape, the distribution characteristics are used for measuring the similarity of adjacent areas under different angles, the significance of the strip distribution characteristics of the grinding marks under different direction angles can be measured, when the direction angle theta (shown in figure 3) is consistent with the direction of the grinding marks, the similarity of the adjacent areas is the highest, and the strip distribution characteristics of the grinding marks are the most significant. The direction angle is an included angle formed by any ray and a horizontal right ray by taking the image central point O as a vertex.
In actual operation, the area blocks along the direction and the angular direction of the image are difficult to select and complicated to calculate, so the present invention adopts a method of rotating the image to turn the direction angle to the horizontal direction and convert the area blocks along the direction and the angular direction into horizontal processing (as shown in fig. 4). Clockwise rotating the gray scale abrasion pattern f by theta degrees around the central point O of the image to obtain a rotary abrasion pattern fθAssuming that it is m in sizeθ×nθAt this time, calculating the similarity of the region blocks in the horizontal direction is equivalent to calculating the similarity of the region blocks of the gray-scale pattern f in the azimuth of the direction angle θ.
Step S3, reference area aθAnd a contrast region DθSelecting:
considering that the edge area of the grinding spot is easily interfered by noise and grinding dust, and the strip-shaped characteristic of the central area is most obvious, the central area of the grinding spot is selected as a reference area to measure the grinding mark similarity when the direction angle theta.
Specifically, the rotational wear pattern f obtained in step S2θHas a central point O (coordinate of
Figure GDA0001425146420000061
) Arbitrarily selecting a certain area as the center, called as a reference area AθWherein, in the step (A),
Figure GDA0001425146420000062
and
Figure GDA0001425146420000063
are all integers, [ 2 ]]Representing a rounding operation.
Reference area AθThe shape of the grinding surface is a circle, a rhombus or a square, the length of the side length or the diameter of the grinding surface is 2 omega +1 pixel units, wherein the minimum elementary feature and the operation speed of the grinding mark are considered, and the value range of omega is 3-20. In the present embodiment, the reference region AθIs square, its area
Figure GDA0001425146420000064
Is of a size of
Figure GDA0001425146420000065
Where ω is 5.
In order to compare the similarity of the region blocks in the central horizontal direction, the reference region A is used in the central horizontal directionθAs the center, symmetrically selecting and referencing areas A to both sidesθS area blocks having the same size and shape as the comparison area DθThe value range of s is 1-10; then 2s contrast areas D are obtainedθComparison region DθNumbered from left to right as
Figure GDA0001425146420000071
Figure GDA0001425146420000072
Denotes the k-th contrast region D from the leftθ。
At the same time, selecting the contrast area DθWhen the two adjacent area blocks are partially overlapped or not overlapped, the distance between the center points of the two adjacent area blocks is preferably 2 ω + 1.
According to the characteristics s of the grinding spot image, the number is 3, and the adjacent area blocks do not overlap.
Step S4, calculating the wear scar axial angle:
to reduce the influence of noise, test specification and other factors, the comparison regions D are respectively calculatedθThe k-th contrast region of
Figure GDA0001425146420000073
And a reference area AθDegree of difference therebetween
Figure GDA0001425146420000074
I.e. calculating the contrast areas separately
Figure GDA0001425146420000075
Itself and eight adjacent directional area blocks and reference area AθThe gray difference between the two areas is obtained, and the minimum value of the nine obtained gray differences is taken as the k-th contrast area
Figure GDA0001425146420000076
And a reference area AθDegree of difference therebetween
Figure GDA0001425146420000077
Calculating all the contrast areas D with the same direction angle thetaθAnd a reference area AθThe difference therebetween is taken as the wear scar difference C of the direction angle thetaθ(ii) a When the difference degree of grinding marks CθWhen the minimum value is reached, the direction angle theta is the axial angle of the grinding mark
Figure GDA0001425146420000078
Contrast region
Figure GDA0001425146420000079
The eight direction neighborhood blocks are contrast areas
Figure GDA00014251464200000710
Upper left, upper right, upper left, right, lower leftAnd the area block is obtained after slightly moving 1 pixel unit in the eight directions at the lower right.
The calculation process of the axial angle of the grinding mark is shown as the formula (1).
Figure GDA00014251464200000711
Wherein i and j are integers, which are the row number and the column number of the pixel coordinates of the comparison area and the reference area respectively,
Figure GDA00014251464200000712
alpha and beta are respectively contrast regions
Figure GDA00014251464200000713
The amount of fine shift in the row direction and the column direction of (a) is taken as α ═ 1,0,1 and β ═ 1,0, 1; a. theθ(i, j) is a reference region A of the direction angle thetaθThe gray value of the pixel point (i, j) in (1);
Figure GDA00014251464200000714
the k-th contrast region at the direction angle theta
Figure GDA00014251464200000715
The gray value of the pixel point (i + α, j + β + k (2 ω +1)) in (1).
In the present embodiment, the first and second electrodes are,
Figure GDA00014251464200000716
step S5, rotating the grayscale shading map:
rotating the gray scale abrasion pattern f obtained in the step S0 counterclockwise
Figure GDA0001425146420000081
Then, a grinding spot correction chart as shown in FIG. 5 was obtained
Figure GDA0001425146420000082
So that the direction of the grinding marks is vertical and the size of the grinding marks is
Figure GDA0001425146420000083
In the present embodiment, the first and second electrodes are,
Figure GDA0001425146420000084
step S6, extracting grinding trace seed points:
the pixel points in the grinding mark area have the characteristics of column direction similarity and row direction dissimilarity in the neighborhood; further, the grinding wheel mark correction map obtained in step S5
Figure GDA0001425146420000085
Randomly selecting an area block, simultaneously taking the area block and adjacent area blocks in eight directions as research areas, and extracting a grinding spot rectification diagram according to formula (2)
Figure GDA0001425146420000086
And obtaining a grinding crack seed graph Sa by using the seed pixels of the middle grinding crack area, specifically:
Figure GDA0001425146420000087
when Sa (x, y) is 1, the pixel point is a grinding crack seed point; when Sa (x, y) ═ 0, the pixel point is represented as a non-grinding-mark seed point, wherein mu and ν respectively represent the pixel point
Figure GDA0001425146420000088
Offset of rows and columns.
Step S7, primary image extraction of the worn spot area:
calculating the gray level mean value of all the grinding mark seed point pixels in the grinding mark seed map Sa, taking the gray level mean value as an adaptive threshold, and correcting the grinding mark obtained in the step S5
Figure GDA0001425146420000089
Dividing to obtain an initial graph H of the grinding area as shown in FIG. 6, wherein the calculation formula is shown in formula (3):
Figure GDA00014251464200000810
wherein, T1The self-adaptive threshold value is obtained, and the value is the gray level mean value of all the pixels of the grinding mark seed points in the grinding mark seed graph Sa; when H (x, y) is 1, the pixel point is a seed point of the wear-scar area; and when H (x, y) is 0, the pixel point is a seed point of the non-grinding area. In this example, T1＝127。
Step S8, performing grinding mark seed point expansion operation:
to connect the pixel seed points in the adjacent regions, the wear scar seed map Sa obtained in step S6 is subjected to morphological dilation operation to obtain a wear scar seed dilation map M as shown in fig. 7g. According to the morphological characteristics of the grinding area, the structural operator of the expansion operation selects a circular structural operator of 5 multiplied by 5-15 multiplied by 15. In this example, 7 × 7 circular structural elements are taken.
Step S9, extracting an abnormal brightness region of the scrub spot:
for the initial pattern H (obtained from step S7) and the expansion pattern M of the grinding mark seedgPerforming a logical operation (obtained in step S8) to eliminate an abnormal brightness region caused by the photographing environment and obtain an abnormal pattern M of the wear markslThe calculation formula is shown in formula (4):
Figure GDA0001425146420000091
when M islWhen (x, y) is 1, the abnormal graph M representing the pixel point as the grinding spotlPixel points; when M islWhen (x, y) is 0, the abnormal graph M shows that the pixel point is non-grinding spotlAnd (6) pixel points.
Step S10, difference processing between the abnormal map and the initial map of the scrub area:
for the initial pattern H and abnormal pattern M of the grinding spot arealDifference processing was performed to obtain a shading difference map M shown in FIG. 8pThe calculation formula is shown in formula (5):
Figure GDA0001425146420000092
wherein M isl(x, y) and Mp(x, y) respectively represent an abnormality map MlAnd the speckle differential map MpThe pixel value of the middle pixel point (x, y).
Step S11, scrub spot area extraction:
for the difference image M of the grinding spotspPerforming morphological close operation, i.e. expansion and then corrosion, to obtain a pattern M of the abraded area as shown in FIG. 9r. And according to the morphological characteristics of the grinding area, selecting a 5 multiplied by 5-15 multiplied by 15 circle by a structural operator of the expansion operation. In this example, 7 × 7 circular structural elements are taken.
Step S12, calculating equivalent radius and centroid coordinates:
according to the pattern M of the wear scar arearHas an uneven area (the area of the uneven area is M)rThe number of all pixels with the middle pixel value of 1) is calculated, and the equivalent radius r of the grinding spot area is calculateddCalculating the centroid O of the grinding spot area according to the homogeneity characteristic of the grinding spot arearCoordinate (x) ofc,yc)，xcAnd ycAre respectively a centroid OrThe row and column numbers of; then, the equivalent radius r is converted according to the unit pixel length hdThe true value of (d) in mm.
In this example, rd＝0.23551mm，xc＝659，yc＝685。
Step S13, calculating a radius index parameter based on the unit pixel length:
the direction radius is defined as the centroid O under different ray directionsrThe farthest distance between the outer contour points of the grinding spots is designated by the symbol rrAnd (4) showing. Assuming that the rays are directed clockwise and outward in steps of 1 deg., the radius of the direction r isrWhich contains 360 data.
According to the direction radius rrCalculating the variance S of the directional radiusrRadial deviation from the sum direction Jr. Extreme difference in radius JrDefined as the radius of direction rrDifference between maximum value and minimum value of(ii) a Then converting the variance S of the direction radius according to the unit pixel length hrRadial deviation from the sum direction JrThe true distance of (c).
In this example, Sr＝0.02310mm2，Jr＝0.1408mm。
Step S14, calculating an axis parameter in combination with the unit pixel length:
the axis parameters include the major axis, minor axis, axis coordinates, axis difference and eccentricity of the scrub spot area. Major axis MrThe length of the longest line segment of the middle wear scar region along the wear scar direction is denoted by symbol laRepresents; the minor axis is the length of the longest line segment in the vertical direction of the grinding mark in the grinding mark region and is denoted by symbol lbRepresents; the intersection point of the long axis and the short axis is the axis OzBy (x)z,yz) Coordinates representing the axis; the axial difference is the major axis laAnd a minor axis lbBy the difference of (a), with the symbol lcAnd (4) showing. The eccentricity is defined as the scrub center OzAnd centroid OrEuropean distance betweendThen, the axial difference l is converted according to the unit pixel length hcAnd degree of centrifugation ldThe true distance of (c).
In this example, xz＝839，yz＝602，lc＝0.09711mm，ld＝0.13280mm。
Step S15, abnormality determination:
the abnormal occurrence of the wear pattern is mainly caused by the abnormal operation of a tester or the abnormal abrasion resistance of a lubricant. According to long-term research experience and expert research, the rule of abnormity judgment is as follows: when l iscNot less than 0.1mm or SrNot less than 0.01mm or JrNot less than 0.05mm or rdNot less than 0.35mm or ldAnd when the size is more than or equal to 0.1mm, judging the image to be an abnormal grinding spot image.
In this example, Jr＝0.14082mm≥0.05mm、ldThe abrasion mark image in the embodiment is judged to be an abnormal abrasion mark image because 0.13280mm is larger than or equal to 0.1mm, and the further analysis is carried out, which is caused by the poor lubricating effect of the lubricant; according to the method specified by the national standard, if the average of two measurements of a bottom ball is equal to all six measurementsIf the mean deviation is more than 0.04mm, the abnormality of the test result cannot be judged, and therefore, the method provided by the invention has an obvious effect.
According to the technical scheme of the invention, advantages and disadvantages of the method are compared with those of the current manual judgment method.
Firstly, the time and the labor are saved. The invention provides a method for automatically detecting an abnormal grinding spot image, which can automatically analyze the image and appearance of the grinding spot image, find the abnormal grinding spot image and save a great deal of time and energy of testers; secondly, it is accurate and effective. Compared with the existing method for judging the anomaly of the wear-scar image, the method is characterized in that the anomaly of the wear-scar image is judged by adding indexes such as equivalent radius, radius range, shaft difference, centrifugal degree and the like on the basis of following the existing standard, so that the judgment result is more accurate and effective.

Claims (8)

1. A four-ball friction test grinding spot image abnormity detection method is characterized by comprising the following steps:
firstly, collecting a steel ball to be measured and a white scale by using a scanning electron microscope to obtain a wear scar image F; carrying out gray processing on the grinding spot image F by a weighted average method to obtain a gray grinding spot image F;
secondly, calculating the actual length h of a unit pixel in the grinding spot image;
thirdly, clockwise rotating the gray scale abrasion pattern f by theta degrees around the central point O of the image to obtain a rotary abrasion pattern fθ；
The fourth step, rotating the abrasion pattern fθA certain area is selected as a reference area A by taking the central point O as the centerθ(ii) a In the central horizontal direction, with reference to the area AθAs the center, symmetrically selecting and referencing areas A to both sidesθS area blocks having the same size and shape as the comparison area DθIn total, 2s contrast regions D are obtainedθ；
Fifthly, comparing the areas D according to the same direction angle thetaθAnd a reference area AθDegree of difference C of grinding marksθCalculating the axial angle of the grinding mark
Figure FDA0002857681240000011
Sixthly, rotating the gray scale abrasion pattern f obtained in the first step counterclockwise
Figure FDA0002857681240000012
And (4) obtaining a grinding spot correction diagram
Figure FDA0002857681240000013
Seventh, on the above-mentioned grinding spot correcting map
Figure FDA0002857681240000014
Extracting seed pixels of the grinding crack area to obtain a grinding crack seed image Sa;
eighthly, calculating the gray mean value of all the grinding mark seed point pixels in the grinding mark seed map Sa, and taking the gray mean value as an adaptive threshold value to correct the grinding marks
Figure FDA0002857681240000015
Dividing to obtain an initial pattern H of the wear scar area;
ninth, performing morphological expansion operation on the grinding crack seed graph Sa to obtain a grinding crack seed expansion graph Mg；
The tenth step is to carry out the initial graph H of the grinding spot area and the expansion graph M of the grinding trace seedsgPerforming difference operation to obtain an abnormal speckle pattern Ml；
The tenth step is that an initial pattern H and an abnormal pattern M of the grinding spots are comparedlPerforming difference processing to obtain a speckle difference image Mp；
The twelfth step, the speckle difference map M is comparedpPerforming morphological close operation to obtain a mottled area map Mr；
The tenth step, according to the abrasion spot area diagram MrThe area of the grinding spot area, and calculating a grinding spot area diagram MrEquivalent radius r ofdDirection radius rrAnd centroid OrCoordinate (x) ofc,yc) And according to the unit imageEquivalent radius r is converted from element length hdThe true distance of (d);
fourteenth, according to the pattern M of the wear-scar regionrRadius in the direction of (r)rCalculating the variance S of the radius of the directionrAnd extreme difference J of direction radiusrThen, the variance S of the direction radius is calculated according to the unit pixel length hrRadial deviation from the sum direction JrThe true distance of (d);
the fifteenth step, according to the abrasion spot area diagram MrLong axis l ofaMinor axis lbAnd axis OzCalculating the pattern M of the mottled arearAxial difference lcAnd degree of centrifugation ldThen, the axial difference l is converted according to the unit pixel length hcAnd degree of centrifugation ldThe true distance of (d);
sixteenth, according to the variance S of the direction radiusrDirection radius extremely different JrAxial difference lcOr degree of centrifugation ldAnd judging the abnormality of the grinding spot image.
2. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the fourth step, the reference region AθIs a square of 2 omega +1 pixels on a side, the area of which
Figure FDA0002857681240000021
Is of a size of
Figure FDA0002857681240000022
Wherein the value range of omega is 3-20; in the selected contrast area DθWhen the positions of two adjacent area blocks are partially overlapped or not overlapped.
3. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the fifth step, the axial angle of the grinding mark
Figure FDA0002857681240000029
The calculation process of (2) is shown in the formula (1)：
Figure FDA0002857681240000023
Wherein i and j are integers, which are the row number and the column number of the pixel coordinates of the comparison area and the reference area respectively,
Figure FDA0002857681240000024
alpha and beta are respectively k-th contrast area
Figure FDA0002857681240000025
The amount of fine shift in the row direction and the column direction of (a) is taken as α ═ 1,0,1 and β ═ 1,0, 1; a. theθ(i, j) is a reference region A of the direction angle thetaθThe gray value of the pixel point (i, j) in (1);
Figure FDA0002857681240000026
the k-th contrast region at the direction angle theta
Figure FDA0002857681240000027
The gray value of the pixel point (i + alpha, j + beta + k (2 omega +1)) in (1); the value range of omega is 3-20.
4. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the seventh step, a grinding spot rectification chart is extracted according to the formula (2)
Figure FDA00028576812400000210
And obtaining a grinding crack seed graph Sa by using the seed pixels of the middle grinding crack area, specifically:
Figure FDA0002857681240000028
when Sa (x, y) is 1, the pixel point is a grinding crack seed point; when Sa (x, y) is 0, it indicates that the pixel isNon-grinding-mark seed points, wherein mu and v respectively represent pixel points
Figure FDA0002857681240000034
Offset of rows and columns.
5. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the eighth step, an initial pattern H of the wear-mark area is obtained according to the formula (3):
Figure FDA0002857681240000031
wherein, T1The self-adaptive threshold value is obtained, and the value is the gray level mean value of all the pixels of the grinding mark seed points in the grinding mark seed graph Sa; when H (x, y) is 1, the pixel point is a seed point of the wear-scar area; and when H (x, y) is 0, the pixel point is a seed point of the non-grinding area.
6. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the tenth step, an anomaly map M of the whets is obtained according to the formula (4)l：
Figure FDA0002857681240000032
When M islWhen (x, y) is 1, the abnormal graph M representing the pixel point as the grinding spotlPixel points; when M islWhen (x, y) is 0, the abnormal graph M shows that the pixel point is non-grinding spotlAnd (6) pixel points.
7. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: the tenth step, obtaining a speckle difference map M according to the formula (5)p：
Figure FDA0002857681240000033
Wherein M isl(x, y) and Mp(x, y) respectively represent an abnormality map MlAnd the speckle differential map MpThe pixel value of the middle pixel point (x, y).
8. The method for detecting the image abnormality of the four-ball friction test wear scar according to claim 1, characterized in that: in the sixteenth step, when lc≥0.1mm、Sr≥0.01mm、Jr≥0.05mm、rdNot less than 0.35mm or ldAnd when the size is more than or equal to 0.1mm, judging that the grinding spot image is abnormal.
CN201710597206.0A 2017-07-20 2017-07-20 Four-ball friction test speckle image anomaly detection method Expired - Fee Related CN107358604B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710597206.0A CN107358604B (en) 2017-07-20 2017-07-20 Four-ball friction test speckle image anomaly detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710597206.0A CN107358604B (en) 2017-07-20 2017-07-20 Four-ball friction test speckle image anomaly detection method

Publications (2)

Publication Number Publication Date
CN107358604A CN107358604A (en) 2017-11-17
CN107358604B true CN107358604B (en) 2021-04-06

Family

ID=60284517

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710597206.0A Expired - Fee Related CN107358604B (en) 2017-07-20 2017-07-20 Four-ball friction test speckle image anomaly detection method

Country Status (1)

Country Link
CN (1) CN107358604B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111383225B (en) * 2020-03-20 2023-04-07 长安大学 Method for detecting abnormal state of tire by comparing standard deformation rates
CN111429428B (en) * 2020-03-20 2022-05-13 长安大学 Method for detecting abnormal grinding image based on appearance characteristics

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103712565A (en) * 2013-12-31 2014-04-09 长安大学 Grinding crack diameter measurement method based on steel ball grinding crack gradient
CN105701816A (en) * 2016-01-13 2016-06-22 上海海事大学 Automatic image segmentation method
CN106770296A (en) * 2017-01-11 2017-05-31 长安大学 A kind of four ball friction tests mill spot image polishing scratch deflection automatic measuring method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103712565A (en) * 2013-12-31 2014-04-09 长安大学 Grinding crack diameter measurement method based on steel ball grinding crack gradient
CN105701816A (en) * 2016-01-13 2016-06-22 上海海事大学 Automatic image segmentation method
CN106770296A (en) * 2017-01-11 2017-05-31 长安大学 A kind of four ball friction tests mill spot image polishing scratch deflection automatic measuring method

Also Published As

Publication number Publication date
CN107358604A (en) 2017-11-17

Similar Documents

Publication Publication Date Title
JP6598162B2 (en) Visual identification method of multi-type BGA chip based on linear clustering
CN108960237B (en) Reading identification method for pointer type oil level indicator
CN107239742B (en) Method for calculating scale value of instrument pointer
CN108613630B (en) Two-wire tube level bubble offset measurement method based on image processing technology
CN110906875B (en) Visual processing method for aperture measurement
CN107341802A (en) It is a kind of based on curvature and the compound angular-point sub-pixel localization method of gray scale
CN107358604B (en) Four-ball friction test speckle image anomaly detection method
CN103292701A (en) Machine-vision-based online dimensional measurement method of precise instrument
CN116188496B (en) Remote sensing image self-adaptive segmentation method based on land utilization type
CN106770296B (en) A kind of four ball friction tests mill spot image polishing scratch deflection automatic measuring method
CN111445438B (en) Double-neighborhood grinding crack angle automatic detection method
CN105046257A (en) Character recognition method and system for special marking form of main gear of rear axle
CN116563279B (en) Measuring switch detection method based on computer vision
CN112085675A (en) Depth image denoising method, foreground segmentation method and human motion monitoring method
CN114693659B (en) Copper pipe surface cleaning effect evaluation method and system based on image processing
CN103712565B (en) A kind of wear scar diameter measuring method based on steel ball polishing scratch gradient
CN115272349A (en) Method for detecting welding defects of computer PCB main board
CN111932494A (en) Tire wear degree evaluation method and device
CN111412864B (en) Automatic grinding crack angle detection method based on grinding crack gray scale similarity
CN106841575B (en) A kind of four ball friction tests mill spot image polishing scratch direction automatic positioning method
CN107392899B (en) Automatic detection method for horizontal angle of grinding mark of steel ball grinding spot image
CN103761729A (en) Steel ball grinding crack detection method based on neighborhood gray level similarity
CN107403445B (en) Morphology comparison analysis method for wear resistance of lubricant
CN108846839B (en) Sagger crack detection method based on sector neighborhood difference histogram
CN111429428A (en) Abnormal grinding image detection method based on appearance characteristics

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210406