CN110517276A - A kind of grinding wheel dressing method and device based on machine vision - Google Patents
A kind of grinding wheel dressing method and device based on machine vision Download PDFInfo
- Publication number
- CN110517276A CN110517276A CN201910631969.1A CN201910631969A CN110517276A CN 110517276 A CN110517276 A CN 110517276A CN 201910631969 A CN201910631969 A CN 201910631969A CN 110517276 A CN110517276 A CN 110517276A
- Authority
- CN
- China
- Prior art keywords
- straight line
- group
- point
- circular arc
- poor
- 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.)
- Pending
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B24—GRINDING; POLISHING
- B24B—MACHINES, DEVICES, OR PROCESSES FOR GRINDING OR POLISHING; DRESSING OR CONDITIONING OF ABRADING SURFACES; FEEDING OF GRINDING, POLISHING, OR LAPPING AGENTS
- B24B53/00—Devices or means for dressing or conditioning abrasive surfaces
- B24B53/017—Devices or means for dressing, cleaning or otherwise conditioning lapping tools
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- 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
-
- 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
Abstract
The present invention relates to a kind of grinding wheel dressing method and device based on machine vision, the dressing method of the grinding wheel obtains the contour line information of plurality of positions when rotating by grinding wheel, the contour line information that will acquire obtains basic line segment by segment processing, using the very poor bounce as working lining of maximum of the basic line segment of same position, working lining jerk value can accurately be calculated, to be modified accordingly, reduce working lining bounce, improve grinding wheel contour accuracy, so that the grinding wheel after finishing has better ground effect, it can be used for accurate grinding, the stability of wheel grinding performance after helping to improve finishing.
Description
Technical field
The present invention relates to a kind of grinding wheel dressing method and device based on machine vision.
Background technique
Grinding wheel is usually used in being ground article, such as electroplating abrasion wheel, ceramics, resin, metal wheel, electroplating abrasion wheel typically refer to use
Electric plating method by single layer of abrasive particles and matrix metal co-deposition on steel substrate, to produce stone layer.It compares
In extra hard material grinding wheels such as resin, ceramics and sintering metal bonding agents, electroplating abrasion wheel have bonding agent it is strong to abrasive grain holding power,
Abrasive grain reveals the advantages that sword is high, grinding wheel chip space is big；In addition to this, with the development of Numeric Control Technology, steel substrate can be made into respectively
The complicated curved surface of kind, i.e., electroplating abrasion wheel excircle configuration can be made into the curved surface of various complexity, therefore be widely used in form grinding.So
And due to electroplating abrasion wheel is not easy to control etc. there are wear particle concentration and height of protrusion consistency, lead to the bounce of stone layer
It is poor with cross section profile precision, it is typically considered to be only applicable to the corase grinding of part complex profile.
In order to modify to grinding wheel, need to know the profile of the working lining of grinding wheel to be trimmed, in order to obtain the profile, now
There is technology to acquire profile information by way of machine vision, is irradiated again particular by backlight along the tangential of grinding wheel to be trimmed
The profile information of grinding wheel to be trimmed can be obtained by the acquisition of corresponding CCD camera；It can be right according to obtained profile information
The correction of the flank shape situation of grinding wheel to be trimmed is detected, and can also be obtained trim amount according to profile information and be modified to grinding wheel.But
For the accurate grinding wheel shaping of small profile variations, the prior art is when handling collected outline data, processing essence
It spends lower, the requirement of accurate repairing type is not achieved.
Summary of the invention
The object of the present invention is to provide a kind of grinding wheel dressing method based on machine vision, to solve it is existing to grinding wheel into
The poor problem of the precision of row finishing；The present invention provides a kind of wheel dresser based on machine vision, existing to solve
The poor problem of the precision that device when in use modifies grinding wheel.
To achieve the goals above, the present invention provides a kind of grinding wheel dressing method based on machine vision, including following step
It is rapid:
1) one group of original image when the rotation of acquisition grinding wheel, every width original image include a wheel coplanar with grinding wheel shaft
The information of profile；Each width original image is pre-processed to obtain the pixel edge point for indicating the contour line；
2) contour line is split according to the Curvature varying of the pixel edge point, obtains N item basis line segment, ground line
Section is straightway and/or arc section；
If 3) basic line segment includes straightway, straight line fitting, all images are carried out to pixel edge point in each straightway
The straight line of same section is one group on middle expression contour line, and each group straight line is very poor after digital simulation；If basic line segment includes circle
Segmental arc then carries out round fitting to pixel edge point section in each arc section, the circle of same section on contour line is indicated in all images
Arc is one group, and each group circular arc is very poor after digital simulation；It is described very poor for outermost layer basis line segment and innermost layer basis line segment
Difference；
4) when pixel edge point is only divided into straightway, take the very poor maximum of each group straight line as trim amount to sand
Wheel working lining is modified；When pixel edge point is only divided into arc section, takes the very poor maximum of each group circular arc to be used as and repair
Whole amount modifies stone layer；When pixel edge point is divided into straightway and arc section, each group straight line and circular arc are taken
Very poor maximum stone layer is modified as trim amount.
Beneficial effect is that grinding wheel obtains the contour line information of plurality of positions when rotating, and the contour line information that will acquire is passed through
Segment processing obtains basic line segment, can be accurate using the very poor bounce as working lining of maximum of the basic line segment of same position
Calculate working lining jerk value, to be modified accordingly, reduce working lining bounce, improve grinding wheel contour accuracy, make
Grinding wheel after must modifying has better ground effect, can be used for accurate grinding, the wheel grinding after helping to improve finishing
The stability of energy.
Further, in order to which operation has the extraction of better robustness and pixel edge point more accurate, the pre- place
Reason are as follows:
1) gray processing processing is carried out to each width original image, obtains the gray value of each pixel, generate grayscale image；
2) establish coordinate system by coordinate origin of the pixel in the lower left corner of grayscale image, wherein X positive direction horizontally to the right, root
According to the pixel edge point of each grayscale image of Canny operator extraction integrated in Matlab.
Further, in order to accurately obtain the very poor calculation amount of straight line, the very poor calculating process of any group of straight line is as follows:
1) starting point coordinate of each straight line and terminating point coordinate in the group are obtained,
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 21: starting point P21(xl21,yl21), terminating point P '21(x′l21,y′l21)；
……
Straight line N1: starting point PN1(xlN1,ylN1), terminating point P 'N1(x′lN1,y′lN1)；
Wherein, N is the number of original image；
2) x is enabledrange1=max { xl11,xl21,......,xlN1}
x′range1=min { x 'l11,x′l21,......,x′lN1}
Cross (xrange1, 0) and be straight line L1 perpendicular to X-axis, then each straight line has 1 intersection point in straight line L1 and the group,
The set of these intersection points is denoted as:
I1={ (x1N,y1N)|y1N=knxrange1+bn}
I1In each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
Cross (x 'range1, 0) and be straight line L2 perpendicular to X-axis, then every straight line has 1 intersection point in straight line L2 and the group,
The set of these intersection points is denoted as:
I′1={ (x1N,y1N)|y1N=knx′range1+bn}
I1' in each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
3) the very poor of this group of straight line is calculated are as follows:
Further, in order to accurately obtain the very poor calculation amount of circular arc, the very poor calculating process of any group of circular arc is as follows:
1) it is fitted to obtain the radius of each section of circular arc in the group by Least Square Circle, be denoted as: R1={ r11,r21,......,
rN1, wherein N is the number of original image；
2) radial difference of each section of circular arc and first segment circular arc is calculated:
KN1=rN1-r11；
3) the very poor of this group of circular arc is calculated are as follows:
DI1=max KN1-min KN1。
To achieve the goals above, the present invention also provides a kind of wheel dressers based on machine vision, including storage
Device, processor and storage in memory and the computer program that can run on a processor, described in the processor execution
It is performed the steps of when program
1) one group of original image when the rotation of acquisition grinding wheel, every width original image include a wheel coplanar with grinding wheel shaft
The information of profile；Each width original image is pre-processed to obtain the pixel edge point for indicating the contour line；
2) contour line is split according to the Curvature varying of the pixel edge point, obtains N item basis line segment, ground line
Section is straightway and/or arc section；
If 3) basic line segment includes straightway, straight line fitting, all images are carried out to pixel edge point in each straightway
The straight line of same section is one group on middle expression contour line, and each group straight line is very poor after digital simulation；If basic line segment includes circle
Segmental arc then carries out round fitting to pixel edge point section in each arc section, the circle of same section on contour line is indicated in all images
Arc is one group, and each group circular arc is very poor after digital simulation；It is described very poor for outermost layer basis line segment and innermost layer basis line segment
Difference；
4) when pixel edge point is only divided into straightway, take the very poor maximum of each group straight line as trim amount to sand
Wheel working lining is modified；When pixel edge point is only divided into arc section, takes the very poor maximum of each group circular arc to be used as and repair
Whole amount modifies stone layer；When pixel edge point is divided into straightway and arc section, each group straight line and circular arc are taken
Very poor maximum stone layer is modified as trim amount.
Beneficial effect is that grinding wheel obtains the contour line information of plurality of positions when rotating, and the contour line information that will acquire is passed through
Segment processing obtains basic line segment, can be accurate using the very poor bounce as working lining of maximum of the basic line segment of same position
Calculate working lining jerk value, to be modified accordingly, reduce working lining bounce, improve grinding wheel contour accuracy, make
Grinding wheel after must modifying has better ground effect, can be used for accurate grinding, the wheel grinding after helping to improve finishing
The stability of energy, in addition the device provides certain hardware structure to be easy to implement above-mentioned method.
Further, in order to which operation has the extraction of better robustness and pixel edge point more accurate, in the device
The pretreatment are as follows:
1) gray processing processing is carried out to each width original image, obtains the gray value of each pixel, generate grayscale image；
2) establish coordinate system by coordinate origin of the pixel in the lower left corner of grayscale image, wherein X positive direction horizontally to the right, root
According to the pixel edge point of each grayscale image of Canny operator extraction integrated in Matlab.
Further, in order to accurately obtain the very poor calculation amount of straight line, any group of the very poor of straight line was calculated in the device
Journey is as follows:
1) starting point coordinate of each straight line and terminating point coordinate in the group are obtained,
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 21: starting point P21(xl21,yl21), terminating point P '21(x′l21,y′l21)；
……
Straight line N1: starting point PN1(xlN1,ylN1), terminating point P 'N1(x′lN1,y′lN1)；
Wherein, N is the number of original image；
2) x is enabledrange1=max { xl11,xl21,......,xlN1}
x′range1=min { x 'l11,x′l21,......,x′lN1}
Cross (xrange1, 0) and be straight line L1 perpendicular to X-axis, then each straight line has 1 intersection point in straight line L1 and the group,
The set of these intersection points is denoted as:
I1={ (x1N,y1N)|y1N=knxrange1+bn}
I1In each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
Cross (x 'range1, 0) and be straight line L2 perpendicular to X-axis, then every straight line has 1 intersection point in straight line L2 and the group,
The set of these intersection points is denoted as:
I′1={ (x1N,y1N)|y1N=knx′range1+bn}
I1' in each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
3) the very poor of this group of straight line is calculated are as follows:
Further, in order to accurately obtain the very poor calculation amount of circular arc, any group of the very poor of circular arc was calculated in the device
Journey is as follows:
1) it is fitted to obtain the radius of each section of circular arc in the group by Least Square Circle, be denoted as: R1={ r11,r21,......,
rN1, wherein N is the number of original image；
2) radial difference of each section of circular arc and first segment circular arc is calculated:
KN1=rN1-r11；
3) the very poor of this group of circular arc is calculated are as follows:
DI1=max KN1-min KN1。
Detailed description of the invention
Fig. 1 is a kind of principle schematic diagram of grinding wheel dressing method based on machine vision of the invention；
Fig. 2 is the width original image that stone layer radial section high resolution CCD camera of the invention is shot；
Fig. 3 is the Canny operator extraction edge that one radial section of stone layer of the invention utilizes Matlab to integrate
Binary map afterwards；
Fig. 4 is after pixel edge point coordinate points are converted to actual size by one radial section of stone layer of the invention
Profile diagram；
Fig. 5 is the image processing process that grinding wheel of the invention determines trim amount based on working lining bounce；
In figure, 1 is grinding wheel to be trimmed, and 2 be rotary shaft, and 3 be workbench, and 4 be main shaft, and 5 be contact roller for dressing, and 6 be computer.
Specific embodiment
The present invention will be further described in detail with reference to the accompanying drawing.
Embodiment of the method:
The present invention provides a kind of grinding wheel dressing method based on machine vision, and this method needs to acquire one when grinding wheel rotation
Group original image is arranged in Fig. 1 as shown in Figure 1, the present invention provides the schematic illustration for obtaining the original image of the grinding wheel profile
There are high resolution CCD camera, camera lens, back side light source, main shaft 4, rotary shaft 2, workbench 3, computer 6, image processing and analyzing soft
Part, motion controller, servo-system.Grinding wheel 1 to be trimmed is mounted in rotary shaft 2, and contact roller for dressing 5 is mounted on main shaft 4, the back side
Light source vertical irradiation in the work layer surface of grinding wheel 1 to be trimmed, adopt with image is mounted with by high resolution CCD camera and controller
The computer 6 of truck is connected, and rotary shaft 2 drives grinding wheel 1 to be trimmed slowly at the uniform velocity to rotate, and high resolution CCD camera is with fixation
Frequency is continuously shot stone layer radial section image i.e. original image, and is stored in computer 6 by image pick-up card
Hard disk in, then using Matlab establishment image processing program to each image carry out calculation process most preferably modified
Amount, and it is transmitted to motion controller, motion controller drives servo system control contact roller for dressing 5 theoretical according to grinding wheel radial section
The finishing path movement that figure and best trim amount generate, and feed in X direction, it is finally complete to grinding wheel 1 to be trimmed in X/Y plane
It is modified at interpolation.Grinding wheel to be trimmed is referred to as grinding wheel below.
Specific process method step is as follows:
It 1) include grinding wheel contour line information in the original image, as shown in Fig. 2, being stone layer radial section high score
One width original image of resolution CCD camera shooting pre-processes each width original image to obtain the pixel side for indicating contour line
Edge point；Wherein, it pre-processes to carry out gray processing processing to each width original image, obtains the gray value of each pixel, generate gray scale
Figure；Establish coordinate system by coordinate origin of the pixel in the lower left corner of grayscale image, wherein X positive direction horizontally to the right, according to
The pixel edge point of each grayscale image of Canny operator extraction integrated in Matlab；Wherein, it is calculated using the Canny that Matlab is integrated
Son obtains binary map after extracting edge, as shown in figure 3, pixel edge point coordinate points to be converted to the profile after actual size again
Figure, as shown in Figure 4.Pretreatment can also carry out image procossing using operators other in Matlab, as long as can extract to obtain
Pixel edge point.
For convenience of subsequent descriptions, all original images can successively be named as Image1, Image2 ... ..., ImageN,
N is the number of original image.
2) it is split according to the Curvature varying of these marginal points, finds out characteristic point, complex figure is resolved into ground line
Section, i.e. straight line and circular arc, then Image1 may be expressed as: { straight line 11, straight line 12 ..., straight line 1n, circular arc 11, circular arc 12 ..., circle
Arc 1n }；Characteristic point is the starting point and ending point of every section of basic line segment, takes out the starting point and ending point of every section of basic line segment
Coordinate:
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 12: starting point P12(xl12,yl12), terminating point P '12(x′l12,y′l12)；
……
Straight line 1n: starting point P1n(xl1n,yl1n), terminating point P '1n(x′l1n,y′l1n)；
Circular arc 11: starting point R11(xr11,yr11), terminating point R '11(x′r11,y′r11)；
Circular arc 12: starting point R12(xr12,yr12), terminating point R '12(x′r12,y′r12)；
……
Circular arc 1n: starting point R1n(xr1n,yr1n), terminating point R '1n(x′r1n,y′r1n)。
Every section of basic line segment is fitted using least square method, to straight line 11, straight line 12 ... ..., straight line 1n is carried out
Straight line fitting；To circular arc 11, circular arc 12 ... ..., circular arc 1n carries out round fitting；Fit procedure can not only use least square
Existing following line fitting method and circle approximating method can also be used in method.
It repeats the above steps, processing fortune is carried out to remaining original image Image2, Image3 ... ..., ImageN respectively
It calculates, show that each image marginal point is split, find out characteristic point, complex figure is decomposed into be in line N1, straight line N2 ... ...,
Straight line Nn carries out straight line fitting；And circular arc N1, circular arc N2 ... ..., circular arc Nn carry out Least Square Circle fitting.I.e. are as follows:
Image2:{ straight line 21, straight line 22 ..., straight line 2n, circular arc 21, circular arc 22 ..., circular arc 2n }；
Image3:{ straight line 31, straight line 32 ..., straight line 3n, circular arc 31, circular arc 32 ..., circular arc 3n }；
……
ImageN:{ straight line N1, straight line N2 ..., straight line Nn, circular arc N1, circular arc N2 ..., circular arc Nn }.
3) the very poor of every section after being fitted in all images basic line segment is calculated, steps are as follows:
1. Image2, Image3 ... ..., the figure obtained after ImageN fitting are added to Image1's by Image1
In coordinate system.
2. calculating the very poor of all straight lines.
The very poor difference for outermost layer basis line segment and innermost layer basis line segment can use following calculation method, can also
To be determined using other modes, such as the distance between outermost layer basis line segment center and innermost layer basis line segment center are used as pole
Difference.
Wherein, the very poor process for calculating the 1st group of straight line is as follows:
Take out set S1={ straight line 11, straight line 21 ... ..., straight line N1 }, i.e.,
S1={ f (xlN1,ylN1)|ylN1=knxlN1+bn, n is natural number }, knIndicate the slope of straight line N1.
The starting point and ending point of every straight line are as follows:
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 21: starting point P21(xl21,yl21), terminating point P '21(x′l21,y′l21)；
……
Straight line N1: starting point PN1(xlN1,ylN1), terminating point P 'N1(x′lN1,y′lN1)。
It enables, xrange1=max { xl11,xl21,......,xlN1,
x′range1=min { x 'l11,x′l21,......,x′lN1}。
Cross (xrange1, 0) and be straight line L1 perpendicular to X-axis, then straight line L1 and S1In every straight line have 1 intersection point, these
The set of intersection point is denoted as:
I1={ (x1N,y1N)|y1N=knxrange1+bn}
I1In each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
N=N+1, N are natural number；
Cross (x 'range1, 0) and be straight line L2 perpendicular to X-axis, then straight line L2 and S1In every straight line have 1 intersection point, this
The set of a little intersection points is denoted as:
I′1={ (x '1N,y′1N)|y′1N=knx′range1+bn}；
I1' in each intersection point (x '1N,y′1N) and (x '11,y′11) distance are as follows:
N=N+1, N are natural number；
Then the 1st group of straight line is very poor are as follows:
Similarly, calculate the 2nd group it is very poor to N group straight line:
ThenIn maximum be jitter values 1 in stone layer, be denoted as C1:
3. calculating the very poor of all circular arcs.
Wherein, the very poor process for calculating the 1st group of circular arc is as follows:
Set { circular arc 11, circular arc 21 ... ..., circular arc N1 } is taken out, the radius of every section of circular arc is obtained by being fitted in set, i.e.,
Are as follows:
R1={ r11,r21,......,rN1}；
Every section of circular arc seeks difference with first segment circular arc, i.e., are as follows:；
KN1=rN1-r11, N=2,3....n；
Then the 1st group of circular arc is very poor are as follows:
DI1=max KN1-minKN1；
Similarly, the 2nd group it is very poor to N group circular arc are as follows:
DIn=max KNn-minKN1；
Then DInIn maximum be jitter values 2 in stone layer, be denoted as C2:
C2=max DIn。
In above method embodiment, contour line is divided into straightway and two kinds of arc section basic line segments, is then respectively obtained
Working lining jitter values 2 when working lining jitter values 1 and basic line segment when basic line segment is straightway are arc section, take therein
Trim amount of the maximum value as working lining bounce as grinding wheel.
It in actual use, can also only be straightway or arc section by Contour segmentation.If basic line segment is straight line
Section carries out straight line fitting to pixel edge point in each straightway using corresponding fitting algorithm, indicates contour line in all images
The straight line of upper same section is one group, and each group straight line is very poor after digital simulation；Take the very poor maximum conduct of each group straight line
Trim amount modifies stone layer.
If basic line segment is arc section, pixel edge point section in each arc section is justified using corresponding fitting algorithm
It is fitted, indicates that the circular arc of same section on contour line is one group in all images, each group circular arc is very poor after digital simulation；It takes each
The very poor maximum of group circular arc modifies stone layer as trim amount.
Above-mentioned basis line segment is straightway and basic line segment is that the specific method of arc section is still referred to the above method
The process of straightway and arc section is calculated in embodiment, and details are not described herein.
The present invention provides a kind of dressing process of specific grinding wheel, take a kind of working lining profile as the plating of Gothic circular arc
For CBN grinding wheel.The nominal dimension of the grinding wheel and the contour accuracy of working lining are as shown in table 1.Gothic circular arc is by two and half
The equal circular arc of diameter is spliced, wherein left and right contact angle refers to the two of the virtual steel ball that a radius is 3mm and Gothic circular arc
When section circular arc is tangent, two point of contacts of left and right and the line of the steel ball centre of sphere and the angle of vertical direction.
Table 1
| Nominal dimension | CBN granularity | Left and right contact angle | Left and right arc radius |
| 40*10*8-R3.118 | 120/140 | 42±3° | 3.118±0.01mm |
Firstly, electroplating CBN grinding wheel to be mounted in rotary shaft 2 and fasten.
Then, the position between rotary shaft 2 and high resolution CCD camera is adjusted, stone layer radial section is being counted
Imaging clearly on calculation machine 6, fixed CCD camera position at this time, the technical parameter of CCD camera are as shown in table 2.
Rotary shaft 2 is mounted on workbench 3, is equipped with handwheel on workbench 3 and locking device, handwheel can control workbench 3
It is moved in X direction with Y-direction；It is moved in addition, high resolution CCD camera is manually controllable along Z-direction.By adjusting workbench 3
Mutual alignment between CCD camera can make electroplating abrasion wheel working lining radial section imaging clearly on computer 6.
Table 2
| Model | Port View 300 |
| Pixel [pixel] | 1280 (level) × 1024 (vertical) |
| Field range [mm] | 5.69 (level) × 4.55 (vertical) |
Then, by instruction input computer 6, and it is sent to motion controller, motion controller driving servo-system makes to revolve
Shaft 2 drives electroplating abrasion wheel slowly at the uniform velocity to rotate at least one week, and CCD camera is continuously shot stone layer radial direction with fixed frequency
Section original image, and pass sequentially through figure capture card and be stored in the hard disk of computer 6, and these images are named as
Image1, Image2 ... ..., ImageN.
In computer 6, every width original image is carried out at operation using the image processing software worked out based on Matlab
Reason obtains best trim amount Vbest, as shown in figure 5, Section1, Section1 ..., Section1 indicate original image,
Superposing image indicates image after superposition, is later gray level image, steps are as follows:
1. the image Image1, Image2 ... ..., ImageN that high resolution CCD camera photographed are subjected to gray processing processing,
Obtain each image grayscale image.
2. establish coordinate system by coordinate origin of the pixel in the image lower left corner, wherein X positive direction horizontally to the right, according to
The Canny operator integrated in Matlab extracts pixel edge point to every width original image.
3. the Curvature varying according to these marginal points is split, using highest data point as characteristic point, by a point left side
The marginal point of right two sides is divided into two parts in left and right and obtains left and right arc radius respectively with Least Square Circle fitting is carried out, be denoted as
r1And R1.All images are analyzed, the set { r of left arc radius is obtained1, r2…rnAnd right arc radius set { R1, R2…
Rn, the very poor of two set is calculated separately out, and select maximum value as crushing amount V1, it is computed and obtains V1For
0.0152mm。
Finally, generating finishing path according to the theoretical profile of electroplating abrasion wheel working lining radial section and importing in computer 6
CAM software, CAM software will modify coordinates measurement numerical control program, according to numerical control program motion controller drive servo-system control
Contact roller for dressing 5 processed moves, and feeds in X direction, amount of feeding 0.0152mm, finally completes interpolation to electroplating abrasion wheel in X/Y plane
Finishing.The basic fundamental parameter of contact roller for dressing 5 is as shown in table 3.
Table 3
| Specification [mm] | 3F1 125*31.75*6*2 |
| Abrasive grain | Diamond |
| Wear particle size | 170/200 |
| Bonding agent | Sintering metal bonding agent |
| It modifies revolving speed [rpm] | 6500 |
Detection of the present invention also to stone layer contour accuracy, using electroplating CBN wheel grinding after finishing with a thickness of 2mm
Graphite print, with contourgraph HOMMEL T8000RC200-400 detect graphite print, can reflect the profile of stone layer
Precision.
Installation practice:
The present invention provides a kind of wheel dresser based on machine vision, including memory, processor and is stored in
In memory and the computer program that can run on a processor, processor are realized in above method embodiment when executing program
Method and step；Certainly, which can be fitted into computer 6 as shown in Figure 1.
The grinding wheel dressing method can be used for complex profile electroplating abrasion wheel, breach current industry internal cause complex profile plating sand
Wheel, trim amount are unable to accurately control, and are not easy to carry out the limitation of stone layer shaping using the method for interpolation finishing.In the present invention
Based on the image processing software of Matlab establishment, the Method for Accurate Calculation of the trim amount based on the bounce of stone layer is proposed,
It ensure that stone layer contour accuracy, the dressing efficiency of grinding wheel can be greatly promoted.
Claims (8)
1. a kind of grinding wheel dressing method based on machine vision, which comprises the following steps:
1) one group of original image when the rotation of acquisition grinding wheel, every width original image include a contour line coplanar with grinding wheel shaft
Information；Each width original image is pre-processed to obtain the pixel edge point for indicating the contour line；
2) contour line is split according to the Curvature varying of the pixel edge point, obtains N item basis line segment, basic line segment is
Straightway and/or arc section；
If 3) basic line segment includes straightway, straight line fitting, table in all images are carried out to pixel edge point in each straightway
The straight line for showing same section on contour line is one group, and each group straight line is very poor after digital simulation；If basic line segment includes arc section,
Round fitting then is carried out to pixel edge point section in each arc section, indicates that the circular arc of same section on contour line is one in all images
Group, each group circular arc is very poor after digital simulation；The very poor difference for outermost layer basis line segment and innermost layer basis line segment；
4) when pixel edge point is only divided into straightway, take the very poor maximum of each group straight line as trim amount to grinding wheel work
It is modified as layer；When pixel edge point is only divided into arc section, take the very poor maximum of each group circular arc as trim amount
Stone layer is modified；When pixel edge point is divided into straightway and arc section, the pole of each group straight line and circular arc is taken
The maximum of difference modifies stone layer as trim amount.
2. the grinding wheel dressing method according to claim 1 based on machine vision, which is characterized in that the pretreatment are as follows:
1) gray processing processing is carried out to each width original image, obtains the gray value of each pixel, generate grayscale image；
2) establish coordinate system by coordinate origin of the pixel in the lower left corner of grayscale image, wherein X positive direction horizontally to the right, according to
The pixel edge point of each grayscale image of Canny operator extraction integrated in Matlab.
3. the grinding wheel dressing method according to claim 1 or 2 based on machine vision, which is characterized in that any group of straight line
Very poor calculating process it is as follows:
1) starting point coordinate of each straight line and terminating point coordinate in the group are obtained,
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 21: starting point P21(xl21,yl21), terminating point P '21(x′l21,y′l21)；
……
Straight line N1: starting point PN1(xlN1,ylN1), terminating point P 'N1(x′lN1,y′lN1)；
Wherein, N is the number of original image；
2) x is enabledrange1=max { xl11,xl21,......,xlN1}
x′range1=min { x 'l11,x′l21,......,x′lN1}
Cross (xrange1, 0) and be straight line L1 perpendicular to X-axis, then each straight line has 1 intersection point in straight line L1 and the group, these
The set of intersection point is denoted as:
I1={ (x1N,y1N)|y1N=knxrange1+bn}
I1In each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
Cross (x 'range1, 0) and be straight line L2 perpendicular to X-axis, then every straight line has 1 intersection point in straight line L2 and the group, these
The set of intersection point is denoted as:
I′1={ (x1N,y1N)|y1N=knx′range1+bn}
I1' in each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
3) the very poor of this group of straight line is calculated are as follows:
4. the grinding wheel dressing method according to claim 1 or 2 based on machine vision, which is characterized in that any group of circular arc
Very poor calculating process it is as follows:
1) it is fitted to obtain the radius of each section of circular arc in the group by Least Square Circle, be denoted as: R1={ r11,r21,......,rN1,
Wherein, N is the number of original image；
2) radial difference of each section of circular arc and first segment circular arc is calculated:
KN1=rN1-r11；
3) the very poor of this group of circular arc is calculated are as follows:
DI1=max KN1-min KN1。
5. a kind of wheel dresser based on machine vision, including memory, processor and storage are in memory and can
The computer program run on a processor, which is characterized in that the processor performs the steps of when executing described program
1) one group of original image when the rotation of acquisition grinding wheel, every width original image include a contour line coplanar with grinding wheel shaft
Information；Each width original image is pre-processed to obtain the pixel edge point for indicating the contour line；
2) contour line is split according to the Curvature varying of the pixel edge point, obtains N item basis line segment, basic line segment is
Straightway and/or arc section；
If 3) basic line segment includes straightway, straight line fitting, table in all images are carried out to pixel edge point in each straightway
The straight line for showing same section on contour line is one group, and each group straight line is very poor after digital simulation；If basic line segment includes arc section,
Round fitting then is carried out to pixel edge point section in each arc section, indicates that the circular arc of same section on contour line is one in all images
Group, each group circular arc is very poor after digital simulation；The very poor difference for outermost layer basis line segment and innermost layer basis line segment；
4) when pixel edge point is only divided into straightway, take the very poor maximum of each group straight line as trim amount to grinding wheel work
It is modified as layer；When pixel edge point is only divided into arc section, take the very poor maximum of each group circular arc as trim amount
Stone layer is modified；When pixel edge point is divided into straightway and arc section, the pole of each group straight line and circular arc is taken
The maximum of difference modifies stone layer as trim amount.
6. the wheel dresser according to claim 5 based on machine vision, which is characterized in that the pretreatment are as follows:
1) gray processing processing is carried out to each width original image, obtains the gray value of each pixel, generate grayscale image；
2) establish coordinate system by coordinate origin of the pixel in the lower left corner of grayscale image, wherein X positive direction horizontally to the right, according to
The pixel edge point of each grayscale image of Canny operator extraction integrated in Matlab.
7. the wheel dresser according to claim 5 or 6 based on machine vision, which is characterized in that any group of straight line
Very poor calculating process it is as follows:
1) starting point coordinate of each straight line and terminating point coordinate in the group are obtained,
Straight line 11: starting point P11(xl11,yl11), terminating point P '11(x′l11,y′l11)；
Straight line 21: starting point P21(xl21,yl21), terminating point P '21(x′l21,y′l21)；
……
Straight line N1: starting point PN1(xlN1,ylN1), terminating point P 'N1(x′lN1,y′lN1)；
Wherein, N is the number of original image；
2) x is enabledrange1=max { xl11,xl21,......,xlN1}
x′range1=min { x 'l11,x′l21,......,x′lN1}
Cross (xrange1, 0) and be straight line L1 perpendicular to X-axis, then each straight line has 1 intersection point in straight line L1 and the group, these
The set of intersection point is denoted as:
I1={ (x1N,y1N)|y1N=knxrange1+bn}
I1In each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
Cross (x 'range1, 0) and be straight line L2 perpendicular to X-axis, then every straight line has 1 intersection point in straight line L2 and the group, these
The set of intersection point is denoted as:
I′1={ (x1N,y1N)|y1N=knx′range1+bn}
I1' in each intersection point (x1N,y1N) and (x11,y11) distance are as follows:
Wherein, n=N+1；
3) the very poor of this group of straight line is calculated are as follows:
8. the wheel dresser according to claim 5 or 6 based on machine vision, which is characterized in that any group of circular arc
Very poor calculating process it is as follows:
1) it is fitted to obtain the radius of each section of circular arc in the group by Least Square Circle, be denoted as: R1={ r11,r21,......,rN1,
Wherein, N is the number of original image；
2) radial difference of each section of circular arc and first segment circular arc is calculated:
KN1=rN1-r11；
3) the very poor of this group of circular arc is calculated are as follows:
DI1=max KN1-min KN1。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910631969.1A CN110517276A (en) | 2019-07-12 | 2019-07-12 | A kind of grinding wheel dressing method and device based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910631969.1A CN110517276A (en) | 2019-07-12 | 2019-07-12 | A kind of grinding wheel dressing method and device based on machine vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110517276A true CN110517276A (en) | 2019-11-29 |
Family
ID=68623290
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910631969.1A Pending CN110517276A (en) | 2019-07-12 | 2019-07-12 | A kind of grinding wheel dressing method and device based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110517276A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112873051A (en) * | 2021-01-16 | 2021-06-01 | 陈晖� | Distributed grinding wheel dressing system |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06114556A (en) * | 1992-09-30 | 1994-04-26 | Yamaha Corp | Data preparing method for working path of robot |
| US20080156061A1 (en) * | 2006-12-27 | 2008-07-03 | Denso Corporation | Method of producing honeycomb structure molding die |
| US20100245592A1 (en) * | 2009-03-31 | 2010-09-30 | Aisin Seiki Kabushiki Kaisha | Calibrating apparatus for on-board camera of vehicle |
| CN104858788A (en) * | 2015-05-28 | 2015-08-26 | 江苏万达特种轴承有限公司 | Grinding method for external diametral plane large circular arc |
| CN106041740A (en) * | 2016-07-25 | 2016-10-26 | 郑州磨料磨具磨削研究所有限公司 | Automatic tool setting method and device for precision trimming of grinding wheel |
| CN107756252A (en) * | 2017-10-18 | 2018-03-06 | 郑州磨料磨具磨削研究所有限公司 | A kind of grinding wheel dressing method and device based on hydrodynamic driving |
| CN107886972A (en) * | 2012-12-29 | 2018-04-06 | Hoya株式会社 | Glass substrate for disc and disk |
| CN108081134A (en) * | 2017-11-16 | 2018-05-29 | 上海交通大学 | Grinding wheel profile situ detection system, method and grinding wheel profile error compensating method |
| CN108747825A (en) * | 2018-06-19 | 2018-11-06 | 湖南大学 | A kind of view-based access control model detection laser reconditioning forming grinding wheel device and its dressing method |
| CN109531425A (en) * | 2018-12-29 | 2019-03-29 | 厦门创云精智机械设备股份有限公司 | A kind of full-automatic abrasive wheel finishing machine and its finishing grinding wheel method |
| CN109612390A (en) * | 2018-12-17 | 2019-04-12 | 江南大学 | Large-size workpiece automatic measuring system based on machine vision |
| US10894293B2 (en) * | 2014-12-16 | 2021-01-19 | Profilator Gmbh & Co. Kg | Hob peeling method and cutting tool for producing at least partially rounded tooth tips |
-
2019
- 2019-07-12 CN CN201910631969.1A patent/CN110517276A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06114556A (en) * | 1992-09-30 | 1994-04-26 | Yamaha Corp | Data preparing method for working path of robot |
| US20080156061A1 (en) * | 2006-12-27 | 2008-07-03 | Denso Corporation | Method of producing honeycomb structure molding die |
| US20100245592A1 (en) * | 2009-03-31 | 2010-09-30 | Aisin Seiki Kabushiki Kaisha | Calibrating apparatus for on-board camera of vehicle |
| CN107886972A (en) * | 2012-12-29 | 2018-04-06 | Hoya株式会社 | Glass substrate for disc and disk |
| US10894293B2 (en) * | 2014-12-16 | 2021-01-19 | Profilator Gmbh & Co. Kg | Hob peeling method and cutting tool for producing at least partially rounded tooth tips |
| CN104858788A (en) * | 2015-05-28 | 2015-08-26 | 江苏万达特种轴承有限公司 | Grinding method for external diametral plane large circular arc |
| CN106041740A (en) * | 2016-07-25 | 2016-10-26 | 郑州磨料磨具磨削研究所有限公司 | Automatic tool setting method and device for precision trimming of grinding wheel |
| CN107756252A (en) * | 2017-10-18 | 2018-03-06 | 郑州磨料磨具磨削研究所有限公司 | A kind of grinding wheel dressing method and device based on hydrodynamic driving |
| CN108081134A (en) * | 2017-11-16 | 2018-05-29 | 上海交通大学 | Grinding wheel profile situ detection system, method and grinding wheel profile error compensating method |
| CN108747825A (en) * | 2018-06-19 | 2018-11-06 | 湖南大学 | A kind of view-based access control model detection laser reconditioning forming grinding wheel device and its dressing method |
| CN109612390A (en) * | 2018-12-17 | 2019-04-12 | 江南大学 | Large-size workpiece automatic measuring system based on machine vision |
| CN109531425A (en) * | 2018-12-29 | 2019-03-29 | 厦门创云精智机械设备股份有限公司 | A kind of full-automatic abrasive wheel finishing machine and its finishing grinding wheel method |
Non-Patent Citations (2)
| Title |
|---|
| LI XUE等: "Monitoring of wheel dressing operations for precision grinding", 《2002 IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL TECHNOLOGY》 * |
| 彭檐波: "基于机器视觉的激光切向整形V形凹面砂轮测控系统设计", 《中国优秀博硕士学位论文全文数据库(硕士)(工程科技Ⅰ辑)》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112873051A (en) * | 2021-01-16 | 2021-06-01 | 陈晖� | Distributed grinding wheel dressing system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106041946B (en) | Image-processing-based robot polishing production method and production system applying same | |
| CN110815048B (en) | Grinding wheel dressing method and device based on machine vision | |
| AU2016251428B2 (en) | Method and device for machining a tool by removing material | |
| CN110153567A (en) | A kind of laser cutting system based on image recognition | |
| CN110728667A (en) | Automatic and accurate cutter wear loss measuring method based on gray level image probability | |
| CN113941919B (en) | Data processing and control method for automatic steel pipe weld joint polishing system | |
| CN113298833A (en) | Target object point cloud characteristic line and surface extraction method and system | |
| CN102490523A (en) | Automatic laser engraving system based on computer vision and service method of automatic laser engraving system | |
| CN114571326B (en) | Method, device and system for grinding deformed curved surface based on computer vision | |
| CN112017232B (en) | Positioning method, device and equipment for circular patterns in image | |
| CN1730248A (en) | Reverse engineering robot system | |
| Xu et al. | Fast on-machine profile characterization for grinding wheels and error compensation of wheel dressing | |
| Chen et al. | Automated vision positioning system for dicing semiconductor chips using improved template matching method | |
| CN101561250B (en) | Method for intelligent position finding and online measurement of large-dimension cam non-circular grinding | |
| CN111311618A (en) | Circular arc workpiece matching and positioning method based on high-precision geometric primitive extraction | |
| CN108089544A (en) | A kind of orbit generation method and control system of sole glue spraying robot | |
| KR20190121244A (en) | Method for generating grinding of a gear wheel workpiece and grinding machine having a controller for generating grinding of a gear wheel workpiece | |
| CN105643395A (en) | Grinding forming method for optical free-form surface | |
| CN110517276A (en) | A kind of grinding wheel dressing method and device based on machine vision | |
| Xu et al. | Methodology for the immediate detection and treatment of wheel wear in contour grinding | |
| CN103764344B (en) | Method for manufacturing a component using forging | |
| CN115937468A (en) | Automatic generation method for machining program of countless-module robot | |
| CN117677474A (en) | System and method for treating a work surface | |
| CN212320647U (en) | Rotary scanning equipment for circular object | |
| CN110837715B (en) | Complex curved surface machining error compensation method based on reverse engineering technology |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191129 |