CN109829897A - A kind of gear burr detection method and gear high-precision vision measuring system - Google Patents
A kind of gear burr detection method and gear high-precision vision measuring system Download PDFInfo
- Publication number
- CN109829897A CN109829897A CN201910042298.5A CN201910042298A CN109829897A CN 109829897 A CN109829897 A CN 109829897A CN 201910042298 A CN201910042298 A CN 201910042298A CN 109829897 A CN109829897 A CN 109829897A
- Authority
- CN
- China
- Prior art keywords
- gear
- image
- line
- subsystem
- burr
- 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.)
- Granted
Links
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of gear burr detection method and gear high-precision vision measuring systems.Gear measuring system of the invention includes: First look subsystem, the second vision subsystem, server and board subsystem, wherein the First look subsystem can obtain the first image of tested gear along the axis direction of the gear；Second vision subsystem can obtain the second image of tested gear along the axis direction perpendicular to the gear；The server can be handled the first image and the second image, to measure the formal parameter of tested gear；The board subsystem can send tested gear to the designated position for being suitble to acquisition image；The server detects burr using gear burr detection method of the invention.System of the invention can be convenient quick and accurately carry out gear vision measurement, improves measurement accuracy, improves environmental suitability.
Description
Technical field
The present invention relates to contactless workpiece calibration field more particularly to a kind of gear vision measurement systems.
Background technique
Industrial gear be widely used in machine driving and production equipment in.The precision of gear affects making for mechanical equipment
With performance and working life.
With the continuous improvement of the market demand, more and more enterprise requirements carry out full inspection to gear.It is in the prior art
Although the detection accuracy of three coordinates is met the requirements, but detection time is very long, sampling site it is few need a few minutes, the need scanned comprehensively
Dozens of minutes is wanted, the requirement of full inspection is unable to satisfy.Usually by manually carrying out full inspection after the completion of traditional Gear Production, assembly line is made
Industry great work intensity, worker cause the measurement of gear deviation occur, reduce vulnerable to the influence of the factors such as working environment and psychology
Production efficiency.
Vision-based detection is to realize to manufacture intelligentized effective way, is able to achieve the transition and upgrade of mechanical industry production.
The prior art has been able to detect with machine vision and the technology of image procossing to gear geometry parameter, to realize non-
Contact measurement.
Application No. is 201710968282.8, entitled " one kind are based on machine vision revolving body workpieces detection method "
Chinese invention patent application disclose a kind of detection method of revolving body workpieces, utilize the CMOS area array cameras of 5,000,000 pixels
Static shooting is carried out to revolving body workpieces.Although the inventive technique is in workpiece size smaller (maximum gauge is less than 25mm), shape
There are better effects in the case where rule (for example cylindrical type, or be circular ring type), required precision not high (> 0.05mm), but it is right
In this required precision of gear very high (< 0.02mm), the work of outer ring apical tooth root with teeth, diameter range larger (200mm-350mm)
For part, so that the technology of the invention is difficult to carry out, this is just needed, and a kind of environmental suitability is stronger, the higher vision measurement of precision
Technology.
Summary of the invention
The purpose of the present invention is to provide one kind can be convenient quick and accurate gear vision measurement system, improves measurement
Precision improves environmental suitability.
Technical scheme is as follows.
One aspect of the present invention provides a kind of gear burr detection method, comprising the following steps:
Step S1 acquires gear thickness gray level image；
Step S2 carries out Threshold segmentation to the gray level image, extracts gear region；
Step S3 carries out an opening operation to the area-of-interest, and is once expanded；
Step S4, the zonule after opening operation is subtracted using the big region after step S3 expansion, is obtained comprising edge
The image of band；
Step S5 is connected to image obtained in the step S4, extracts gear or more according to length and orientation
The marginal belt on surface；
Step S6 cuts original-gray image using the marginal belt of the gear upper and lower surface, obtains gear boundary
Spiculation visualization；
Step S7 disconnects the connection between the burr of the spiculation visualization, obtains discrete burr；
Step S8 judges whether to meet the requirements to each discrete burr.
Preferably, the step S7 is to the spiculation visualization Threshold segmentation, and carries out lateral encroaching, to obtain discrete
Burr.
Preferably, the step S8 judges whether it meets by the minimum circumscribed circle for the burr for asking each discrete and wants
It asks.
Another aspect provides a kind of gear measuring systems, including First look subsystem, the second vision
System, server and board subsystem, wherein
The First look subsystem can obtain the first image of tested gear along the axis direction of the gear；
Second vision subsystem can obtain the second of tested gear along the axis direction perpendicular to the gear
Image；
The server can be handled the first image and the second image, to measure the outer parameter of tested gear
Number；
The board subsystem can send tested gear to the designated position for being suitble to acquisition image；
It is characterized in that, the server uses gear burr detection side according to any one of claim 1-3
The burr of method detection gear.
Preferably, the First look subsystem includes: the first line-scan digital camera, the first doubly telecentric camera lens, magic hand, and line is swept
Light source, the first bar shaped backlight.
Preferably, the board central axes of the board subsystem are passed through in the linear acquisition visual field of first line-scan digital camera；
The distance of first doubly telecentric distance of camera lens gear under test upper surface is arranged to its camera lens object distance；First line-scan digital camera
It is mounted on first level sliding block by upright slide block；
The first bar shaped backlight, which is placed on turntable, to be fixedly connected, and is arranged on turntable central axes, with First Line
The acquisition of array camera, which is stitched, to be overlapped；
The line clears off source and is hold by magic hand, and is tilted a certain angle, and light is made just to fall in gear upper surface
On central axes；
The line clears off source and beats over from the direction for carrying chamfering pair, and below the first line-scan digital camera, chamfering is in right side, institute
Source is cleared off with line to squeeze into from left side.
Preferably, second vision subsystem includes: the second line-scan digital camera, the second doubly telecentric camera lens, the second bar shaped back
Light source.
Preferably, second line-scan digital camera is fixedly mounted on the second horizontal slider, and sliding with second level
Block passes through transparent turntable central axes together；
The second bar shaped backlight is fixedly mounted on board, and is arranged on the central axes of transparent turntable, face
The visual field pickup area of the second line-scan digital camera is completely covered in the visual field of second line-scan digital camera.
Preferably, the lower surface installation site of the second bar shaped backlight is lower than the table top of the board.
Preferably, the board subsystem includes: that PLC controller, first servo motor, the second servo motor, third are watched
Take motor, the 4th servo motor, feeding pawl disk, blanking pawl disk, transparent turntable；
The first servo motor can drive the first level sliding block to slide in the horizontal direction；The second servo electricity
Machine can drive the upright slide block to slide in vertical direction；The third servo motor can drive second horizontal slider
It slides in the horizontal direction；4th servo motor can drive the transparent turntable rotation；
When measurement, PLC controller can control feeding pawl disk and send gear to the transparent turntable, then start the 4th
Servo motor, the 4th servo motor is able to drive the transparent turntable and rotates, and after stabilization of speed, the PLC controller will
Starting measuring signal is sent to the server.
By above technical scheme, the present invention being capable of On-line rapid measurement gear parameter.The present invention can be examined once simultaneously
Survey 11 gear parameters, comprising: internal diameter, bottom of the tooth diameter, tooth top diameter, internal diameter circle bounce, tooth top diameter circle bounce, concentricity, common normal,
Face song amount (flank chamfering is wide), gear thickness, flatness and burr.And the present invention for the same gear internal diameter and
The repetition detection accuracy of tooth top diameter can achieve 0.005mm, and vision measurement result error compared with the measurement result of three coordinates is small
In 0.02mm, meet required precision.
Detailed description of the invention
Fig. 1 is system structure diagram of the invention；
Fig. 2 is the details position display diagram in Fig. 1；
Fig. 3 is to grayscale image horizontal segmentation and to extract gear parts schematic diagram；
Fig. 4 is to establish measured zone schematic diagram to the gear parts extracted in Fig. 3；
Fig. 5 is to extract marginal point schematic diagram to the measured zone established in Fig. 4；
Fig. 6 is to grayscale image Threshold segmentation and to extract gear parts schematic diagram
Fig. 7 is to the gear parts opening operation schematic diagram extracted in Fig. 6；
Fig. 8 is to establish measured zone schematic diagram to opening operation result in Fig. 7；
Fig. 9 is to seek marginal point schematic diagram to the measured zone in Fig. 8；
Figure 10 is to draw flank top edge rectangle measured zone schematic diagram to Fig. 7 opening operation result；
Figure 11 is to extract marginal point schematic diagram to the rectangle measured zone of Figure 10；
Figure 12 is to draw flank lower edge rectangle measured zone schematic diagram to Fig. 7 opening operation result；
Figure 13 is to extract marginal point schematic diagram to the measured zone of Figure 12；
Figure 14 is to draw chamfering rectangle measured zone schematic diagram above flank to Fig. 7 opening operation result；
Figure 15 is the rectangle measured zone Threshold segmentation schematic diagram to Figure 14；
Figure 16 is to measure curved surface amount schematic diagram to the region of Figure 15 segmentation；
Figure 17 is to extract gear parts schematic diagram to acquired image；
Figure 18 is to obtain minimum circumscribed rectangle according to Figure 17 to original image to cut rectangular area schematic diagram；
Figure 19 is the gear parts opening operation result schematic diagram to extraction；
Figure 20 is to create measured zone schematic diagram to the opening operation result of Figure 19；
Figure 21 is to seek eating root zone domain marginal point schematic diagram to the measured zone of Figure 20；
Figure 22 a, 22b, 22c are to extract marginal point schematic diagram again later to image mosaic；
Figure 23 is to seek diametric representation to fitting circle；
Figure 24 is to seek fitting circle concentricity schematic diagram；
Figure 25 is that base tangent length schematic diagram is sought in tooth combination；
Figure 26 is to gear thickness gray level image Threshold segmentation schematic diagram；
Figure 27 is to draw rectangle measured zone schematic diagram to the result of Figure 26 segmentation；
Figure 28 is the extracted region marginal point schematic diagram drawn to Figure 27；
Figure 29 is to draw support claw strip rectangle schematic diagram to gear thickness gray level image；
Figure 30 is the rectangular extraction support claw schematic diagram to original image according to Figure 29；
Figure 31 is to draw measured zone rectangle schematic diagram in gear lower edge；
Figure 32 is extraction marginal point schematic diagram in the region rectangle to Figure 31；
Figure 33 is to extract gear region opening operation schematic diagram to gear thickness gray level image；
Figure 34 is that primary expansion schematic diagram is carried out to the opening operation result of Figure 33；
Figure 35 is image border band image；
Figure 36 is gear boundary spiculation visualization Threshold segmentation schematic diagram；
Figure 37 is that discrete burr schematic diagram is obtained to Figure 36 lateral encroaching；
Figure 38 a, 38b are the first and second image information processing module processing result schematic diagrames；
Figure 39 a, 39b are the first and second image information processing module parameters input schematic diagrames；
Figure 40 is the first image information processing module schematic diagram as the result is shown；
Figure 41 is the second image information processing module schematic diagram as the result is shown.
The meaning of each appended drawing reference is as follows in figure.
1PLC controller, 2 blanking pawl disks, 3 first ball screws, 4 first servo motors, 5 first bar shaped backlights, 6
One horizontal slider, 7 lines clear off source, 8 magic hands, 9 first doubly telecentric camera lenses, 10 first line-scan digital cameras, 11 second ball-screws, 12
Second servo motor, 13 upright slide blocks, 14 feeding pawl disks, 15 gears, 16 transparent turntables, 17 second doubly telecentric camera lenses, 18 second
Line-scan digital camera, 19 second horizontal sliders, 20 third ball-screws, 21 third servo motors, 22 displays, 23 servers, 24 machines
Platform, 25 retarders, 26 the 4th servo motors, 27 second bar shaped backlights.
Specific embodiment
As shown in Figure 1, a kind of gear high precision measuring system according to the present invention, including First look subsystem, second
Four part of vision subsystem, server and board subsystem.
Gear to be measured can be sent to designated position by board subsystem.Gear 15 is sent to specified by the board subsystem
Behind position, the server 23 can be notified, further the server 23 can notify First look subsystem and second
Vision subsystem starts to acquire photo.It is described after the completion of the First look subsystem and second vision subsystem acquisition
The next gear 15 of board Sub-system Replacement, while the server starts to the First look subsystem and second view
Feel that the gear photo of subsystem shooting is handled, result is saved after the completion of processing and is shown on display screen 22.
In a preferred embodiment, the First look subsystem includes: the first line-scan digital camera 10, the first doubly telecentric
Camera lens 9, magic hand 8, line clear off source 7, the first bar shaped backlight 5.
24 central axes of board of the board subsystem are passed through in the linear acquisition visual field of first line-scan digital camera 10.It is described
Distance of the first doubly telecentric camera lens 9 apart from 15 upper surface of gear under test is arranged to its camera lens object distance, and the first line-scan digital camera
The 10 linear acquisition visual field can cover minimum diameter gear to be measured to the gear 15 of maximum outside diameter.First line-scan digital camera
10 are mounted on first level sliding block 6 by upright slide block 13.It is sliding by mobile first level if can not once be completely covered
Block 6 adjusts the visual field position of the first line-scan digital camera 10, is completely covered with realizing.
First bar shaped backlight 5, which is placed on turntable, to be fixedly connected, and is arranged on turntable central axes, with the first linear array phase
The acquisition of machine 10, which is stitched, to be overlapped.In order to make light fully enter the visual field of the first line-scan digital camera 10, the first bar shaped backlight 5 is set
For with specific length, so that in the case that the position of the first line-scan digital camera 10 changes in a certain range, the first bar shaped
Backlight 5 still is able to that the visual field of the first line-scan digital camera 10 is completely covered.
Line clears off source 7 and is hold by magic hand 8, and is tilted a certain angle, and light is made just to fall in 15 upper surface of gear
On central axes.It is possible thereby to which line is made to clear off the visual field that 7 reflected light of source enters the first line-scan digital camera 10 to greatest extent.
Line clears off source 7 and beats over from the direction for carrying chamfering pair, as shown in Fig. 2, below the first line-scan digital camera 10, chamfering
On right side, squeezed into so line clears off source 7 from left side.The purpose being arranged in this way is to try to line is made to clear off source 7 to beat in chamfering
Light reflection less than the first line-scan digital camera 10 the visual field inside, thus make shooting gray value of image pull open difference: background is most bright
(5 direct projection of the first bar shaped backlight), 15 upper surface of gear are taken second place, and most secretly (line clears off for (line clears off source 7 and reflects to greatest extent), chamfering
Source 7 most descends limit to reflect).
Second vision subsystem includes: the second line-scan digital camera 18, the second doubly telecentric camera lens 17, the second bar shaped backlight
27。
Second line-scan digital camera 18 is fixedly mounted on the second horizontal slider 19, and with second horizontal slider 19
Pass through transparent 16 central axes of turntable together.
Second bar shaped backlight 27 is fixedly mounted on board 24, and is arranged on the central axes of transparent turntable 16, just
To the visual field of the second line-scan digital camera 18, the visual field pickup area of the second line-scan digital camera 18 is completely covered.Second bar shaped backlight 27
Lower surface installation site be lower than table top.
The server 23 includes: image pick-up card, Ethernet interface, display 22, and configured at image analysis
Reason system.
The board subsystem includes: PLC controller 1, first servo motor 4, the second servo motor 12, third servo electricity
Machine 21, the 4th servo motor 26, first ball screw 3, the second ball-screw 11, third ball-screw 20, feeding pawl disk 14, under
Expect pawl disk 2, transparent turntable 16.
Enabling signal can be sent to server 23 by Ethernet interface by PLC controller 1, and server 23 notifies first
First line-scan digital camera 10 of vision subsystem and the second line-scan digital camera 18 of the second vision subsystem start to shoot photo.
The first servo motor 4 can drive first level sliding block 6 to slide in the horizontal direction.Second servo motor
12 can drive upright slide block 13 to slide in vertical direction.The third servo motor 21 can drive the second horizontal slider 19 to exist
Horizontal direction sliding.4th servo motor 26 can drive the transparent turntable 16 to rotate.
Collected picture is given described the by image pick-up card by the first line-scan digital camera 10 of First look subsystem
Collected picture is passed through image pick-up card by the second line-scan digital camera 18 of one image information processing module, the second vision subsystem
Give the second image information processing module.The two is all completed to notify 1 blanking of PLC controller by Ethernet interface later, and
Next waiting of gear 15 is replaced to measure next time.
Processing result is preserved and is sent by the first image information processing module and the second image information processing module
To on display screen 22.
It after system starting, is completed if be successfully tested with gear type selecting, PLC controller 1 controls first servo motor first
4, the second servo motor 12 and third servo motor 21 move to the designated position of each servo motor recorded under the gear model
Place, and no longer shift position before shutting down, wait it is to be measured, in order to avoid need to re-scale.
After starting measurement, PLC controller 1 controls feeding pawl disk 14 and send gear 15 to measured zone, i.e., transparent turntable 16
On, then start the 4th servo motor 26.4th servo motor 26 drives transparent turntable 16 to rotate by speed changer, transparent turntable
After 16 stabilizations of speed, PLC controller 1 is sent to server 23 for measuring signal is started by Ethernet interface.
In a preferred embodiment, described image analysis process system includes: image mosaic module, the first image letter
It ceases processing module, the second image information processing module, save submodule, camera parameters calibration submodule and Parameter File
Setting and reading submodule.
The picture that first line-scan digital camera 10 acquires can be spliced into complete gear graph by described image splicing module
Piece.Since the first line-scan digital camera 10 acquisition picture is the accumulation of a line a line, when camera resolution is extra high, acquisition one
The photo of whole circle gear can be especially big.In order to improve the software speed of service, the present invention by a whole bracing cable sweep picture be divided into it is more
Part, marginal point is extracted respectively to every portion.But will lead to a problem in this way, it is exactly that might have at photo decomposition
Tooth is divided into two, and can not carry out the accurate extraction of marginal point at this time.Therefore, in a preferred embodiment, side is being extracted
The tooth that both ends are first temporarily abandoned when edge point, after it the part lost can be finally stitched together, then extracts an edge
Point, as shown in Figure 22 a, Figure 22 b and Figure 22 c.
The first image message processing module includes: the first Image Acquisition submodule, the first image procossing and marginal point
Extracting sub-module, marginal point polar coordinate transform submodule, fitting circle submodule, diameter and circle bounce submodule, concentricity submodule
Block, common normal submodule.
Second image information processing module includes: the second Image Acquisition submodule, the second image procossing and marginal point
Extracting sub-module, facewidth computational submodule, flatness computational submodule, burr computational submodule.
The first image message processing module and the first Image Acquisition submodule of the second image information processing module and
The collected picture of image pick-up card can be saved preparation processing in memory by the second Image Acquisition submodule.
Particularly, the first image procossing of the first image message processing module further includes with marginal point extracting sub-module
Inner circle edges of regions point extraction module, the edge point extraction module of tooth top diameter, the edge point extraction module of common normal, face song amount
Peak width computing module and tooth root region edge point extraction module.
The inner circle edges of regions point extraction module extracts inner circle edges of regions point using following steps.
Firstly, collected grayscale image is carried out horizontal segmentation at n sections (n < picture height/measurement operator step pitch), such as Fig. 3
It is shown.
Low ash angle value part is extracted (because of gear ratio to the carrying out image threshold segmentation after each piece of segmentation using Do statement
Background wants secretly more), as shown in Figure 3.
Operation is opened and closed to the region extracted, and is connected to, wherein maximum region (the i.e. gear part of area is selected
Point), minimum circumscribed rectangle is asked to the gear region extracted, then creates a measurement in the left side edge of minimum circumscribed rectangle
Edge is sought in region, as shown in Figure 4.
The marginal point that left side metering model on each block of image obtains is taken together, end loop.
A measured zone is created, that is, draws a measurement rectangle, measures rectangle approximately perpendicular to boundary, start-stop point coordinate
It is artificially given with width.So that the true edge of objects in images is located in so-called measured zone, then in the survey drawn
It measures in region and extracts marginal point using sub- edge detection algorithm, as shown in Figure 5.
The edge point extraction module of the tooth top diameter extracts the marginal point of tooth top diameter using following steps.
Acquired image is subjected to Threshold segmentation first, extracting the lesser part of gray value (i.e. dark part) is to feel emerging
Interesting region, that is, gear parts, then carry out a closed operation gear regional connectivity once, as shown in Figure 6.Sense to extracting
Interest region carries out rectangle opening operation, and rectangle parameter calculation formula is as follows:
Wherein: w indicates the rectangle width that Katyuan is calculated, and unit is number of pixels；H indicates the rectangular elevation that Katyuan is calculated, unit
For number of pixels；D indicates reference diameter, and unit is millimeter；Di indicates interior circular diameter, and unit is millimeter；P indicates pixel ruler
Very little, unit is micron；W ' indicates transverse tooth thickness, and unit is millimeter.
Opening operation is carried out to target area, as a result as shown in Figure 7.
Then, regional connectivity is carried out, single stripe region is obtained, the corresponding stripe region of each tooth, then pass through
Do statement seeks minimum circumscribed rectangle to each stripe region, then creates one in the right side edge of minimum circumscribed rectangle
Measured zone seeks marginal point, as shown in Figure 8.The result for extracting marginal point is as shown in Figure 9.
Finally, the marginal point at each tooth top is summarized, and end loop.
The edge point extraction module of the common normal extracts the marginal point of common normal using following steps.
Step i extracts top flank edge.It is obtained according to the tooth top opening operation during tooth top diameter edge extracting each
The minimum circumscribed rectangle of a stripe region draws the rectangle measured zone for extracting flank top edge, as shown in Figure 10.Rectangle parameter
Calculation formula is as follows
Wherein: col0For the column coordinate of rectangle measured zone；row0For the row coordinate of rectangle measured zone；col2For strip
The column coordinate of the lower right angle point of region minimum circumscribed rectangle；row1For the upper left side angle point of stripe region minimum circumscribed rectangle
Row coordinate；Da is tooth top diameter, and unit is millimeter；W1 is rectangle width, and unit is number of pixels；M is modulus, and unit is millimeter；
H1 is rectangular elevation.
Marginal point is extracted as a result, as shown in figure 11 using sub-pixel edge detective operators in measurement rectangle.
Step ii extracts lower section flank edge.This extraction process and top edges are extracted similarly, and the two is about in tooth top
Axisymmetrical, as shown in Figure 12 and Figure 11.
Four, specific step is as follows for the peak width extraction of face song amount: being equally according to previous step with common normal operation
The minimum circumscribed rectangle of each stripe region obtained after tooth top opening operation draws the rectangle extracted above flank at chamfering
Region, such as figure Figure 14.Then the rectangular area, and Threshold segmentation are cut in original image, are partitioned into most dark region, such as Figure 15,
Then namely chamfering seeks the external parallelogram of minimum of the chamfering, such as Figure 16 can find out the width of the chamfering, industry
It is upper to be referred to as face song amount.
Five, tooth root region marginal point extraction specifically refers to, and acquired image is carried out Threshold segmentation first, extracts gray scale
Being worth lesser part (i.e. dark part) is area-of-interest i.e. gear parts, then carries out a closed operation and gear region is connected
Lead to, as shown in Figure 6.Then seek its minimum circumscribed rectangle, from original image on the basis of minimum circumscribed rectangle cut a square
The image in shape region, as shown in figure 17.Wherein, right margin of the right margin of the rectangular area of cutting than gear minimum circumscribed rectangle
To right avertence 2m, i.e. two moduluses, the number of pixels of deviation is 2m*100/p.Then Threshold segmentation is carried out to the image after cutting,
Extract background, i.e., bright gray areas.As shown in figure 18.Rectangle opening operation, rectangle ginseng are carried out to the area-of-interest extracted
Number calculation formula is as follows:
Wherein: w1Indicate the rectangle width that Katyuan is calculated, unit is number of pixels；h1Indicate the rectangular elevation that Katyuan is calculated, unit
For number of pixels.
Opening operation result is as shown in figure 19.Then as at tooth top, regional connectivity is carried out, single strip area is obtained
Domain, the corresponding stripe region of each tooth, then by Do statement, minimum circumscribed rectangle is asked to each stripe region, so
A measured zone is created in the left side edge of minimum circumscribed rectangle afterwards, as shown in figure 20.Marginal point, side are sought in measured zone
It is as shown in figure 21 that edge point extracts result.Finally the marginal point at each tooth root is summarized, and end loop.
The marginal point polar coordinate transform submodule of the first image message processing module, to extract inner circle, tooth top
Point on circle, root circle is ranked up from small to large with row coordinate, carries out polar coordinate transform, polar coordinate transform formula after sequence again
It is as follows.
Wherein: d0Distance for the board centre of gyration apart from camera image boundary, row and col are inner circle, outside circle, tooth
The row ordinate of edge collecting point on root circle, α are the installation angle of the first line-scan digital camera acquisition seam and board horizontal median axis,
The above unit is number of pixels.Angl is the corresponding actual angle of every a line of the first line-scan digital camera acquisition, and unit is degree.
The fitting circle submodule of the first image message processing module, first fits the initial center of circle using least square method
And radius, it is then round come iterative fitting using steepest descent method using the huber algorithm based on least square method optimized.
Step 1: seeking initial center of circle radius using common least square method fitting circle, equation of a circle be can be written as:
(x-xc)2+(y-yc)2=R2
Common least square fitting requires the quadratic sum of distance minimum.Namely
It is minimum.For the ease of seeking analytic solutions, above formula be can simplify are as follows:
F=∑ ((xi-xc)2+(yi-yc)2-R2)2
Enable g (x, y)=(x-xc)2+(y-yc)2-R2
So above formula can be expressed as f=∑ g (xi,yi)2。
It knows when f obtains minimum, it should meet following condition:
Initial central coordinate of circle (x can be found out by above equationc,yc), initial radius of circle R.
Step 2: initial center of circle radius, which is updated to huber algorithm, is iterated optimization.
It is iterated using the least square method optimized, the distance of single marginal point to target circle
eiIt can just bear, be brought into huber function as independent variable:
Contribution one positive sample error of overall error can be obtained, independent variable parameter vector to be asked is
Objective function are as follows:
The present invention is solved using steepest descent method, and the core of steepest descent method is that independent variable changes along negative gradient iteration
When, function decline is most fast.
It enables
ThenGradientFor
Step 3: the setting of iteration step length
According toCalculate current independent variableUnder target loss functionValue,
With the target loss function of last iterationCompare, if become smaller, returns to step 2 and carry out next iteration, until
Reach the upper limit less than a certain error threshold or cycle-index；If become larger, αk=αkη, η are descending factors (such as 0.5), are returned
It arrivesUntil meeting operation termination condition, circulation terminates, and returns to target circle.
Two conditions that operation stops:
Until cycle-index is greater than given threshold；
The difference of the value of the objective function of the value of objective function and last iteration is less than a certain given threshold.
The diameter of the first image message processing module and circle bounce submodule can be with after previous step fitting circle
The center of circle of edge discrete point after obtaining polar coordinate transform and radius, then press irradiation angle from these discrete points and uniformly extract N
Then part (wherein N be even number), the width (angular range β, wherein β < 360/N) of every portion utilize Do statement, find out often
Next a mean radius will can find out a diameter about the symmetrical two radiuses summation in the center of circle again, may finally
Find out N/2 diameter, end loop.As shown in figure 23.
When circle being asked to beat, to the N/2 diameter just acquired, maximum value and minimum value are extracted, then maximum value subtracts most
Small value can find out round bounce.
The concentricity submodule of the first image message processing module, according to requiring concentricity measurement inner circle and outside circle
Concentricity.After fitting circle, the center of circle of the edge discrete point after available polar coordinate transform and radius, then according to tooth
Discrete point is divided into Z parts (wherein Z is tooth top number), as shown in figure 24, then utilizes Do statement, ask by the quantity on top and position
The difference of the inner circle and outside circle mean radius of every portion out.The maximum value for finally finding out semidiameter subtracts minimum value, acquired results
As inner circle and outside circle concentricity.
The common normal submodule of the first image message processing module, in image procossing and marginal point extracting sub-module
The marginal point of common normal extract, marginal point coordinate required for common normal is measured under available rectangular coordinate system, how many
A tooth can be obtained by how many groups of common normal edge point datas, then utilize Do statement, to a teeth of i-th of tooth and the i-th+n into
Row combination, the number of teeth that wherein n uses common normal micrometer to block when measuring common normal in the process by actual processing.To each group of group
It closes, the polar coordinate transform of corresponding angle is carried out, so that this two tooth in combining all is laid flat, up and down about x-axis positive axis pair
Claim distribution, then utilize the row coordinate maximum value of lower section tooth, subtracts the row coordinate minimum value of top tooth, common normal that you can get it
Length, as shown in figure 25.Wherein the upper left angle point of picture is the origin that row ordinate is 0.
The image procossing of second image information processing module and marginal point extracting sub-module, include two parts, and first
Part is gear thickness top edge extraction module, and second part is gear thickness lower edge extraction module.
The extraction step of gear thickness top edge extraction module is as follows.
Threshold segmentation is carried out to collected gear thickness gray level image first and extracts dark part, so as shown in figure 26
After seek its minimum circumscribed rectangle, measurement rectangular area is drawn with the left margin of minimum circumscribed rectangle.Then sub-pixel edge is utilized
Detection method then extracts marginal point, as shown in figure 28 as shown in figure 27.
The extraction step of gear thickness lower edge extraction module is as follows.
The strip square of support claw is drawn in the fixed position in gear lower surface to collected gear thickness gray level image first
Shape, as shown in figure 29.It is obtained using rectangle interception original image and cuts picture, and Threshold segmentation is carried out to the picture after cutting, mentioned
Dark part is taken, is support claw.As shown in figure 30.Next support claw is skipped, draws measured zone square in gear lower edge
Shape, as shown in figure 31.Finally marginal point is extracted using sub-pixel edge method in the region rectangle of drafting.As shown in figure 32.
The gear THICKNESS CALCULATION submodule of second image information processing module, can obtain and be adopted by common normal submodule
The marginal point collected is divided into M group, and seeks column coordinate mean value, finally that top edge and lower edge row coordinate is opposite
The average value for two groups of the column coordinate answered is subtracted each other to get to the gear thickness at the group, and it is thick to will eventually get M group gear
Degree.
The flatness computational submodule of second image information processing module, can mention according to image procossing and marginal point
The gear thickness for taking submodule to obtain, subtracting minimum value using maximum value is required flatness.
The burr computational submodule of second image information processing module can calculate whether burr meets the requirements.First
To collected gear thickness gray level image, Threshold segmentation is carried out, dark part, i.e. gear region is extracted, then carries out one
Secondary opening operation, as shown in figure 33.Then it is once expanded, as shown in figure 34.Opening operation is subtracted using the big region after expansion
Zonule afterwards obtains the image of an image border band, as shown in figure 35.Followed by connection, according to length and orientation
The marginal belt of gear upper and lower surface is extracted, then region is gone to cut original image using the edge, obtains the hair of gear boundary
Image is pierced, after Threshold segmentation, as shown in figure 36.Followed by lateral encroaching, the connection between burr is disconnected, available one
Each and every one discrete burr, as shown in figure 37.Finally, its minimum circumscribed circle is sought to each discrete burr, to judge it
Whether meet the requirements.
The preservation submodule can save result to designated position after image information processing module terminates.Together
When, qualified and substandard product is retained separately, substandard product will indicate gear model and detection number, and indicating is which
One unqualified.Wherein the processing result of the first image information processing module and the second image information processing module is retained separately,
As shown in Figure 38 a and Figure 38 b.
The camera parameters calibration submodule is for demarcating the parameter of video camera.When calibration, a mark is first provided
Quasi- exemplar records each measurement result, and compare with three coordinates after the completion of two line-scan digital camera focusings, according to difference, fits
When the distance at fine tuning rotation of rotary table center to the first line-scan digital camera visual field boundary, the first line-scan digital camera line style pickup area and revolution
The angle of platform central axes, the pixel dimension of the first line-scan digital camera, the second line-scan digital camera line style acquisition seam and the angle of vertical direction,
Second line-scan digital camera pixel dimension.Until vision measurement result and three coordinates are coincide, click " exiting " terminates calibration.
The Parameter File setting and reading submodule, the tolerance that 11 projects of gear can be received in operation panel are defeated
Enter, wherein other point of the parameter of the first image information processing module and the second image information processing module input, such as Figure 39 a and figure
Shown in 39b.
The display 22 can be processed to as the result is shown after the first image information processing module terminates first
In user interface, including tooth top diameter, bottom of the tooth diameter, inner circle, inner circle out of roundness, outer round bounce, inside and outside circle concentricity, common normal,
Face song amount, as a result as shown in figure 40.
The display 22 can also be processed to as the result is shown after the second image information processing module terminates
In two user interfaces, including gear workpieces thickness, flatness, upper and lower surface burr, as a result as shown in figure 41.
The present invention can be using 7s as one gear of a cycle On-line rapid measurement, and intermediate station rotates one week shooting time
About 3s, analysis processing time about 4s.The present invention can simultaneously disposably detect 11 gear parameters, comprising: internal diameter, bottom of the tooth diameter,
Tooth top diameter, the bounce of internal diameter circle, tooth top diameter circle bounce, concentricity, common normal, face song amount (flank chamfering is wide), gear thickness, plane
Degree and burr.And the present invention can achieve the internal diameter of the same gear and the repetition detection accuracy of tooth top diameter
0.005mm, vision measurement result error compared with the measurement result of three coordinates are less than 0.02mm, meet required precision.
The above description is only a preferred embodiment of the present invention, is not intended to restrict the invention, for those skilled in the art
For, the invention can have various changes and changes.All any modifications made within the spirit and principles of the present invention are equal
Replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of gear burr detection method, comprising the following steps:
Step S1 acquires gear thickness gray level image；
Step S2 carries out Threshold segmentation to the gray level image, extracts gear region；
Step S3 carries out an opening operation to the area-of-interest, and is once expanded；
Step S4, the zonule after opening operation is subtracted using the big region after step S3 expansion, is obtained comprising marginal belt
Image；
Step S5, is connected to image obtained in the step S4, extracts gear upper and lower surface according to length and orientation
Marginal belt；
Step S6 cuts original-gray image using the marginal belt of the gear upper and lower surface, obtains the burr of gear boundary
Image；
Step S7 disconnects the connection between the burr of the spiculation visualization, obtains discrete burr；
Step S8 judges whether to meet the requirements to each discrete burr.
2. gear burr detection method according to claim 1, which is characterized in that the step S7 is to the spiculation visualization
Threshold segmentation, and lateral encroaching is carried out, to obtain discrete burr.
3. gear burr detection method according to claim 1, which is characterized in that the step S8 by ask each from
The minimum circumscribed circle of scattered burr judges whether it meets the requirements.
4. a kind of gear measuring system, including First look subsystem, the second vision subsystem, server and board subsystem,
Wherein
The First look subsystem can obtain the first image of tested gear along the axis direction of the gear；
Second vision subsystem can obtain the second image of tested gear along the axis direction perpendicular to the gear；
The server can be handled the first image and the second image, to measure the formal parameter of tested gear；
The board subsystem can send tested gear to the designated position for being suitble to acquisition image；
It is characterized in that, the server is examined using gear burr detection method according to any one of claim 1-3
Survey the burr of gear.
5. gear measuring system according to claim 4, which is characterized in that the First look subsystem includes: first
Line-scan digital camera, the first doubly telecentric camera lens, magic hand, line clear off source, the first bar shaped backlight.
6. gear measuring system according to claim 5, which is characterized in that the linear acquisition view of first line-scan digital camera
The board central axes of the board subsystem are passed through in open country；The distance quilt of first doubly telecentric distance of camera lens gear under test upper surface
It is set as its camera lens object distance；First line-scan digital camera is mounted on first level sliding block by upright slide block；
The first bar shaped backlight, which is placed on turntable, to be fixedly connected, and is arranged on turntable central axes, with the first linear array phase
The acquisition of machine, which is stitched, to be overlapped；
The line clears off source and is hold by magic hand, and is tilted a certain angle, and light is made just to fall in gear upper surface axis
On line；
The line clears off source and beats over from the direction for carrying chamfering pair, and below the first line-scan digital camera, chamfering is on right side, so line
Source is cleared off to squeeze into from left side.
7. gear measuring system according to claim 6, which is characterized in that second vision subsystem includes: second
Line-scan digital camera, the second doubly telecentric camera lens, the second bar shaped backlight.
8. gear measuring system according to claim 7, which is characterized in that second line-scan digital camera is fixedly mounted on
On two horizontal sliders, and pass through transparent turntable central axes together with second horizontal slider；
The second bar shaped backlight is fixedly mounted on board, and is arranged on the central axes of transparent turntable, face second
The visual field pickup area of the second line-scan digital camera is completely covered in the visual field of line-scan digital camera.
9. gear measuring system according to claim 8, which is characterized in that pacify the lower surface of the second bar shaped backlight
Holding position is lower than the table top of the board.
10. gear measuring system according to claim 9, which is characterized in that the board subsystem includes: PLC control
It is device, first servo motor, the second servo motor, third servo motor, the 4th servo motor, feeding pawl disk, blanking pawl disk, transparent
Turntable；
The first servo motor can drive the first level sliding block to slide in the horizontal direction；The second servo motor energy
The upright slide block is enough driven to slide in vertical direction；The third servo motor can drive second horizontal slider in water
Square to sliding；4th servo motor can drive the transparent turntable rotation；
When measurement, PLC controller can control feeding pawl disk and send gear to the transparent turntable, then start the 4th servo
Motor, the 4th servo motor is able to drive the transparent turntable and rotates, and after stabilization of speed, the PLC controller will start
Measuring signal is sent to the server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042298.5A CN109829897B (en) | 2019-01-17 | 2019-01-17 | Gear burr detection method and gear high-precision vision measurement system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042298.5A CN109829897B (en) | 2019-01-17 | 2019-01-17 | Gear burr detection method and gear high-precision vision measurement system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109829897A true CN109829897A (en) | 2019-05-31 |
| CN109829897B CN109829897B (en) | 2020-11-24 |
Family
ID=66861727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910042298.5A Expired - Fee Related CN109829897B (en) | 2019-01-17 | 2019-01-17 | Gear burr detection method and gear high-precision vision measurement system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109829897B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110102826A (en) * | 2019-06-17 | 2019-08-09 | 燕山大学 | A kind of submerged-arc welded (SAW) pipe lining inner burr removing apparatus and method |
| CN117522784A (en) * | 2023-10-23 | 2024-02-06 | 北京新光凯乐汽车冷成型件股份有限公司 | Gear part image detection method and system based on tooth distance segmentation compensation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1885317A (en) * | 2006-07-06 | 2006-12-27 | 上海交通大学 | Adaptive edge detection method based on morphology and information entropy |
| CN102737377A (en) * | 2012-05-14 | 2012-10-17 | 上海瑞伯德智能系统科技有限公司 | Improved method for extracting sub-pixel edge |
| CN104034263A (en) * | 2014-07-02 | 2014-09-10 | 北京理工大学 | Non-contact measurement method for sizes of forged pieces |
| US20150146918A1 (en) * | 2013-11-26 | 2015-05-28 | National Chung Cheng University | Video device for realtime pedaling frequency estimation |
| CN108723510A (en) * | 2017-04-13 | 2018-11-02 | 克林格伦贝格股份公司 | Burr remover and CNC Gear cutting machines including the burr remover |
-
2019
- 2019-01-17 CN CN201910042298.5A patent/CN109829897B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1885317A (en) * | 2006-07-06 | 2006-12-27 | 上海交通大学 | Adaptive edge detection method based on morphology and information entropy |
| CN102737377A (en) * | 2012-05-14 | 2012-10-17 | 上海瑞伯德智能系统科技有限公司 | Improved method for extracting sub-pixel edge |
| US20150146918A1 (en) * | 2013-11-26 | 2015-05-28 | National Chung Cheng University | Video device for realtime pedaling frequency estimation |
| CN104034263A (en) * | 2014-07-02 | 2014-09-10 | 北京理工大学 | Non-contact measurement method for sizes of forged pieces |
| CN108723510A (en) * | 2017-04-13 | 2018-11-02 | 克林格伦贝格股份公司 | Burr remover and CNC Gear cutting machines including the burr remover |
Non-Patent Citations (1)
| Title |
|---|
| 段振云，单紫薇，赵文辉，杨旭: "数字图像边缘的快速提取算法研究", 《组合机床与自动化加工技术》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110102826A (en) * | 2019-06-17 | 2019-08-09 | 燕山大学 | A kind of submerged-arc welded (SAW) pipe lining inner burr removing apparatus and method |
| CN117522784A (en) * | 2023-10-23 | 2024-02-06 | 北京新光凯乐汽车冷成型件股份有限公司 | Gear part image detection method and system based on tooth distance segmentation compensation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109829897B (en) | 2020-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109632007A (en) | A kind of edge point extracting method and gear high-precision vision measuring system | |
| CN102590218B (en) | Device and method for detecting micro defects on bright and clean surface of metal part based on machine vision | |
| CN108956645A (en) | A kind of the optical mirror slip defect detecting device and method of more vision systems | |
| CN111750805A (en) | Three-dimensional measuring device and method based on binocular camera imaging and structured light technology | |
| CN107869954B (en) | Binocular vision volume weight measurement system and implementation method thereof | |
| CN210604434U (en) | Online detection device for typical defects of curved glass of mobile phone | |
| CN110501347A (en) | A kind of rapid automatized Systems for optical inspection and method | |
| US20070115464A1 (en) | System and method for inspection of films | |
| CN113418933B (en) | Flying shooting visual imaging detection system and method for detecting large-size object | |
| CN113916156B (en) | High-speed high-precision three-dimensional detection system and method | |
| CN109829897A (en) | A kind of gear burr detection method and gear high-precision vision measuring system | |
| CN113259579A (en) | Image acquisition device, method and equipment | |
| CN108278972B (en) | Rotary member measuring device and background control method thereof | |
| US8390731B2 (en) | System and method for measuring a border of an image of an object | |
| TW201837477A (en) | Probe card detection method and system capable of automatically detecting needle offset, needle missing, needle tip oxidization, wear, damage and melt, etc. on the whole area of a probe card | |
| CN109521022A (en) | Touch screen defect detecting device based on the confocal camera of line | |
| CN112378797A (en) | Full-automatic indentation measuring device and method | |
| CN110657750B (en) | Detection system and method for passivation of cutting edge of cutter | |
| CN106200036B (en) | Lighting detection equipment | |
| JP2011145160A (en) | Device and method for multi-focus inspection | |
| CN112017181A (en) | Cylinder product surface detection method and related equipment thereof | |
| CN111122565A (en) | Device and method for simultaneously shooting multiple surfaces of object and detection system | |
| CN116256366A (en) | Chip defect detection method, detection system and storage medium | |
| CN112834517B (en) | Image detection method for bearing appearance | |
| CN212300252U (en) | Image measuring instrument capable of automatically adjusting position of sample to be measured |
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: 20201124 Termination date: 20220117 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |