CN104793494A - Method for automatic mounting and positioning of fatigue testing machine test piece and based on machine vision - Google Patents
Method for automatic mounting and positioning of fatigue testing machine test piece and based on machine vision Download PDFInfo
- Publication number
- CN104793494A CN104793494A CN201510171981.0A CN201510171981A CN104793494A CN 104793494 A CN104793494 A CN 104793494A CN 201510171981 A CN201510171981 A CN 201510171981A CN 104793494 A CN104793494 A CN 104793494A
- Authority
- CN
- China
- Prior art keywords
- hole
- test specimen
- circle
- fuzzy
- image
- 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
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
Abstract
The invention discloses a method for automatic mounting and positioning of a fatigue testing machine test piece and based on machine vision. The method includes: acquiring images of a lower clamp hole and a test piece hole through a CCD (charge coupled device); performing system calibration, image preprocessing, image analysis and image recognition on the images; detecting the test piece hole and the lower clamp hole through a circle fitting method respectively; calculating pixel distance between two round holes through a position measuring algorithm; converting the pixel distance into actual size through a system calibration value; taking hole difference data as a triggering signal for controlling a motor, controlling the motor to move for positioning so as to enable the lower clamp hole and the test piece hole to be aligned completely to realize automatic mounting and positioning of the fatigue testing machine test piece. The method realizes automatic mounting and positioning of the test piece, and is time saving, labor saving and high in accuracy.
Description
Technical field
The present invention relates to machine vision, image procossing and automation field, is a kind of fatigue tester test specimen Auto-mounting localization method based on machine vision concretely.
Background technology
Fatigure failure is the principal mode that component of machine lost efficacy, owing to still can not be studied by effective theoretical method completely at present, therefore by fatigue tester, fatigue crack propagation test is carried out to specific material, explore the rule of parts spreading fracture process, tool is of great significance.Owing to needing to carry out a large amount of fatigue crack propagation tests to the test specimen of difformity and different materials, so just need to install on fatigue tester frequently, dismantle test specimen.HF fatigue testing machine is generally that upper fixture (or lower clamp) maintains static, and (or upper fixture) can it moves up and down lower clamp by motor motion control.When test specimen is installed, first with pin, test specimen is fixed on upper fixture, then controls motor rotating forward or reverse to regulate lower clamp to move up and down, make lower clamp hole and test specimen hole centering, be fixed with pin again, complete the installation positioning work of test specimen, then carry out required fatigue experiment.
At present, most of HF fatigue testing machine, from installation sample to record data, all relies on manual operation to complete substantially.Location test specimen process is installed in traditional craft generally needs two people, hole difference between one of them people's Real Time Observation lower clamp hole and test specimen hole changes and the instruction assigned lower clamp rising or decline, and another person then needs to judge that motor rotates forward or reverses and control table motor in real time according to instruction.Detect by an unaided eye lower clamp hole and the basic centering in test specimen hole time, locate with pin, if the work repeated again above cannot be located, till test specimen can being fixed on lower clamp with pin, the accurate location and installation of time guarantee test specimen of two people's co-operatings at least 15 minutes.
Summary of the invention
In order to overcome existing fatigue tester test specimen, the wasting time and energy of locator meams, deficiency that precision is lower are installed, the invention provides a kind of fatigue tester test specimen Auto-mounting localization method based on machine vision, use machine vision and the hole between image processing techniques calculation testing piece hole and fixture hole poor, on the basis of position machine image acquisition and motor motion control in realization, controlling the lifting of worktable lower clamp makes fixture hole aim at test specimen hole, realizes test specimen Auto-mounting location.
In order to the technical scheme solving the problems of the technologies described above employing is:
A kind of fatigue tester test specimen Auto-mounting localization method based on machine vision, the image in lower clamp hole and test specimen hole is obtained by CCD, system calibrating is carried out to image, Image semantic classification, graphical analysis, image recognition operations, test specimen hole and fixture hole is detected respectively by the method for circle matching, calculate the pixel distance between two circular holes by position measurement algorithm again and be translated into physical size by system calibrating value, hole difference data is as the trigger pip of Electric Machine Control, control motor movement is located, make the entirely on the center of lower clamp hole and test specimen hole, realize the Auto-mounting location of fatigue tester test specimen.
Further, described system calibrating process is: first carry out Image semantic classification and graphical analysis to the image that image capturing system obtains, adopt the horizontal pixel distance of IMAQ Clamp Horizontal Max function measurement clamp edges again, just obtain the calibration value of system by fixture real standard distance divided by horizontal pixel distance;
Further, described Image semantic classification process is: first carry out medium filtering to image and carry out vertical direction coboundary Edge contrast again, then carries out histogram equalization and removes noise.
Further again, described image analysis process is: adopt histogram valley point to determine that the method for segmentation threshold carries out binaryzation to image, then the morphology operations of the post-etching that first expands to image.
Further, institute's position measurement algorithmic procedure is as follows:
1) coordinate system is set up to the crescent moon figure obtained after image procossing;
2) because lower clamp hole is on a Cylinder Surface, therefore can distortion be there is when it is taken pictures, so the size in the size of lower clamp hole on photo and test specimen hole is unequal, the closed region Modling model of the circular arc composition that two radiuses therefore formed lower clamp hole and test specimen hole differ, wherein the central coordinate of circle in test specimen hole is O
1: (X0, Y1) radius is r1, and the central coordinate of circle in lower clamp hole is O
2: (X0, Y2) radius is r2, and their center of circle coexists on the straight line of X=X0, then the lower half circle intersecting point coordinate of the circle in straight line X=X0 and test specimen hole and the circle in lower clamp hole is respectively P1 (X0, Y3) with P2 (X0, Y4), can obtain thus:
Y3=Y1-r1
Y4=Y2-r2
There is again P1 and P2 on the straight line of X=X0
Obtain hole gap from computing formula:
P1P2=Y3-Y4=y1-r1-(Y2-r2)=Y1-Y2-(r1-r2)
3) use the Find Circular Edge circle fitting function in NI IMAQ Vision kit, simulate test specimen hole circle and fixture hole circle respectively and calculate its central coordinate of circle and radius;
4) apply above described holes gap and obtain the pixel value of coordinate system mesopore difference and the value of P1P2 from computing formula, the actual relative position obtained between test specimen hole and fixture hole that is multiplied with system calibrating value.
Described step 3) in, choose crescent moon part as target search region, from left margin, vertical sweep from top to bottom, continue scanning when running into pixel and being 0, until run into stored in register when pixel is the point of 1 first, the point of these pixels from 0 to 1 is the point set the matched curve of test specimen hole circle, remove isolated point and matching is justified to it, thus obtain circle O
1central coordinate of circle and radius; The circle that the in like manner point set matching of pixel from 1 to 0 obtains is fixture hole circle O
2and obtain its central coordinate of circle and radius.
Control motor movement location by fuzzy PID control strategy, the process of fuzzy PID control strategy is as follows:
1) adopt fuzzy controller, system is three parameter: ratio k of the fuzzy controller on-line tuning PID that three dual input lists export
p, integration k
iwith differential k
d, PID controller is input as setpoint distance and the deviation e and the deviation variation rate ec that feed back distance, and output quantity is Δ k
p, Δ k
iwith Δ k
d, be used for respectively adjusting the parameter k of PID controller
p, k
i, k
dvalue;
2) motor fuzzy inference system
Input quantity e, ec and output quantity Δ k
p, Δ k
i, Δ k
dcorresponding Fuzzy Linguistic Variable is E, EC, Δ K
p, Δ K
iwith Δ K
dif its discrete-time fuzzy domain is all {-6 ,-5 ,-4 ,-3 ,-2 ,-1,0,1,2,3,4,5,6}.The basic domain of e is (-6,6), then quantizing factor is a
e=6/6=1, computing formula e (t)=r (t)-y (t), r (t) is default value, and y (t) is measured value; The basic domain of ec is (-0.3,0.3), then quantizing factor is a
ec=6/0.3=20; Computing formula ec (k)=e (k)-e (k-1); Δ k
pbasic domain is (-10,10), then quantizing factor is a
kp=6/10=0.6; Δ k
ibasic domain is (-0.2,0.2), then quantizing factor is a
ki=6/0.2=30; Δ k
dbasic domain is (-1,1), then quantizing factor is a
kp=6/1=6; Above [-6,6] between, the variable of change is divided into 7 grades as required: negative large (NB), negative in (NM), negative little (NS), zero (Z0), just little (PS), center (PM), honest (PB), each grade is as a fuzzy variable, and a corresponding fuzzy subset or membership function.
Adopt two inputs single output Mamdani MAX-MIN inference method, de-fuzzy then adopts the gravity model appoach that precision is higher, and concrete fuzzy inference rule is as follows:
Rule 1:IF A
iaND B
itHEN C
i(i=1,2,3 ... n).
Beneficial effect of the present invention is: realize test specimen Auto-mounting location, laborsaving, precision is higher.
Accompanying drawing illustrates:
Fig. 1 is fatigue tester test specimen Auto-mounting positioning system structure figure of the present invention.
Fig. 2 is fatigue tester test specimen Auto-mounting positioning system algorithm flow chart of the present invention.
Fig. 3 is fatigue tester test specimen Auto-mounting positioning system process flow diagram of the present invention.
Fig. 4 is test specimen position measurement illustraton of model of the present invention.
Fig. 5 is fuzzy-adaptation PID control structural drawing.
Embodiment
Below in conjunction with accompanying drawing, the present invention will be further described.
With reference to Fig. 1 ~ Fig. 5, a kind of fatigue tester test specimen Auto-mounting localization method based on machine vision, the image in lower clamp hole and test specimen hole is obtained by CCD, system calibrating is carried out to image, Image semantic classification, graphical analysis, image recognition operations, test specimen hole and fixture hole is detected respectively by the method for circle matching, calculate the pixel distance between two circular holes by position measurement algorithm again and be translated into physical size by system calibrating value, hole difference data is as the trigger pip of Electric Machine Control, control motor movement is located, make the entirely on the center of lower clamp hole and test specimen hole, realize the Auto-mounting location of fatigue tester test specimen.
Further, described system calibrating process is: first carry out Image semantic classification and graphical analysis to the image that image capturing system obtains, adopt the horizontal pixel distance of IMAQ Clamp Horizontal Max function measurement clamp edges again, just obtain the calibration value of system by fixture real standard distance divided by horizontal pixel distance;
Further, described Image semantic classification process is: first carry out medium filtering to image and carry out vertical direction coboundary Edge contrast again, then carries out histogram equalization and removes noise.
Further again, described image analysis process is: adopt histogram valley point to determine that the method for segmentation threshold carries out binaryzation to image, then the morphology operations of the post-etching that first expands to image.
Further, institute's position measurement algorithmic procedure is as follows::
1) coordinate system is set up to the crescent moon figure obtained after image procossing;
2) with reference to Fig. 4, because lower clamp hole is on a Cylinder Surface, therefore can distortion be there is when it is taken pictures, so the size in the size of lower clamp hole on photo and test specimen hole is unequal, the closed region Modling model of the circular arc composition that two radiuses therefore formed lower clamp hole and test specimen hole differ, wherein the central coordinate of circle in test specimen hole is O
1: (X0, Y1) radius is r1, and the central coordinate of circle in lower clamp hole is O
2: (X0, Y2) radius is r2, and their center of circle coexists on the straight line of X=X0, then the lower half circle intersecting point coordinate of the circle in straight line X=X0 and test specimen hole and the circle in lower clamp hole is respectively P1 (X0, Y3) with P2 (X0, Y4), can obtain thus:
Y3=Y1-r1
Y4=Y2-r2
There is again P1 and P2 on the straight line of X=X0
Obtain hole gap from computing formula:
P1P2=Y3-Y4=y1-r1-(Y2-r2)=Y1-Y2-(r1-r2)
3) use the Find Circular Edge circle fitting function in NI IMAQ Vision kit, simulate test specimen hole circle and fixture hole circle respectively and calculate its central coordinate of circle and radius;
4) apply above described holes gap and obtain the pixel value of coordinate system mesopore difference and the value of P1P2 from computing formula, the actual relative position obtained between test specimen hole and fixture hole that is multiplied with system calibrating value.
Described step 3) in, choose crescent moon part as target search region, from left margin, vertical sweep from top to bottom, continue scanning when running into pixel and being 0, until run into stored in register when pixel is the point of 1 first, the point of these pixels from 0 to 1 is the point set the matched curve of test specimen hole circle, remove isolated point and matching is justified to it, thus obtain circle O
1central coordinate of circle and radius; The circle that the in like manner point set matching of pixel from 1 to 0 obtains is fixture hole circle O
2and obtain its central coordinate of circle and radius.
Control motor movement location by fuzzy PID control strategy, the process of fuzzy PID control strategy is as follows:
1) adopt fuzzy controller, system is three parameter: ratio k of the fuzzy controller on-line tuning PID that three dual input lists export
p, integration k
iwith differential k
d, PID controller is input as setpoint distance and the deviation e and the deviation variation rate ec that feed back distance, and output quantity is Δ k
p, Δ k
iwith Δ k
d, be used for respectively adjusting the parameter k of PID controller
p, k
i, k
dvalue;
2) motor fuzzy inference system
Input quantity e, ec and output quantity Δ k
p, Δ k
i, Δ k
dcorresponding Fuzzy Linguistic Variable is E, EC, Δ K
p, Δ K
iwith Δ K
dif its discrete-time fuzzy domain is all {-6 ,-5 ,-4 ,-3 ,-2 ,-1,0,1,2,3,4,5,6}.The basic domain of e is (-6,6), then quantizing factor is a
e=6/6=1, computing formula e (t)=r (t)-y (t), r (t) is default value, and y (t) is measured value; The basic domain of ec is (-0.3,0.3), then quantizing factor is a
ec=6/0.3=20; Computing formula ec (k)=e (k)-e (k-1); Δ k
pbasic domain is (-10,10), then quantizing factor is a
kp=6/10=0.6; Δ k
ibasic domain is (-0.2,0.2), then quantizing factor is a
ki=6/0.2=30; Δ k
dbasic domain is (-1,1), then quantizing factor is a
kp=6/1=6.Above [-6,6] between, the variable of change is divided into 7 grades as required: negative large (NB), negative in (NM), negative little (NS), zero (Z0), just little (PS), center (PM), honest (PB), each grade is as a fuzzy variable, and a corresponding fuzzy subset or membership function, the degree of membership of each fuzzy subset is as shown in table 1.
Table 1 e, ec and Δ k
p, Δ k
i, Δ k
ddegree of membership rule list
When different e and ec, the controlled process of native system is to Δ k
p, Δ k
i, Δ k
dself-tuning System require be:
(1), when e is comparatively large, in order to there be preferably tracking performance fast, and avoids the change of the moment because starting time error e may cause differential supersaturation greatly and make control action exceed tolerance band, larger k should be got
pless k
d.Simultaneously in order to prevent saturation integral, avoid system responses to occur larger overshoot, reply integral action is limited, and now gets k
i=0.
(2), when e and ec is median size, there is less overshoot, k for making system responses
p, k
iand k
dall can not take large values, less k should be got
ivalue, k
pand k
dvalue size moderate, to ensure the response speed of system, wherein k
dvalue larger on the impact of the response speed of system.
(3), when e is less, for making system have good steady-state behaviour, larger k should be got
iand k
d, simultaneously for avoiding system to occur vibration near setting value, and consider the interference free performance of system, k
dshould select according to ec, its principle is: when ec is less, k
dvalue desirable larger, be usually taken as median size; When ec is larger, k
dvalue should get smaller.
(4), the physical significance of change of error amount ec is the fast slow rate showing change of error, and the value of ec is larger, k
pvalue less, k
ivalue larger.According to PID Rule adjusting, in conjunction with the Volume control of real material fatigue tester object in the design, establish respectively for k
p, k
iand k
dthe fuzzy control rule table of 3 parameter tunings, 3 parameters of PID controller respectively have 49 fuzzy relations in the present system, and table 2 is Δ k
pfuzzy reasoning table, table 3 is Δ k
ifuzzy reasoning table, table 4 is Δ k
dfuzzy reasoning table, as follows:
Table 2
Table 3
Table 4
Adopt such as formula
with
shown two inputs are single exports Mamdani MAX-MIN inference methods, de-fuzzy then adopt gravity model appoach that precision is higher such as formula
shown in, concrete fuzzy inference rule is as follows:
Rule 1:IF A
1aND B
1tHEN C
1
Rule 2:IF A
2aND B
2tHEN C
2
……
Rule n:IF A
naND B
ntHEN C
n
Input x
0aND y
0conclusion z
0
By prerequisite " x
0aND y
0" and various fuzzy rule " A
iaND B
itHEN C
i(i=1,2 ... n) " the reasoning results can be obtained: the wall scroll rule of activation exports the fuzzy subset C of fuzzy variable
i' degree of membership such as formula
the degree of membership of the fuzzy subset C ' of total activation vagueness of regulations variable such as formula
output quantity z after defuzzification
0such as formula
In the present embodiment, fatigure failure is the principal mode that component of machine lost efficacy, and fatigue tester carries out to specific material the Main Means that fatigue crack propagation test is research parts spreading fracture.Owing to needing to carry out a large amount of fatigue crack propagation tests to the test specimen of difformity and different materials, positioning system all the more important is installed in therefore robotization, intelligentized test specimen.When discussing fatigue sample Auto-mounting positioning system, modeling and analysis being carried out to the hole difference in lower clamp hole and test specimen hole, motor has been controlled to fuzzy PID algorithm and also analyzes.Contrived experiment LabVIEW based on virtual instrument technique is that development platform has write relative program, has done deep research from theoretical and experiment two aspects to fatigue tester test specimen Auto-mounting localization method.
The present invention utilizes computer digital image treatment technology and LabVIEW to be developing instrument, image capturing system and computer system for fatigue tester test specimen Auto-mounting Location System Design, image capturing system mainly gathers image to test specimen and lower clamp, is transferred in computer system and carries out image procossing.Image semantic classification in main use digital image processing techniques, comprises image filtering denoising, histogram equalization and edge sharpening; Graphical analysis, comprises Iamge Segmentation, morphology processing; Image recognition, detects test specimen hole and fixture hole respectively by circle approximating method, then calculates by position measurement algorithm the poor pixel distance be converted into physical size by demarcation of portalling.Physical size data are transferred to the control program of the motor in computer system, wherein control strategy is fuzzy-adaptation PID control, carries out Electric Machine Control adjustment working table movement thus drives lower clamp to move thereupon, finally completing the centering in lower clamp hole and test specimen hole.
The present invention establishes the test specimen Auto-mounting positioning system of a set of Vision Builder for Automated Inspection and complete set, well solves traditional-handwork and installs the problem of locating test specimen and wasting time and energy.Fig. 1 is the Organization Chart of fatigue tester test specimen Auto-mounting positioning system, comprises HF fatigue testing machine, camera, annular light source, camera lens, image pick-up card and computer system, and these are hardware devices required for the present invention.Native system also comprises other modules used by computer system, comprises image input module, system calibrating module, image processing module and automatic control module.Camera wherein with annular light source is used for gathering test specimen hole and fixture hole pattern picture, and image is transferred to computer system by image input module, computer system utilizes image processing module to carry out image procossing, and utilize automatic control module to control working table movement thus drive lower clamp motion, complete the centering in fixture hole and test specimen hole.
Fig. 2 is fatigue tester test specimen Auto-mounting positioning system process flow diagram, it comprises the following steps: the image transmitting first obtained by image capturing system is to computer system, relevant treatment carried out to image and adopts the horizontal pixel distance of IMAQ Clamp Horizontal Max function measurement clamp edges, just obtaining system calibrating value by fixture real standard distance divided by horizontal pixel distance; Adopt first medium filtering to carry out the sharpening of vertical direction coboundary again, then the method for carrying out histogram equalization removal noise carry out Image semantic classification; Histogram valley point is adopted to determine that the method for segmentation threshold is carried out Iamge Segmentation and obtained binary image; Adopt and first to image, closed edge operation is carried out to the image morphology operations corroded again that expands; Use the Find Circular Edge circle fitting function in NI IMAQ Vision kit, matching is justified to test specimen hole and fixture hole and obtains central coordinate of circle and radius; Measure the pixel distance of two circle minimum points and change into actual range according to system calibrating value; Adopt fuzzy PID algorithm to control motor and make working table movement, finally complete the centering in test specimen hole and fixture hole.
Example: the hardware device such as HX-A30-D70-W-24V annular light source, the PENTAX-B2514D camera lens of Japanese Bin get company, the PCI-1410 image pick-up card of America NI company of the PLG-100 HF fatigue testing machine adopting Tianshui city Red Hill factory to produce, SONY XC-HR70 camera, German WEILANG company, utilize image processing techniques and Fuzzy PID Control Technique, carry out Study system process flow diagram as shown in Figure 3 for CT test specimen and PLG-100 HF fatigue testing machine lower clamp centering.
Step 1. gathers image.Opening program carries out system initialization, opens annular light source and makes it be radiated at test specimen and chucking surface uniformly, and adjustment camera position can obtain more details, carries out image acquisition to it.
Step 2. serial communication.Serial communication RS232 is used between the host computer of server end and field controller in this experiment.First, server sends checking command to controller, returns the instruction of transmission from controller, and whether inspection communication is normal.The data that both sides send are hexadecimal data, and each data length is 1 byte.The instruction that both sides send is made up of the hexadecimal data of more than 3 or 3.Serial communication protocol is formulated as follows:
Epigynous computer section:
Communication checking command: FF 03 FA
If return data FF is FA, for communication is normal.
Static load loading command: FF 05 0D X 02
FF 05 0D X 01
FF 05 0D X 00
Wherein X is static loading speed, is the sexadecimal number of 1 byte.Last byte represents three kinds of different operating modes of motor, is respectively rotating forward, reversion, stops.
Query State instruction: FF 03 0C
Terminate vibration instruction: FF 04 09 02
Closed communication instruction: FF 04 03 07
Slave computer controller part:
What slave computer sent is status information instruction, and its instruction length is 1015 bytes.1st, 2 bytes are respectively FA and 08, and wherein FA represents packet header, and 08 represents check code; 13rd byte representation status information, that only uses wherein is low 4; What the 16 to 1015 byte representation this section of hexadecimal data represented is Wave data, and overall length is 1000 bytes, and every 2 bytes are a wave numerics.Conversion from hexadecimal data to wave numerics is carried out according to demarcation file.
Step 3. system calibrating.The image obtained image capturing system carries out medium filtering and then carries out edge sharpening in vertical direction, carry out again histogram equalization remove noise and utilize histogram valley point to determine the method for segmentation threshold carries out binaryzation to image, the morphology operations of the post-etching that then image first expanded; The horizontal pixel distance finally adopting IMAQ Clamp Horizontal Max function measurement clamp edges is 373.247 pixels, just obtains system calibrating value 0.06751mm/ pixel with fixture real standard distance 25.2mm divided by horizontal pixel distance.
Step 4. position measurement and Electric Machine Control.The test specimen obtain image capturing system and fixture image adopt the image processing method in step 3 system calibrating to process, use the Find Circular Edge circle fitting function in NI IMAQVision kit, simulate fixture hole and two, test specimen hole circle and obtain central coordinate of circle and radius, the pixel distance being obtained hole difference by distance calculating method is multiplied by the actual range that system calibrating value 0.06751mm/ pixel obtains hole difference; To the feed back input of the poor actual range in the hole obtained as fuzzy control motor algorithm be detected, the initial k of setting Fuzzy PID
p, k
i, k
dvalue and desired distance are 0, and desired distance and the distance detected are subtracted each other the deviation e` of deviation e, e and previous chronomere subtracts each other to obtain deviation variation rate ec, thus call Δ k
p, Δ k
i, Δ k
dquestion blank obtains exact value, is added respectively is self-adjusting current k with corresponding setting value
p, k
i, k
dvalue, controls motor through serial communication accordingly.
Step 5. data and image storage.Store the calibration value that system calibrating obtains, next time can directly call these data and not need to repeat the operation of step 3.
Claims (8)
1. the fatigue tester test specimen Auto-mounting localization method based on machine vision, it is characterized in that: the image being obtained lower clamp hole and test specimen hole by CCD, system calibrating is carried out to image, Image semantic classification, graphical analysis, image recognition operations, test specimen hole and fixture hole is detected respectively by the method for circle matching, calculate the pixel distance between two circular holes by position measurement algorithm again and be translated into physical size by system calibrating value, hole difference data is as the trigger pip of Electric Machine Control, control motor movement is located, make the entirely on the center of lower clamp hole and test specimen hole, realize the Auto-mounting location of fatigue tester test specimen.
2. as claimed in claim 1 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: described system calibrating process is: first Image semantic classification and graphical analysis are carried out to the image that image capturing system obtains, adopt the horizontal pixel distance of IMAQ Clamp Horizontal Max function measurement clamp edges again, just obtain the calibration value of system by fixture real standard distance divided by horizontal pixel distance.
3. as claimed in claim 1 or 2 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: described Image semantic classification process is: first medium filtering is carried out to image and carry out vertical direction coboundary Edge contrast again, then carry out histogram equalization and remove noise.
4. as claimed in claim 1 or 2 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: described image analysis process is: adopt histogram valley point to determine that the method for segmentation threshold carries out binaryzation to image, then the morphology operations of post-etching that image is first expanded.
5., as claimed in claim 1 or 2 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: institute's position measurement algorithmic procedure is as follows::
1) coordinate system is set up to the crescent moon figure obtained after image procossing;
2) because lower clamp hole is on a Cylinder Surface, therefore can distortion be there is when it is taken pictures, so the size in the size of lower clamp hole on photo and test specimen hole is unequal, the closed region Modling model of the circular arc composition that two radiuses therefore formed lower clamp hole and test specimen hole differ, wherein the central coordinate of circle in test specimen hole is O
1: (X0, Y1) radius is r1, and the central coordinate of circle in lower clamp hole is O
2: (X0, Y2) radius is r2, and their center of circle coexists on the straight line of X=X0, then the lower half circle intersecting point coordinate of the circle in straight line X=X0 and test specimen hole and the circle in lower clamp hole is respectively P1 (X0, Y3) with P2 (X0, Y4), can obtain thus:
Y3=Y1-r1
Y4=Y2-r2
There is again P1 and P2 on the straight line of X=X0
Obtain hole gap from computing formula:
P1P2=Y3-Y4=y1-r1-(Y2-r2)=Y1-Y2-(r1-r2)
3) use the Find Circular Edge circle fitting function in NI IMAQ Vision kit, simulate test specimen hole circle and fixture hole circle respectively and calculate its central coordinate of circle and radius;
4) apply above described holes gap and obtain the pixel value of coordinate system mesopore difference and the value of P1P2 from computing formula, the actual relative position obtained between test specimen hole and fixture hole that is multiplied with system calibrating value.
6. as claimed in claim 5 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: described step 3) in, choose crescent moon part as target search region, from left margin, vertical sweep from top to bottom, continues scanning when running into pixel and being 0, until run into first when pixel is the point of 1 stored in register, the point of these pixels from 0 to 1 is the point set the matched curve of test specimen hole circle, removes isolated point and justifies matching to it, thus obtains circle O
1central coordinate of circle and radius; The circle that the in like manner point set matching of pixel from 1 to 0 obtains is fixture hole circle O
2and obtain its central coordinate of circle and radius.
7. as claimed in claim 1 or 2 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: control motor movement location by fuzzy PID control strategy, the process of fuzzy PID control strategy is as follows:
1) adopt fuzzy controller, system is three parameter: ratio k of the fuzzy controller on-line tuning PID that three dual input lists export
p, integration k
iwith differential k
d, PID controller is input as setpoint distance and the deviation e and the deviation variation rate ec that feed back distance, and output quantity is Δ k
p, Δ k
iwith Δ k
d, be used for respectively adjusting the parameter k of PID controller
p, k
i, k
dvalue;
2) motor fuzzy inference system
Input quantity e, ec and output quantity Δ k
p, Δ k
i, Δ k
dcorresponding Fuzzy Linguistic Variable is E, EC, Δ K
p, Δ K
iwith Δ K
dif its discrete-time fuzzy domain is all that {-6 ,-5 ,-4 ,-3 ,-2 ,-1, the basic domain of 0,1,2,3,4,5,6}, e is (-6,6), then quantizing factor is a
e=6/6=1, computing formula e (t)=r (t)-y (t), r (t) is default value, and y (t) is measured value; The basic domain of ec is (-0.3,0.3), then quantizing factor is a
ec=6/0.3=20; Computing formula ec (k)=e (k)-e (k-1); Δ k
pbasic domain is (-10,10), then quantizing factor is a
kp=6/10=0.6; Δ k
ibasic domain is (-0.2,0.2), then quantizing factor is a
ki=6/0.2=30; Δ k
dbasic domain is (-1,1), then quantizing factor is a
kp=6/1=6, above [-6,6] between, the variable of change is divided into 7 grades as required: negative large (NB), negative in (NM), negative little (NS), zero (Z0), just little (PS), center (PM), honest (PB), each grade is as a fuzzy variable, and a corresponding fuzzy subset or membership function;
Adopt two inputs single output Mamdani MAX-MIN inference method, de-fuzzy then adopts the gravity model appoach that precision is higher, and fuzzy inference rule is as follows:
Rule 1:IF A
iaND B
itHEN C
ii=1,2,3 ... n.
8., as claimed in claim 7 based on the fatigue tester test specimen Auto-mounting localization method of machine vision, it is characterized in that: in described motor fuzzy inference system, adopt such as formula
with
shown two inputs are single exports Mamdani MAX-MIN inference methods, de-fuzzy then adopt gravity model appoach that precision is higher such as formula
shown in, concrete fuzzy inference rule is as follows:
Rule 1:IF A
1aND B
1tHEN C
1
Rule 2:IF A
2aND B
2tHEN C
2
……
Rule n:IF A
naND B
ntHEN C
n
Input x
0aNDy
0conclusion z
0
By prerequisite " x
0aNDy
0" and various fuzzy rule " A
iaND B
itHEN C
i(i=1,2 ... n) " the reasoning results is obtained: the wall scroll rule of activation exports the fuzzy subset C ' of fuzzy variable
idegree of membership such as formula
the degree of membership of the fuzzy subset C ' of total activation vagueness of regulations variable such as formula
output quantity z after defuzzification
0such as formula
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510171981.0A CN104793494B (en) | 2015-04-13 | 2015-04-13 | A kind of automatic installation and locating method of fatigue tester test specimen based on machine vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510171981.0A CN104793494B (en) | 2015-04-13 | 2015-04-13 | A kind of automatic installation and locating method of fatigue tester test specimen based on machine vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104793494A true CN104793494A (en) | 2015-07-22 |
CN104793494B CN104793494B (en) | 2018-04-20 |
Family
ID=53558397
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510171981.0A Active CN104793494B (en) | 2015-04-13 | 2015-04-13 | A kind of automatic installation and locating method of fatigue tester test specimen based on machine vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104793494B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110015836A (en) * | 2018-01-09 | 2019-07-16 | 蓝思科技股份有限公司 | A kind of adjusting method and system of circular product concentricity |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102692347A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Camera automatic regulating image acquisition device and method for fatigue crack propagation test |
CN102692188A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Dynamic crack length measurement method for machine vision fatigue crack propagation test |
-
2015
- 2015-04-13 CN CN201510171981.0A patent/CN104793494B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102692347A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Camera automatic regulating image acquisition device and method for fatigue crack propagation test |
CN102692188A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Dynamic crack length measurement method for machine vision fatigue crack propagation test |
Non-Patent Citations (5)
Title |
---|
廖常俊等: "一种用于零件定位***的亚像素边缘检测方法", 《冶金自动化》 * |
沈姗姗: "谐振式疲劳试验***工作载荷自适应模糊控制技术研究", 《中国优秀硕士学位论文全文数据库》 * |
王晓东等: "基于机器视觉的微小型零件测量与装配控制", 《哈尔滨工程大学学报》 * |
范壮等: "基于IMAQ的尺寸、形状检测方法研究", 《电子工业专用设备》 * |
黄帅等: "一种基于机器视觉的自动定位方法及实验", 《测控技术》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110015836A (en) * | 2018-01-09 | 2019-07-16 | 蓝思科技股份有限公司 | A kind of adjusting method and system of circular product concentricity |
Also Published As
Publication number | Publication date |
---|---|
CN104793494B (en) | 2018-04-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20240129368A1 (en) | Server device, learned model providing program, learned model providing method, and learned model providing system | |
US12020415B2 (en) | Method for monitoring manufacture of assembly units | |
KR101917006B1 (en) | Semiconductor Manufacturing Yield Prediction System and Method based on Machine Learning | |
US10254750B2 (en) | Machining machine system which determines acceptance/rejection of workpieces | |
CN110109441B (en) | Laser inertial measurement unit fault prediction method and system | |
CN105014678A (en) | Robot hand-eye calibration method based on laser range finding | |
CN104019843B (en) | Automobile meter pointer zero-bit automatic correction device based on image procossing and metaprogramming techniques and method | |
CN109991918A (en) | Parallel control method based on the sampling of multicycle difference with the twin technology of number | |
CN110906875A (en) | Visual processing method for aperture measurement | |
CN113449703B (en) | Quality control method and device for environment online monitoring data, storage medium and equipment | |
CN115683235B (en) | Industrial robot servo motor vibration fault detection method and device | |
CN107167172A (en) | A kind of on-line monitoring method of bus type automobile digital instrument pointer functionality | |
CN108764278A (en) | A kind of the self study industrial intelligent detecting system and method for view-based access control model | |
CN104793494A (en) | Method for automatic mounting and positioning of fatigue testing machine test piece and based on machine vision | |
CN113532341A (en) | Method and device for determining a measurement strategy for measuring a measurement object, and program | |
CN112434600A (en) | Unmanned aerial vehicle inspection method and system | |
CN116168019B (en) | Power grid fault detection method and system based on machine vision technology | |
CN112034048A (en) | Beam structure crack positioning method based on multiple frequency response function estimation | |
US20200202178A1 (en) | Automatic visual data generation for object training and evaluation | |
US20180264570A1 (en) | Method and device for automated machining of gearwheel components | |
CN108733859B (en) | Early warning method for deformation of support plate in heat treatment process of steam generator | |
Cao | Robot Learning for Object Handling in an Unstructured Environment | |
CN117392182B (en) | Film pasting precision detection method, device, equipment and storage medium | |
CN117549317B (en) | Robot grabbing and positioning method and system | |
Baicu et al. | Control of the yeast growth process using an image processing-based transducer |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |