CN110008955A - A kind of automotive brake pads face character coining quality inspection method - Google Patents
A kind of automotive brake pads face character coining quality inspection method Download PDFInfo
- Publication number
- CN110008955A CN110008955A CN201910255460.1A CN201910255460A CN110008955A CN 110008955 A CN110008955 A CN 110008955A CN 201910255460 A CN201910255460 A CN 201910255460A CN 110008955 A CN110008955 A CN 110008955A
- Authority
- CN
- China
- Prior art keywords
- image
- character
- character zone
- region
- round orifice
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Abstract
The invention discloses a kind of automotive brake pads face characters to imprint quality inspection method.This method is based on OTSU algorithm and carries out image segmentation；To extract two positioning round orifice regions calculate its center of circle image coordinate；The positioning of two sides character zone is realized according to relative positional relationship between the two location hole centers of circle and two sides character zone and is extracted；Binaryzation is carried out using adaptive threshold fuzziness in extracted two sides character zone；To the brake picture for being used as matching template, matching template is obtained based on sciagraphy；To brake block image to be tested, matching template is used to carry out template matching in extracted two sides character zone one by one by character imprint sequence from top to bottom, and be filled to best match region, until completing the matching of all templates；Realize that disc surface character imprints quality inspection according to black picture element points in result images.Quality inspection is imprinted using the achievable automotive brake pads face character of the present invention, provides a kind of new approaches for the zone location of vision-based detection object.
Description
Technical field
The present invention relates to a kind of quality inspection method, especially a kind of automotive brake pads face character imprints quality inspection side
Method.
Background technique
With the continuous improvement of people's living standards, automobile is on continuous as its ownership of the vehicles used in everyday
The trend risen, the traffic safety of automobile also become focus concerned by people.Automotive brake pads running as an influence
The important spare part of safety, quality directly affects the personal safety of vehicle operator and passenger, therefore, to its matter
It is particularly important that amount carries out unified management management.The pressed characters for representing brake block model or batch are often printed on brake block,
It is to realize the important step of quality of brake pad management that quality inspection is carried out to it.Automotive brake pads face character imprints quality inspection
It mainly include that character zone positioning and character imprint quality inspection two parts.
At present in machine vision realize character zone positioning means mainly pass through gradient, character texture characteristics or
The methods of the special math-morphological features (such as license plate) that character zone has.These methods are required to entire image all
It performs corresponding processing, carries out judgement further according to corresponding feature to extract character zone, although can get preferable
Locating effect, but positioning time is longer, the unfixed image in localization region suitable for processing localization region rule relatively.So
And this kind of for automotive brake pads character zone edge is irregular, position vision relatively-stationary for components itself
Test object needs a kind of highly efficient, accurate character zone localization method.
After carrying out character zone positioning, need to carry out character coining quality inspection.Current trial method usually elder generation is one by one
Character recognition is carried out, then judges whether character coining is correct according to character identification result, and then realizes character coining quality inspection
It tests.This method is based on single character recognition algorithm, and the accuracy rate of single character recognition will will have a direct impact on face character coining matter
Measure the accuracy and reliability examined.However, such as alphabetical " I " and digital " 1 ", letter " o " and digital " 0 " etc. shapes very phase
As character, be easier to identification mistake occur, and character carries out knowledge method for distinguishing one by one, due to needing to characters all in character repertoire
It is compared, and character quantity is more in character repertoire, causes the character coining quality inspection time longer, therefore, it is necessary to one kind more
For easy, efficient and accurate whole character code method, to ensure character coining quality inspection accuracy and efficiency.
The automatic inspection of automotive brake pads face character coining quality can be achieved in the present invention, can be the area of vision-based detection object
Domain positioning provides a kind of effective method, imprints quality inspection for character and provides a kind of new thinking.
Summary of the invention
The purpose of the present invention is to provide a kind of automotive brake pads face characters to imprint quality inspection method, to realize automobile
The automatic inspection of the automatic positioning of disc surface character zone and character coining quality.
The technical solution adopted by the present invention is that:
The present invention includes the following steps:
1.1) character zone positions: carrying out brake block or so using a kind of automotive brake pads character zone automatic positioning method
The automatic positioning and extraction of two sides character zone；Automotive brake pads binary map is extracted by OTSU automatic threshold segmentation algorithm first
As B, noise reduction is carried out to automotive brake pads bianry image based on gaussian filtering and corrosion expansion algorithm；Then it is mentioned by BLOB analysis
Two positioning round orifice of brake block is taken, to all pixels point application least square method fitting circle formulas Extraction brake block in circular hole region
Two positioning round orifice center of circle image coordinate；Respective angles are carried out based on two positioning round orifice circle center line connectings and image abscissa angle
Image rotation correction, keep two positioning round orifice circle center line connecting of brake block parallel with image abscissa；Finally by brake block two sides
Fixed relative positional relationship calculates the boundary rectangle characteristic point for determining two sides character zone between character zone and two positioning round orifice
The image coordinate of (for determining the point of two sides character zone boundary rectangle position in image coordinate system) realizes brake block two sides
The automatic positioning of character zone；
1.2) character zone extracts: the two sides character zone boundary rectangle position and preset determined according to step 1.1)
Boundary rectangle size m row × n column, extracts left and right sides character zone image Z from original image O；
1.3) character zone image noise reduction: the left and right two that step 1.2) is obtained using the corrosion and expansion algorithm of grayscale image
Side character zone image Z carries out noise reduction process, obtains the noise-reduced image L of left and right sides character zoneZ；
1.4) character zone image binaryzation: the noise-reduced image that step 1.3) is obtained using auto-thresholding algorithm
LZBinary conversion treatment is carried out, binary image D is obtained；Auto-thresholding algorithm is by calculating separately its point to each pixel
It cuts threshold value and realizes that image segmentation, its implementation are as follows: seeking the weighted mean of grey scale pixel value in each pixel N neighborhood；It should
The weighted mean of pixel subtracts constant C and obtains the binarization segmentation threshold value T of the pixel；If current image processed is as subsequent
Character imprints the template image of quality inspection, then enters step upon completion of this step 1.5)；If current image processed be to
1.8) checking image then enters step upon completion of this step；
1.5) upright projection: two sides character zone binary image D resulting to step 1.4) carries out upright projection respectively,
Obtain vertical projective histogram TX, vertical projective histogram TXHeight to indicate that binary image D correspond in each column gray value be 0
Pixel number；In vertical projective histogram TXHeight be 0 corresponding with non-zero separation in binary image D at respective column
Image cutting is carried out, vertical projection histogram T is retainedXHighly non-zero corresponding region, obtains upright projection segmented image SX；
1.6) floor projection: to upright projection segmented image SXFloor projection is carried out, horizontal projective histogram T is obtainedY, water
Flat projection histogram TYWidth means upright projection segmented image SXThe pixel number that gray value is 0 in corresponding each row；It is thrown in level
Shadow histogram TYWidth be 0 corresponding with non-zero separation in upright projection segmented image SXImage is carried out at middle corresponding row to cut
It cuts, retention level projection histogram TYThe non-zero corresponding region of width, obtains Character mother plate image；
1.7) matching template saves: saving to the Character mother plate image that step 1.6) obtains, as successive character pressure
Print the matching template of quality inspection；
1.8) character imprints quality inspection: imprinting correction judgement method by character and realizes automotive brake pads face character
Imprint quality inspection；This method by template matching algorithm respectively to the left and right sides character zone in binary image D, according to
Character imprint sequence from top to bottom, is matched using corresponding matching template one by one and obtains best match region and (existed
In character zone with the highest subregion of matching template coincidence factor), it is by area filling that each matching template is corresponding best
Whether matching area is filled with white, be more than finally that threshold value realizes braking automobile according to black picture element number remaining in image after filling
Piece face character imprints quality inspection, if remaining black picture element number is less than threshold value in image, character coining is up-to-standard, no
Then, character coining is off quality.
A kind of automotive brake pads character zone automatic positioning method, implementation method are as follows in the step 1.1):
2.1) image segmentation: using the OTSU automatic threshold image segmentation algorithm based on gray level image to automotive brake pads into
Row image segmentation obtains bianry image B；Non- brake panel region and two positioning round orifice region of brake block are background in bianry image B,
Brake panel region is prospect；
2.2) Gaussian Blur image noise reduction: is carried out to bianry image B；Corrosion expansion behaviour is carried out on the basis of Gaussian Blur
Make, to eliminate the noise spot in image, obtains filtering image LB；
2.3) two positioning round orifice extracted region: being analyzed by BLOB, extracts filtering image LBIn maximum background area, and
It is negated, converts foreground area for the maximum background area, remaining background area is two setting circle porose area of brake block
Domain, to realize two positioning round orifice extracted region of brake block；Read the image for the pixel that gray value is 0 in two positioning round orifice regions
Transverse and longitudinal coordinate, since two positioning round orifice regions are divided at left and right sides of column brake block, the corresponding abscissa in image coordinate system
Range is different, therefore, when positioning round orifice area pixel abscissa columns half total less than image, it is believed that the pixel belongs to left side
Positioning round orifice region；When positioning round orifice area pixel abscissa columns half total greater than image, it is believed that the pixel belongs to right side
Positioning round orifice region；
2.4) two location hole central coordinate of circle and distance computation: to all pixels point application least square method in circular hole region
Fitting circle formula seeks the central coordinate of circle value (X of two positioning round orifice of left and rightL, YL)、(XR, YR), and two center of circle spacing L are sought, such as formula (1)
It is shown:
2.5) brake block tilt angle calculate: by two positioning round orifice circle center line connecting of brake block and image level reference axis institute at
Inclination angle of the angle value α as brake block is calculated as shown in formula (2):
2.6) image rotation: by image around the midpoint A of two positioning round orifice circle center line connectings, rotation-α angle, so that two positioning
Circular hole circle center line connecting is parallel with image level reference axis；Postrotational left and right sides positioning round orifice central coordinate of circle is respectively (Xl,
Yl)、(Xr, Yr), it calculates as shown in formula (3):
2.7) character zone boundary rectangle characteristic point coordinate calculates: the right vertices of definition left side character zone boundary rectangle
For the characteristic point of the character zone, the left upper apex of right side character zone boundary rectangle is the characteristic point of the character zone；Due to
Left and right character zone is fixed with the left and right positioning round orifice center of circle in the relative positional relationship of disc surface, accordingly, can be according to left and right
Positioning round orifice central coordinate of circle calculates the image coordinate (x of left and right sides character zone characteristic pointl, yl)、(xr, yr), such as formula (4)
It is shown:
In formula: a, b are preset ratio coefficient, are constant；
2.8) length and width of character zone boundary rectangle calculate (vertical direction is length, and horizontal direction is width): due to brake block
The size of surface imprint character region is fixed, therefore, can be according to the two positioning round orifice centers of circle of the resulting fixation of step 2.4)
Spacing L determines the long m and width n of character zone boundary rectangle in proportion, as shown in formula (5):
M=cL, n=dL (5)
In formula: c, d are preset ratio coefficient, are constant；
2.9) character zone positions: according to left and right sides character zone characteristic point coordinate (xl, yl)、(xr, yr) and character
The long m of region boundary rectangle and width n calculates remaining 3 apex coordinate of left and right sides character zone boundary rectangle, completes left and right two
The positioning of side character zone, as shown in formula (6):
In formula: (xllu, yllu)、(xlld, ylld)、(xlrd, ylrd)、(xrru, yrru)、(xrrd, yrrd)、(xrld, yrld) respectively
For the upper left angle point of left side character zone, lower-left angle point, bottom right angle point and the upper right angle point of right side character zone, the lower right corner
Point, lower-left angle point the plane of delineation in coordinate.
Character imprints correction judgement method in the step 1.8), and implementation method is as follows:
3.1) it template matching: by template matching algorithm respectively to the left and right sides character zone in binary image D, presses
According to character imprint sequence (can also be adjusted according to the actual situation) from top to bottom, carried out one by one using corresponding matching template
Matching, and obtain best match region (i.e. in character zone with the highest subregion of matching template coincidence factor)；
3.2) best match region is filled: image completion is carried out to best match region, i.e., it will be in the best match region
Whole pixels be filled with white, that is, gray value is all adjusted to 255；Filling region size is big in matching template
On the basis of small, on it, under, 3 pixels are widened on left and right four direction respectively, to ensure that check character is covered completely
Lid, avoids character from being repeated inspection；
3.3) return step 3.1), until completing the matching of all characters in the character zone of the left and right sides, it is transferred to step
3.4):
3.4) the sum of all pixels S that gray value is 0 in the character zone of the statistics left and right sidesN；
3.5) the sum of all pixels S for being 0 according to left and right sides character zone gray valueN, judge whether character coining quality closes
Lattice: if SN< TN(TNFor preset threshold), then character coining is correct, and character coining is up-to-standard；Otherwise, character imprints mistake, word
Symbol coining is off quality.
The invention has the advantages that: the present invention realizes the coining quality inspection of automotive brake pads face character, together
Shi Kewei realizes that the automatic positioning of vision-based detection object region provides a kind of new method, can print quality inspection for character printing
A kind of new approaches are provided.
Detailed description of the invention
Fig. 1 is automotive brake pads character coining Quality Inspection System composition schematic diagram.
Fig. 2 is automotive brake pads character coining quality inspection method flow chart.
Fig. 3 is brake block master drawing.
Fig. 4 is brake block master drawing image segmentation result.
Fig. 5 is brake block master drawing noise reduction result.
Fig. 6 is the brake block setting circle hole pattern extracted after inversion operation.
Fig. 7 is image rotation correction schematic diagram.
Fig. 8 is that character zone boundary rectangle characteristic point coordinate calculates schematic diagram.
Fig. 9 is the character zone figure extracted after brake block character zone positions.
Figure 10 is character zone binary map.
Figure 11 is two sides character zone vertical projection histogram.
Figure 12 is two sides character zone horizontal projective histogram.
Figure 13 is the character match Prototype drawing extracted.
In Fig. 1: 1 is camera, and 2 be computer, and 3 imprint quality inspection software for automotive brake pads face character, 4,5,7,8
It is automotive brake pads for 4 strip sources, 6,9 be optical lens.
Specific embodiment
Present invention will be further explained below with reference to the attached drawings and examples.
As Fig. 1 illustrates a specific embodiment of automotive brake pads face character positioning system.Including automotive brake pads
Character zone positioning system, image received device,.The pros that lighting system is constituted using the red bar light 4,5,7,8 of 4 30w
Shape lighting system.Image received device uses black and white camera 1, black and white camera model JHSM500Bf-E, and maximum resolution is
2592 × 1944, band caching, 1/2.5 " CMOS.Camera lens 9 is mega pixel camera lens, model JHLD1108-5M.Computer 2 is
7 operating system of WIN, image processing algorithm programmed environment are Microsoft Visual Studio 2010.Automotive brake pads 5
For excellent power (JURID) brake block, model 200FF.
The positioning of automotive brake pads character zone is implemented as follows:
Lighting system is adjusted, so that the illumination intensity of image acquisition region is moderate and uniform, the shelves of red bar light 4,5,7,8
Position is respectively 6.2,6.8,6.8,6.6；Electronic image is converted by the optical imagery received by camera lens 9 and black and white camera 1
It is input in computer 2；Automotive brake pads character zone positioning software 3 in computer 2 realizes oneself of brake block character zone
Dynamic positioning.
As shown in Fig. 2, in automotive brake pads character zone positioning software 5 automotive brake pads character zone automatic positioning side
Method is implemented as follows:
1.1) character zone positions: carrying out brake block or so using a kind of automotive brake pads character zone automatic positioning method
The automatic positioning and extraction of two sides character zone；Automotive brake pads binary map is extracted by OTSU automatic threshold segmentation algorithm first
As B, noise reduction is carried out to automotive brake pads bianry image based on gaussian filtering and corrosion expansion algorithm；Then it is mentioned by BLOB analysis
Two positioning round orifice of brake block is taken, to all pixels point application least square method fitting circle formulas Extraction brake block in circular hole region
Two positioning round orifice center of circle image coordinate；Respective angles are carried out based on two positioning round orifice circle center line connectings and image abscissa angle
Image rotation correction, keep two positioning round orifice circle center line connecting of brake block parallel with image abscissa；Finally by brake block two sides
Fixed relative positional relationship calculates the boundary rectangle characteristic point for determining two sides character zone between character zone and two positioning round orifice
The image coordinate of (for determining the point of two sides character zone boundary rectangle position in image coordinate system) realizes brake block two sides
The automatic positioning of character zone；
1.2) character zone extracts: the two sides character zone boundary rectangle position and preset determined according to step 1.1)
Boundary rectangle size m row × n column, extracts two sides rectangular broken line area in the left and right sides character zone image Z, i.e. Fig. 8 from original image O
Domain, as shown in Figure 9.
1.3) character zone image noise reduction: the left and right two that step 1.2) is obtained using the corrosion and expansion algorithm of grayscale image
Side character zone image Z carries out noise reduction process, obtains the noise-reduced image L of left and right sides character zoneZ；
Left side character zone image uses size for 10 × 10 rectangle convolution kernel, and right side character zone image is using size
For 6 × 6 rectangle convolution kernel；
1.4) character zone image binaryzation: the noise-reduced image that step 1.3) is obtained using auto-thresholding algorithm
LZBinary conversion treatment is carried out, binary image D is obtained；Auto-thresholding algorithm is by calculating separately its point to each pixel
It cuts threshold value and realizes that image segmentation, its implementation are as follows: seeking the weighted mean of grey scale pixel value in each pixel N neighborhood；It should
The weighted mean of pixel subtracts constant C and obtains the binarization segmentation threshold value T of the pixel；The present invention is used as using the mean value of neighborhood to be added
The result of weight average.Shown in the mathematical expression of adaptive threshold fuzziness such as formula (8):
In formula, F (i, j) is the pixel gray value of the i-th row jth column in N × N neighborhood of the pixel.It is used in the present invention
The neighborhood that size is 31 × 31, being subtracted constant is 20.The character of part of brake block is shallower due to imprinting, so that in image
The gray difference of brake block is smaller near the gray scale and character of character, also allows for this parts of images and carries out adaptive threshold fuzziness
When will appear over-segmentation, i.e. partial character can also be taken as background segment and fall.Therefore, this parts of images is devised accordingly
Method of adjustment, to after segmentation image carry out gray value be 0 pixel carry out quantity statistics, if gray value be 0 pixel
Points are less than 1300, then it is assumed that over-segmentation, at this point, by being subtracted in adaptive threshold fuzziness after constant is adjusted to 15, then it is right
Image before carrying out adaptivenon-uniform sampling carries out adaptive threshold fuzziness again.Segmentation result is as shown in Figure 10.If currently being located
Managing image is the template image that quality inspection is imprinted as successive character, then enters step upon completion of this step 1.5)；If working as
Preceding image processed is image to be tested, then enters step upon completion of this step 1.8)；
1.5) upright projection: two sides character zone binary image D resulting to step 1.4) carries out upright projection respectively,
Obtain vertical projective histogram TX, as shown in figure 11, vertical projective histogram TXHeight indicate binary image D correspond to each column
The pixel number that middle gray value is 0；In vertical projective histogram TXHeight be 0 with non-zero separation (in Figure 11 white area with
Separation of the black region on axis of abscissas) the progress image cutting at respective column in binary image D is corresponded to, it is vertical to retain
Projection histogram TXHighly non-zero corresponding region (black region in Figure 11), obtains upright projection segmented image SX；
1.6) floor projection: to upright projection segmented image SXFloor projection is carried out, horizontal projective histogram T is obtainedY, such as
Shown in Figure 12, horizontal projective histogram TYWidth means upright projection segmented image SXThe pixel that gray value is 0 in corresponding each row
Number；In horizontal projective histogram TYWidth be that 0 (white area and black region are in ordinate in Figure 12 with non-zero separation
On separation) it is corresponding in upright projection segmented image SXImage cutting, retention level projection histogram T are carried out at middle corresponding rowY
The non-zero corresponding region (black region in Figure 12) of width, obtains Character mother plate image；
1.7) matching template saves: saving to the Character mother plate image that step 1.6) obtains, as successive character pressure
The matching template of quality inspection is printed, as shown in figure 13；
1.8) character imprints quality inspection: imprinting correction judgement method by character and realizes automotive brake pads face character
Imprint quality inspection；This method by template matching algorithm respectively to the left and right sides character zone in binary image D, according to
Character imprint sequence from top to bottom, is matched using corresponding matching template one by one and obtains best match region and (existed
In character zone with the highest subregion of matching template coincidence factor), it is by area filling that each matching template is corresponding best
Whether matching area is filled with white, be more than finally that threshold value realizes braking automobile according to black picture element number remaining in image after filling
Piece face character imprints quality inspection, if remaining black picture element number is less than threshold value in image, character coining is up-to-standard, no
Then, character coining is off quality.
A kind of automotive brake pads character zone automatic positioning method, implementation method are as follows in the step 1.1):
2.1) image segmentation: using the OTSU automatic threshold image segmentation algorithm based on gray level image to automotive brake pads into
Row image segmentation obtains bianry image B；Non- brake panel region and two positioning round orifice region of brake block are background in bianry image B,
Brake panel region is prospect, as shown in figure 4, background area and two positioning round orifice region of brake block are black in figure, section of braking
Domain is white；
2.2) image noise reduction: the rectangle Gaussian filter that Gaussian Blur uses 9 × 9 is carried out to bianry image B, is laterally filtered
Coefficient and longitudinal filter factor are 2..3 are carried out using different different size of rectangle kernels on the basis of Gaussian Blur
Wheel corrosion expansive working (first corrode and expand afterwards, first expand post-etching, first corrode and expand afterwards), used rectangle kernel size difference
It is 15 × 15,20 × 20,10 × 10, further eliminates the noise spot in image, obtain filtering image LB, as shown in Figure 5；
2.3) two positioning round orifice extracted region: by the BLOB analysis model of 3 × 3 sizes, filtering image L is extractedBIn
Maximum background area, and it is negated, foreground area is converted by the maximum background area, remaining background area is to brake
Two positioning round orifice region of piece, to realize two positioning round orifice extracted region of brake block, (black and white area distribution in figure as shown in Figure 6
With actual conditions on the contrary, convenient for observation)；The image transverse and longitudinal coordinate for reading the pixel that gray value is 0 in two positioning round orifice regions, by
Divide at left and right sides of column brake block in two positioning round orifice regions, corresponding abscissa range is different in image coordinate system, because
This, when positioning round orifice area pixel abscissa columns half total less than image, it is believed that the pixel belongs to left side setting circle porose area
Domain；When positioning round orifice area pixel abscissa columns half total greater than image, it is believed that the pixel belongs to right positioner circular hole area
Domain；
2.4) two location hole central coordinate of circle and distance computation: to all pixels point application least square method in circular hole region
Fitting circle formula seeks the central coordinate of circle value (X of two positioning round orifice of left and rightL, YL)、(XR, YR), and two center of circle spacing L are sought, such as formula (1)
It is shown:
2.5) brake block tilt angle calculate: by two positioning round orifice circle center line connecting of brake block and image level reference axis institute at
Inclination angle of the angle value α as brake block is calculated as shown in formula (2):
2.6) image rotation: by image around the midpoint A of two positioning round orifice circle center line connectings, rotation-α angle, so that two positioning
Circular hole circle center line connecting is parallel with image level reference axis；Postrotational left and right sides positioning round orifice central coordinate of circle is respectively (Xl,
Yl)、(Xr, Yr), it calculates as shown in formula (3):
2.7) character zone boundary rectangle characteristic point coordinate calculates: the right vertices of definition left side character zone boundary rectangle
For the characteristic point of the character zone, the left upper apex of right side character zone boundary rectangle is the characteristic point of the character zone；Due to
Left and right character zone is fixed with the left and right positioning round orifice center of circle in the relative positional relationship of disc surface, accordingly, can be according to left and right
Positioning round orifice central coordinate of circle calculates the image coordinate (x of left and right sides character zone characteristic pointl, yl)、(xr, yr), such as formula (4)
It is shown:
In formula: a, b are preset ratio coefficient, are constant, tests determined a=0.495, b=0.29；；
2.8) length and width of character zone boundary rectangle calculate (vertical direction is length, and horizontal direction is width): due to brake block
The size of surface imprint character region is fixed, therefore, can be according to the two positioning round orifice centers of circle of the resulting fixation of step 2.4)
Spacing L determines the long m and width n of character zone boundary rectangle in proportion, as shown in formula (5):
M=cL, n=dL (5)
In formula: c, d are preset ratio coefficient, are constant, tests determined c=0.19, d=0.079；
2.9) character zone positioning: according to left and right sides character zone characteristic point coordinate (xl, yl)、(xr, yr) and character
The long m of region boundary rectangle and width n calculates remaining 3 apex coordinate of left and right sides character zone boundary rectangle, completes left and right two
The positioning of side character zone, as shown in formula (6):
In formula: (xllu, yllu)、(xlld, ylld)、(xlrd, ylrd)、(xrru, yrru)、(xrrd, yrrd)、(xrld, yrld) respectively
For the upper left angle point of left side character zone, lower-left angle point, bottom right angle point and the upper right angle point of right side character zone, the lower right corner
Point, lower-left angle point the plane of delineation in coordinate.
Character imprints correction judgement method in the step 1.8), and implementation method is as follows:
3.1) template matching: by calling the template matching function (matchTemplate) in the library OpenCV respectively to two
Left and right sides character zone in value image D is matched using corresponding matching template one by one in a certain order,
And obtain best match region (i.e. in character zone with the highest subregion of matching template coincidence factor)；It is matched for carrying out
Characteristic parameter is CV_TM_SQDIFF, i.e. the difference of two squares, shown in mathematical expression such as formula (7)；
In formula, T (x ', y ') indicates the gray value of xth ' row y ' column pixel in template image, and I (x+x ', y+y ') table
Show the gray value of xth+x ' row y+y ' column pixel in images to be recognized, and template image upper left angle point is in images to be recognized
Be located at (x+1)th row y+1 arrange.
Since with other characters error hiding may occur for left side character " I " and " J " in matching process, in left side word
Template matching is carried out according to the sequence of " D ", " R ", " U ", " J ", " I " in the template matching process of symbol.The mould of right side character
Plate matching process then carries out template matching according to the sequence of (i.e. " F ", " F ", " 0 ", " 0 ", " 2 ") from top to down；
3.2) best match region is filled: image completion is carried out to best match region, i.e., it will be in the best match region
Whole pixels be filled with white, that is, gray value is all adjusted to 255；Filling region size is big in matching template
On the basis of small, on it, under, 3 pixels are widened on left and right four direction respectively, to ensure that check character is covered completely
Lid, avoids character from being repeated inspection；
3.3) return step 3.1), until completing the matching of all characters in the character zone of the left and right sides, it is transferred to step
3.4):
3.4) the sum of all pixels S that gray value is 0 in the character zone of the statistics left and right sidesN；
3.5) the sum of all pixels S for being 0 according to left and right sides character zone gray valueN, judge whether character coining quality closes
Lattice: if SN< TN(TNFor preset threshold), then character coining is correct, and character coining is up-to-standard；Otherwise, character imprints mistake, word
Symbol coining is off quality.It is tests determined, preset threshold TN=60.
Face character is carried out to 216 width to be measured brake picture according to above-mentioned steps and method and imprints quality inspection, word
The accuracy rate for according with zone location is 100%, and the accuracy rate of character code is 100%, and inspection rate is 1.692s/ width, Neng Gougao
It imitates and is accurately realized the coining quality inspection of automotive brake pads face character.
Claims (3)
1. a kind of automotive brake pads face character imprints quality inspection method, which comprises the steps of:
1.1) character zone positions: being carried out at left and right sides of brake block using a kind of automotive brake pads character zone automatic positioning method
The automatic positioning and extraction of character zone；Automotive brake pads bianry image B is extracted by OTSU automatic threshold segmentation algorithm first,
Noise reduction is carried out to automotive brake pads bianry image based on gaussian filtering and corrosion expansion algorithm；Then it analyzes to extract by BLOB and stop
Two positioning round orifice of vehicle piece, it is fixed to all pixels point application least square method fitting circle formulas Extraction brake block two in circular hole region
Circle of position hole center of circle image coordinate；The figure of respective angles is carried out based on two positioning round orifice circle center line connectings and image abscissa angle
As rotation correction, keep two positioning round orifice circle center line connecting of brake block parallel with image abscissa；Finally by brake block two sides character
Fixed relative positional relationship, which calculates, between region and two positioning round orifice determines that the boundary rectangle characteristic point of two sides character zone (is used for
Determine the point of two sides character zone boundary rectangle position in image coordinate system) image coordinate, realize brake block two sides character
The automatic positioning in region；
1.2) character zone extracts: the two sides character zone boundary rectangle position and preset external determined according to step 1.1)
Rectangle size m row × n column, extracts left and right sides character zone image Z from original image O；
1.3) character zone image noise reduction: the left and right sides word that step 1.2) is obtained using the corrosion and expansion algorithm of grayscale image
It accords with area image Z and carries out noise reduction process, obtain the noise-reduced image L of left and right sides character zoneZ；
1.4) character zone image binaryzation: the noise-reduced image L that step 1.3) is obtained using auto-thresholding algorithmZInto
Row binary conversion treatment obtains binary image D；Auto-thresholding algorithm divides threshold by calculating separately it to each pixel
Value realizes that image segmentation, its implementation are as follows: seeking the weighted mean of grey scale pixel value in each pixel N neighborhood；By the pixel
Weighted mean subtract constant C and obtain the binarization segmentation threshold value T of the pixel；If current image processed is as successive character
1.5) template image for imprinting quality inspection, then enter step upon completion of this step；If current image processed is to be tested
1.8) image then enters step upon completion of this step；
1.5) upright projection: two sides character zone binary image D resulting to step 1.4) carries out upright projection respectively, obtains
Vertical projective histogram TX, vertical projective histogram TXHeight indicate that binary image D correspond to gray value in each column as 0 picture
Prime number；In vertical projective histogram TXHeight be 0 it is corresponding with non-zero separation in binary image D at respective column progress
Image cutting, retains vertical projection histogram TXHighly non-zero corresponding region, obtains upright projection segmented image SX；
1.6) floor projection: to upright projection segmented image SXFloor projection is carried out, horizontal projective histogram T is obtainedY, level throwing
Shadow histogram TYWidth means upright projection segmented image SXThe pixel number that gray value is 0 in corresponding each row；It is straight in floor projection
Side figure TYWidth be 0 corresponding with non-zero separation in upright projection segmented image SXImage cutting is carried out at middle corresponding row, is protected
Stay horizontal projective histogram TYThe non-zero corresponding region of width, obtains Character mother plate image；
1.7) matching template saves: saving to the Character mother plate image that step 1.6) obtains, imprints matter as successive character
Measure the matching template of inspection；
1.8) character imprints quality inspection: imprinting correction judgement method by character and realizes automotive brake pads face character coining
Quality inspection；This method by template matching algorithm respectively to the left and right sides character zone in binary image D, according to from upper
Character imprint sequence down is matched using corresponding matching template one by one and obtains best match region (i.e. in character
In region with the highest subregion of matching template coincidence factor), by area filling by the corresponding best match of each matching template
Whether area filling is white, be more than finally that threshold value realizes automotive brake pads table according to black picture element number remaining in image after filling
Face character imprints quality inspection, if remaining black picture element number is less than threshold value in image, character coining is up-to-standard, otherwise, word
Symbol coining is off quality.
2. a kind of automotive brake pads face character as described in claim 1 imprints quality inspection method, which is characterized in that described
A kind of automotive brake pads character zone automatic positioning method, comprise the following steps:
2.1) figure image segmentation: is carried out to automotive brake pads using the OTSU automatic threshold image segmentation algorithm based on gray level image
As segmentation, bianry image B is obtained；Non- brake panel region and two positioning round orifice region of brake block are background, brake in bianry image B
Panel region is prospect；
2.2) Gaussian Blur image noise reduction: is carried out to bianry image B；Corrosion expansive working is carried out on the basis of Gaussian Blur,
To eliminate the noise spot in image, filtering image L is obtainedB；
2.3) two positioning round orifice extracted region: being analyzed by BLOB, extracts filtering image LBIn maximum background area, and to it
It negating, converts foreground area for the maximum background area, remaining background area is two positioning round orifice region of brake block, from
And realize two positioning round orifice extracted region of brake block；Read the image transverse and longitudinal for the pixel that gray value is 0 in two positioning round orifice regions
Coordinate, since two positioning round orifice regions are divided at left and right sides of column brake block, the corresponding abscissa range in image coordinate system
Difference, therefore, when positioning round orifice area pixel abscissa columns half total less than image, it is believed that the pixel belongs to left side positioning
Circular hole region；When positioning round orifice area pixel abscissa columns half total greater than image, it is believed that the pixel belongs to right positioner
Circular hole region；
2.4) two location hole central coordinate of circle and distance computation: all pixels point application least square method in circular hole region is fitted
Circle formula seeks the central coordinate of circle value (X of two positioning round orifice of left and rightL, YL)、(XR, YR), and two center of circle spacing L are sought, as shown in formula (1):
2.5) brake block tilt angle calculates: by two positioning round orifice circle center line connecting of brake block and image level reference axis angle
Inclination angle of the value α as brake block is calculated as shown in formula (2):
2.6) image rotation: by image around the midpoint A of two positioning round orifice circle center line connectings, rotation-α angle, so that two positioning round orifice
Circle center line connecting is parallel with image level reference axis；Postrotational left and right sides positioning round orifice central coordinate of circle is respectively (Xl, Yl)、
(Xr, Yr), it calculates as shown in formula (3):
2.7) character zone boundary rectangle characteristic point coordinate calculates: the right vertices of definition left side character zone boundary rectangle are should
The characteristic point of character zone, the left upper apex of right side character zone boundary rectangle are the characteristic point of the character zone；Due to left and right
Character zone is fixed with the left and right positioning round orifice center of circle in the relative positional relationship of disc surface, accordingly, can be positioned according to left and right
Circular hole central coordinate of circle calculates the image coordinate (x of left and right sides character zone characteristic pointl, yl)、(xr, yr), as shown in formula (4):
In formula: a, b are preset ratio coefficient, are constant；
2.8) length and width of character zone boundary rectangle calculate (vertical direction is length, and horizontal direction is width): due to disc surface
The size of pressed characters region is fixed, therefore, can be according to two positioning round orifice center of circle spacing L of the resulting fixation of step 2.4)
The long m and width n for determining character zone boundary rectangle in proportion, as shown in formula (5):
M=cL, n=dL (5)
In formula: c, d are preset ratio coefficient, are constant；
2.9) character zone positions: according to left and right sides character zone characteristic point coordinate (xl, yl)、(xr, yr) and character zone
The long m of boundary rectangle and width n calculates remaining 3 apex coordinate of left and right sides character zone boundary rectangle, completes left and right sides word
Zone location is accorded with, as shown in formula (6):
In formula: (xllu, yllu)、(xlld, ylld)、(xlrd, ylrd)、(xrru, yrru)、(xrrd, yrrd)、(xrld, yrld) it is respectively a left side
The upper left angle point of side character zone, lower-left angle point, bottom right angle point and right side character zone upper right angle point, bottom right angle point, a left side
Coordinate in the plane of delineation of lower angle point.
3. a kind of automotive brake pads face character as described in claim 1 imprints quality inspection method, which is characterized in that described
Character imprint correction judgement method, comprise the following steps:
3.1) template matching: by template matching algorithm respectively to the left and right sides character zone in binary image D, according to from
On character imprint sequence (can also be adjusted according to the actual situation) down, carried out using corresponding matching template one by one
Match, and obtains best match region (i.e. in character zone with the highest subregion of matching template coincidence factor)；
3.2) best match region is filled: image completion is carried out to best match region, i.e., it will be complete in the best match region
Portion's pixel is filled with white, that is, gray value is all adjusted to 255；Filling region size is in matching template size
On the basis of, on it, under, 3 pixels are widened on left and right four direction respectively, to ensure that check character is completely covered, is kept away
Exempt from character and is repeated inspection；
3.3) return step 3.1), until completing the matching of all characters in the character zone of the left and right sides, it is transferred to step 3.4):
3.4) the sum of all pixels S that gray value is 0 in the character zone of the statistics left and right sidesN；
3.5) the sum of all pixels S for being 0 according to left and right sides character zone gray valueN, judge whether character coining quality is qualified: if
SN< TN(TNFor preset threshold), then character coining is correct, and character coining is up-to-standard；Otherwise, character imprints mistake, character pressure
It prints off quality.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910255460.1A CN110008955B (en) | 2019-04-01 | 2019-04-01 | Method for testing character imprinting quality of surface of automobile brake pad |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910255460.1A CN110008955B (en) | 2019-04-01 | 2019-04-01 | Method for testing character imprinting quality of surface of automobile brake pad |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110008955A true CN110008955A (en) | 2019-07-12 |
| CN110008955B CN110008955B (en) | 2020-12-15 |
Family
ID=67169167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910255460.1A Expired - Fee Related CN110008955B (en) | 2019-04-01 | 2019-04-01 | Method for testing character imprinting quality of surface of automobile brake pad |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110008955B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110412052A (en) * | 2019-08-12 | 2019-11-05 | 东莞市艾尔玛科技有限公司 | A kind of curved surface hot padding quality determining method and system |
| CN111192273A (en) * | 2019-12-27 | 2020-05-22 | 西北工业大学 | Digital shot blasting coverage rate measuring method based on computer vision technology |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5321995A (en) * | 1993-01-11 | 1994-06-21 | Chrysler Corporation | Parking brake pedal pad |
| CN101334837A (en) * | 2008-07-31 | 2008-12-31 | 重庆大学 | Multi-method integrated license plate image positioning method |
| CN101706873A (en) * | 2009-11-27 | 2010-05-12 | 东软集团股份有限公司 | Identification method and device of digital-class limitation marking |
| CN101739566A (en) * | 2009-12-04 | 2010-06-16 | 重庆大学 | Self-adapting projection template method-based automobile plate positioning method |
| CN102651069A (en) * | 2012-03-31 | 2012-08-29 | 重庆大学 | Contour-based local invariant region detection method |
| CN102939477A (en) * | 2010-06-04 | 2013-02-20 | 丰田自动车株式会社 | Break device, friction couple and break pad for brake system |
| CN103324930A (en) * | 2013-06-28 | 2013-09-25 | 浙江大学苏州工业技术研究院 | License plate character segmentation method based on grey level histogram binaryzation |
| US8851246B2 (en) * | 2009-07-27 | 2014-10-07 | Knorr-Bremse Systeme Fuer Nutzfahrzeuge Gmbh | Brake pad for a disc brake |
| CN105258727A (en) * | 2015-10-28 | 2016-01-20 | 中国计量学院 | Brake backplane automatic detection device based on machine vision |
| CN105528789A (en) * | 2015-12-08 | 2016-04-27 | 深圳市恒科通多维视觉有限公司 | Robot vision positioning method and device, and visual calibration method and device |
| CN105701492A (en) * | 2014-11-25 | 2016-06-22 | 宁波舜宇光电信息有限公司 | Machine vision identification system and implementation method thereof |
| CN106384113A (en) * | 2016-11-07 | 2017-02-08 | 湖南源信光电科技有限公司 | Odd-and-even-line license plate character dividing method based on projection and template matching |
| CN106558089A (en) * | 2015-09-21 | 2017-04-05 | 展讯通信（上海）有限公司 | Image depth method for drafting and device |
| CN107084666A (en) * | 2017-05-10 | 2017-08-22 | 中国计量大学 | Brake block dimension synthesis detection method based on machine vision |
| CN206435500U (en) * | 2017-01-16 | 2017-08-25 | 中国计量大学 | Brake block product quality automatic detection device |
| CN109035225A (en) * | 2018-07-12 | 2018-12-18 | 中国计量大学 | A kind of automotive brake pads presentation quality inspection illumination system layout quality evaluating method |
-
2019
- 2019-04-01 CN CN201910255460.1A patent/CN110008955B/en not_active Expired - Fee Related
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5321995A (en) * | 1993-01-11 | 1994-06-21 | Chrysler Corporation | Parking brake pedal pad |
| CN101334837A (en) * | 2008-07-31 | 2008-12-31 | 重庆大学 | Multi-method integrated license plate image positioning method |
| US8851246B2 (en) * | 2009-07-27 | 2014-10-07 | Knorr-Bremse Systeme Fuer Nutzfahrzeuge Gmbh | Brake pad for a disc brake |
| CN101706873A (en) * | 2009-11-27 | 2010-05-12 | 东软集团股份有限公司 | Identification method and device of digital-class limitation marking |
| CN101739566A (en) * | 2009-12-04 | 2010-06-16 | 重庆大学 | Self-adapting projection template method-based automobile plate positioning method |
| CN102939477A (en) * | 2010-06-04 | 2013-02-20 | 丰田自动车株式会社 | Break device, friction couple and break pad for brake system |
| CN102651069A (en) * | 2012-03-31 | 2012-08-29 | 重庆大学 | Contour-based local invariant region detection method |
| CN103324930A (en) * | 2013-06-28 | 2013-09-25 | 浙江大学苏州工业技术研究院 | License plate character segmentation method based on grey level histogram binaryzation |
| CN105701492A (en) * | 2014-11-25 | 2016-06-22 | 宁波舜宇光电信息有限公司 | Machine vision identification system and implementation method thereof |
| CN106558089A (en) * | 2015-09-21 | 2017-04-05 | 展讯通信（上海）有限公司 | Image depth method for drafting and device |
| CN105258727A (en) * | 2015-10-28 | 2016-01-20 | 中国计量学院 | Brake backplane automatic detection device based on machine vision |
| CN105528789A (en) * | 2015-12-08 | 2016-04-27 | 深圳市恒科通多维视觉有限公司 | Robot vision positioning method and device, and visual calibration method and device |
| CN106384113A (en) * | 2016-11-07 | 2017-02-08 | 湖南源信光电科技有限公司 | Odd-and-even-line license plate character dividing method based on projection and template matching |
| CN206435500U (en) * | 2017-01-16 | 2017-08-25 | 中国计量大学 | Brake block product quality automatic detection device |
| CN107084666A (en) * | 2017-05-10 | 2017-08-22 | 中国计量大学 | Brake block dimension synthesis detection method based on machine vision |
| CN109035225A (en) * | 2018-07-12 | 2018-12-18 | 中国计量大学 | A kind of automotive brake pads presentation quality inspection illumination system layout quality evaluating method |
Non-Patent Citations (2)
| Title |
|---|
| 项荣: "用于刹车片外观质量检验的照明系统质量检验", 《仪器仪表学报》 * |
| 马艳宁等: "基于Halcon的刹车片字符检测算法研究", 《电子科技》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110412052A (en) * | 2019-08-12 | 2019-11-05 | 东莞市艾尔玛科技有限公司 | A kind of curved surface hot padding quality determining method and system |
| CN110412052B (en) * | 2019-08-12 | 2022-02-15 | 艾尔玛科技股份有限公司 | Curved surface hot stamping quality detection method and system |
| CN111192273A (en) * | 2019-12-27 | 2020-05-22 | 西北工业大学 | Digital shot blasting coverage rate measuring method based on computer vision technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110008955B (en) | 2020-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111310558B (en) | Intelligent pavement disease extraction method based on deep learning and image processing method | |
| CN111243032B (en) | Full-automatic detection method for checkerboard corner points | |
| CN111145161B (en) | Pavement crack digital image processing and identifying method | |
| Yu et al. | A novel license plate location method based on wavelet transform and EMD analysis | |
| CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
| CN104392205B (en) | A kind of recognition methods and system of improper license plate | |
| CN114723681B (en) | Concrete crack defect detection method based on machine vision | |
| Wang et al. | Robust and real-time traffic lights recognition in complex urban environments | |
| Yang et al. | A novel approach for license plate character segmentation | |
| CN104915963A (en) | Detection and positioning method for PLCC component | |
| KR101182173B1 (en) | Method and system for recognizing vehicle plate | |
| CN110210448B (en) | Intelligent face skin aging degree identification and evaluation method | |
| CN102693423A (en) | Method for precise positioning of license plate in strong light conditions | |
| CN102419820A (en) | Method for rapidly detecting car logo in videos and images | |
| CN104766344B (en) | Vehicle checking method based on movement edge extractor | |
| Azad et al. | New method for optimization of license plate recognition system with use of edge detection and connected component | |
| Shaikh et al. | A novel approach for automatic number plate recognition | |
| CN106991419A (en) | Method for anti-counterfeit based on tire inner wall random grain | |
| CN109978940A (en) | A kind of SAB air bag size vision measuring method | |
| CN113240626A (en) | Neural network-based method for detecting and classifying concave-convex flaws of glass cover plate | |
| CN110008955A (en) | A kind of automotive brake pads face character coining quality inspection method | |
| CN109635737A (en) | Automobile navigation localization method is assisted based on pavement marker line visual identity | |
| CN108802051B (en) | System and method for detecting bubble and crease defects of linear circuit of flexible IC substrate | |
| CN114719749A (en) | Metal surface crack detection and real size measurement method and system based on machine vision | |
| CN109800641B (en) | Lane line detection method based on threshold value self-adaptive binarization and connected domain analysis |
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: 20201215 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |