CN103279956A - Method for detecting positional accuracy of components of chip mounter - Google Patents
Method for detecting positional accuracy of components of chip mounter Download PDFInfo
- Publication number
- CN103279956A CN103279956A CN2013102099286A CN201310209928A CN103279956A CN 103279956 A CN103279956 A CN 103279956A CN 2013102099286 A CN2013102099286 A CN 2013102099286A CN 201310209928 A CN201310209928 A CN 201310209928A CN 103279956 A CN103279956 A CN 103279956A
- Authority
- CN
- China
- Prior art keywords
- angle point
- image
- circuit board
- coordinate
- components
- 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
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method for detecting positional accuracy of components of a chip mounter. Through the method, angle point coordinates of the components of the chip mounter are extracted. Through coordinate system scale conversion, detailed positions, on a substrate, of the components of the chip mounter are obtained, and the detailed positions are compared with ideal positions. Therefore, accuracy of chip mounting is judged whether to reach the standard or not. Graying and preprocessing are carried out on images, a Harris angle point detecting operator is used as an initial detection algorithm, and thus, a certain number of reference angle points are extracted. Then a self-adapting selection dynamic threshold is used for extracting an appropriate number of alternative angle points. Finally, a Susan angle point detecting operator is used for carrying out final detection on the angle points to be selected, and final angle point coordinates are determined. The coordinates are converted into detailed length information and returned to a computer system. The length information is compared with ideal positions of positioning of the components, evaluation of the positional accuracy of each component of the chip mounter is obtained, and circuit boards with non-standard positional accuracy are screened out. The method is applied to occasions in which the positional accuracy of the components of the chip mounter is detected in flow line production.
Description
Technical field
The present invention relates to the technical field of chip mounter components and parts position probing, particularly relate to the method for the paster components and parts position probing that has corner in the circuit board.
Background technology
The chip automatic positioning technology is one of core technology of automatic placement machine, and its effect is to pick up chip certain altitude above the circuit board welding zone, by optical imaging system and corresponding framing algorithm, makes chip and the accurate contraposition of circuit board welding zone.Bearing accuracy successfully influences most important to paster.Traditional chip mounter localization method mainly contains masterplate matching method and point match method.
The location of automatic placement machine always can be not up to standard, finish after the automatic chip mounting, also need the location of components and parts is detected, on factory's streamline, present stage be the streamline workman on the worktable of oneself, use magnifier that it is detected, pick out location circuit board not up to standard, carry out mark, manually do over again then.The result of investigation sees with regard to present stage, because human cost is lower, the majority circuit plate generates on the streamline, does not use the method for automatic components and parts detection and localization.Therefore, the present invention has adopted a kind of components and parts position finding and detection method based on Corner Detection, finds out bearing accuracy components and parts not up to standard, returns its components and parts number, realizes the automatic detection of components and parts location.
The algorithm of Corner Detection has a lot, must be divided into three major types, based on gray level image, based on binary image with based on contour curve three classes.Method based on gray level image is widely used, wherein the most frequently used Corner Detection Algorithm that is based on Harris template and Susan template.
Summary of the invention
At present stage most of automatic placement machine workflow waterline do not possess the present situation of automatic detection paster components and parts bearing accuracy, the purpose of this invention is to provide a kind of method that detects chip mounter components and parts bearing accuracy, concrete technical scheme is as follows.
A kind of method that detects chip mounter components and parts bearing accuracy, its circuit board level is placed on the zone that the shooting telotism obtains image, after obtaining the components and parts angular coordinate by with the comparison of ideal coordinates value, the location of components and parts is divided into accurately and inaccurate two kinds of situations.
Above-mentioned, the method for detection chip mounter components and parts bearing accuracy, it comprises the steps:
(1) facility environment is set: make the image imaging quality not have tangible barrel distortion or pincushion distortion, if there is significantly distortion in camera lens, then image is corrected; Fix video camera then to the vertical range of circuit board, calculate the physical length (mm/pixel) of unit picture element correspondence, precision should reach 0.1mm/pixel.
(2) image pre-service: the image of camera collection is converted into gray level image;
(3) determine alternative angle point: use Harris Corner Detection operator, gray level image is carried out Corner Detection, threshold value of initialization obtains the Corner Detection result; Add up the angle point number then, compare with desirable angle point number range, if surpass ideal range, according to the stepping amount that pre-sets, increase or reduce threshold value, carry out the Harris Corner Detection again, the angle point number in the image-region is all dropped in the ideal range, or threshold value has reached maximum or the minimum value that allows setting;
(4) determine final angle point: use Susan Corner Detection operator, the alternative angle point that step (3) is determined screens, and removes noise spot, obtains final angle point, and records all angular coordinates;
(5) judge by coordinate system conversion whether the angle point location is accurate: at first, according to judging whether image has obvious rotation, obtain the physical location of angle point on circuit board according to different coordinate transformation formula, compare with the angle point ideal position that prestores then, draw angle point accurate positioning, inaccurate and result that Corner Detection is wrong.
Further optimize, in the step (1), surpass 1% distortion if image exists, then use digital image processing techniques that optical system is carried out distortion correction, with guarantee lens imaging accurately; For the length that guarantees unit picture element correspondence on the image is the ratio of fixing, camera vertical height relative and the circuit board placement platform should be fixed, and circuit board horizontal positioned in image in addition.
Further optimize, step (3) determines that alternative angle point step comprises:
(3.1) the directional derivative I of computed image
x, I
y, and the product of calculated direction derivative
I
Xy, obtain five matrixes identical with the image size;
(3.2) right then
I
XyCarry out gaussian filtering, the autocorrelation matrix M of each point in the computed image,
(3.3) calculated response function R, and R (x, y)=det (M (x, y))-k*trace (M (x, y))
2And (x, y) maximal value in is designated as R to keep R
Max
(3.4) setting threshold, utilize response function R (x, y)〉t*R
MaxAnd non-maximal value suppresses to make a decision, and obtains some angle points;
(3.5) statistics angle point number, judge in whether the angle point number is between desirable number field or whether the threshold value adjustment has reached the border, if do not satisfy above-mentioned condition, adjust threshold value according to default stepping amount, repeating step (3.4), add up the angle point number again, drop on ideal zone up to detected angle point number, perhaps threshold value arrives adjustable border.
Further optimize, step (4) specifically comprises:
(4.1) use circular masterplate to travel through alternative angle point;
(4.2) utilize similarity comparison function c (r, the r of each alternative point
0), calculate its USAN(Univalve Segment Assimilating Nucleus) area n (r
0),
(4.3) two threshold value g are set
Min, g
Max, the calculated response function R,
(4.4) calculate the center of gravity and the distance at masterplate center in USAN zone, as if satisfy the USAN regional barycenter to the distance at masterplate center greater than certain threshold value and R (r
0) 0 o'clock, judge that then it is angle point.
Further optimize, the field color of foregoing circuit plate outside evenly or as far as possible guarantees consistent, and in the step (5), the definition horizontal direction is horizontal ordinate, represents with X, and vertical direction is ordinate, represents with Y, at first takes out the angle point of horizontal ordinate minimum, (X
Min, Y), then take out angle point (X, the Y of ordinate minimum
Min).Make k
1=Y-Y
Min, k
2=X-X
Min,
Utilize k
1, k
2Size, whether the decision circuitry plate horizontal positioned, if circuit board is horizontal positioned in the image, the process that then converts is as follows:
W
X=X
i·K
W
Y=Y
i·K
Wherein, K: the concrete length of representation unit pixel correspondence on circuit board, unit: mm/pixel;
(X
i, Y
i) be the coordinate of this angle point; Coordinate origin is positioned at the position, the upper left corner of substrate, and namely all point coordinate values all are that the image coordinate value deducts substrate upper left corner coordinate figure and obtains;
W
X, W
Y: this angle point is with respect to the distance of the vertical border of circuit board and horizontal boundary;
If circuit board is not horizontal positioned in the image, then define substrate and clockwise turn to positive dirction, namely work as k
1K
2The time, θ〉0; Work as k
1<k
2The time θ<0;
The coordinate transformation relation is as follows:
Wherein, K: the concrete length of representation unit pixel correspondence on circuit board, unit is mm/pixel;
W
L: the physical length of indication circuit plate, unit is mm; (X
Ru, Y
Ru): the coordinate in the indication circuit plate upper right corner;
Behind the known anglec of rotation θ, point (X
i, Y
i) rotate to point ('
i, Y '
i) can get following formula according to rotation relationship:
X′
i=X·cosθ+Y·sinθ
Y′
i=-X·sinθ+Y·cosθ
Thereby this is with respect to the distance W on circuit board border
X, W
YFor:
W
X=X′
i·K;
W
Y=Y′
i·K
By above-mentioned steps, obtain the positional information of angle point on substrate, finished the conversion from the Pixel Information to the length information; Then corner location is judged, at first set a permissible error threshold value, in the angle point ideal coordinates zone that prestores, search for, if can search unique angular coordinate, then be judged to be accurate positioning; If search less than, then set the described permissible error threshold value of several times, search for again, if can determine unique angular coordinate, it is inaccurate then to be judged to be the location, if search for less than, perhaps search plural coordinate, think that then Corner Detection is wrong, the rejecting angle point.
The present invention extracts the angular coordinate of paster components and parts, by the coordinate system ratiometric conversion, obtains paster components and parts particular location on substrate, compares with ideal position, thereby judges whether the paster precision is up to standard.By image is carried out gray processing and pre-service, use Harris Corner Detection operator as the initial survey algorithm, extract the reference angle point of some, use adaptively selected dynamic threshold then, extract the suitable alternative angle point of quantity, utilize Susan Corner Detection operator that above-mentioned angle point to be selected is finally detected at last, determine final angular coordinate, become concrete length information to return to computer system coordinate transformation, compare with the ideal position of components and parts location, obtain the evaluation of paster components and parts bearing accuracy is one by one filtered out bearing accuracy circuit board not up to standard.
Compared with prior art, the present invention has following advantage and technique effect: what the present invention selected for use is the algorithm of the two combination, and the method chosen of combining adaptive threshold value, has improved the adaptability of Corner Detection to environment, has guaranteed to extract the accuracy of angle point.The coordinate figure that Corner Detection provides through coordinate system conversion, draws angle point with respect to level, the vertical length information of circuit board reference point, thereby the paster locating accuracy is judged.
Description of drawings
Fig. 1 is that the present invention detects components and parts positioning flow figure.
Fig. 2 is harris Corner Detection process flow diagram.
Fig. 3 is susan Corner Detection process flow diagram.
Fig. 4 is coordinate system conversion and judges the angle point process flow diagram.
Fig. 5 is a kind of synoptic diagram of the circuit board along continuous straight runs rotation in the image.
Fig. 6 is the another kind of synoptic diagram of the circuit board along continuous straight runs rotation in the image.
Embodiment
Below in conjunction with accompanying drawing enforcement of the present invention is described further, but enforcement of the present invention and protection domain are not limited thereto.
The present invention to the process flow diagram of the detection method of paster components and parts bearing accuracy on the circuit board as shown in Figure 1.Concrete elder generation is placed on the zone that the shooting telotism obtains image to circuit board level, gathers image and also carries out the image pre-service.Then, image is carried out the harris Corner Detection, and obtain alternative angle point by dynamic adjustment threshold value, again by the susan Corner Detection, alternative angle point is carried out last instance afterwards, obtain final angle point.Obtain the concrete distance and position of angle point on circuit board by coordinate transformation at last, and by with the comparison of ideal coordinates position, draw the evaluation to this angular coordinate bearing accuracy.
Specifically comprise following several key step:
(1) facility environment is set.At first, guarantee that the image imaging quality does not exist tangible barrel distortion or pincushion distortion, if there is significantly distortion in camera lens, should correct image earlier.Fix video camera then to the vertical range of circuit board, calculate the physical length (mm/pixel) of unit picture element correspondence, precision should reach 0.1mm/pixel.
(2) image pre-service: the image of camera collection is converted into gray level image, and subsequent step is based on that gray level image carries out.
(3) calculate angular coordinate: at first use Harris Corner Detection operator, gray level image is carried out Corner Detection, next be the iterative process that an adaptive threshold is selected, obtain suitable angle point number, use Susan Corner Detection operator then, the detected alternative angle point of previous step is screened, remove some noise spots, obtain final angle point, and record all angular coordinates.
(4) coordinate system of angle point is converted and check.Obtain the location of pixels of angle point in image coordinate system, the rotation that conversion and the substrate by coordinate system may exist obtains the physical location of angle point on substrate.
Described step (1) facility environment need adjust to adapt to the needs of whole flow processing.At first be the high request to imageing sensor, higher resolution and be no more than 1% distortion.Surpass 1% distortion if image exists, can use digital image processing techniques that optical system is carried out distortion correction, to guarantee the accurate of lens imaging.Be the ratio of fixing for the length that guarantees unit picture element correspondence on the image in addition, camera vertical height relative and the substrate placement platform should be fixed, and the as far as possible horizontal positioned of substrate in image, these working equipments to streamline have also proposed high requirement.
The image that described step (2) video camera obtains according to circumstances, can obtain coloured image after the processing through the camera lens rectification earlier, then coloured image is carried out gray processing and handles, and subsequent step is all finished at gray-scale map.
Described step (3) angle point extracts, and has at first used Harris Corner Detection operator.The basic thought of Harris Corner Detection Algorithm is that along continuous straight runs and vertical direction variation point greatly are judged to be angle point.As shown in Figure 2, the step of Harris Corner Detection is as follows:
1) the directional derivative I of computed image
x, I
yThe method of a gradient calculation perhaps uses the Prewitt/Sobel operator to finish before and after can using, and obtains two matrixes identical with the image size.
2) product of compute gradient
I
XyWith I
x, I
yOneself carries out point multiplication operation respectively, and then multiplication cross, obtains three matrixes identical with the image size.
3) right
I
XyCarry out gaussian filtering, and obtain the autocorrelation matrix M of each point in the image.
* W representative is so that (x does convolution with Gauss's masterplate W centered by y).Each point in the image (x, y) a corresponding matrix M (x, y).
4) calculate respective function R, and R (x, y)=det (M (x, y))-k*trace (M (x, y))
2Wherein k is a fixing threshold value, generally between 0.04~0.06.(x, y) more big, the possibility that this point is angle point is just more big for R.And (x, y) maximal value in is designated as R to keep R
Max
5) setting threshold is judged and is carried out non-maximal value and suppress.Traversing graph is as each pixel, if satisfy R (x, y)〉t*R
MaxCondition (t is initialized as 0.03), and with the R of the square frame of 7*7 size on every side (x y) compares, and the R of this pixel (x, y) response maximum then are defined as it initial survey angle point.
Next, be the process that a self-adaptation is adjusted threshold value.Statistics angle point number is estimated the true angle point number m on the substrate, the desirable angle of determining 2*m left and right sides certain width zone [2*m ?n, the 2*m+n] (n<m) that counts out.Adjust the size of threshold value t, 0.005 the stepping amount of can arranging is regulated, and repeats the 5th) step, add up the angle point number again, drop on ideal zone up to detected angle point number, perhaps threshold value arrives adjustable border.Through this step, setting threshold adapts to the place environment more.
At last, use the Susan operator, above-mentioned alternative angle point is carried out last instance, obtain final angle point.Susan has the strong characteristics of anti-noise ability, can screen noise and the too much noise spot of local detail in some images.
This algorithm is based on the sub-circular masterplate that the pixel field comprises some elements, and each pixel is calculated the region area size similar to him, when this area less than half of whole circular masterplate is, as alternative angle point.Be illustrated in figure 3 as Susan Corner Detection process flow diagram.At first use circular masterplate to travel through alternative angle point.In template, utilize similarity comparison function c (r, r
0), calculate the USAN area of each alternative angle point.Two threshold value g are set
Min, g
Max, calculate the response function R of each alternative angle point again.Calculate the center of gravity in USAN zone and the distance at masterplate center at last.If the distance at center of gravity and masterplate center is greater than certain threshold value and work as this point and satisfy R (r
0) under 0 the condition, think that then this point is correct angle point.Concrete steps are as follows:
1. use circular masterplate to travel through alternative angle point.Circular shuttering is the zone of forming with 37 pixels commonly used.
2. utilize similarity comparison function c (r, the r of each alternative point
0), calculate its USAN area n (r
0).
3. two threshold value g are set
Min, g
Max, the calculated response function R.
g
MaxBe the geometry threshold value that suppresses noise, determined the USAN maximal value of angle point simultaneously.G in addition
MinAnd g
MaxDetermined the angular range of output angle point jointly.Here select g
Max=18, g
Min=6;
4. calculate the center of gravity in USAN zone and the distance at masterplate center.The coordinate figure addition in USAN zone is obtained the barycentric coordinates in USAN zone divided by the number of point.If the distance at center of gravity and masterplate center thinks then that less than certain threshold value it is not correct angle point.
So through above-mentioned steps, alternative angle point is carried out after susan detects, R (r satisfies condition
0) 0 o'clock, judge that then it is angle point.
Described step (4) obtains angular coordinate in the image by described step (3), by coordinate system conversion, obtains this angle point at last with respect to the actual range at circuit board edge.Be illustrated in figure 4 as coordinate system conversion and judge the angle point process flow diagram.The first step: from the angular coordinate that has obtained, find the point of horizontal ordinate and ordinate minimum.Utilize above two point coordinate poor, whether enough decision circuitry plate horizontal pendulum is placed on image acquisition region.Whether level is put according to circuit board, selects different coordinate transformation formula for use, calculates the physical location of angle point on circuit board.Second step: for each angle point through coordinate transformation, according to default permissible error threshold value, in the angle point ideal coordinates zone that prestores, search for, if can search the difference of unique distance with it less than the desirable angle point of error threshold, then judge this angle point accurate positioning, if search less than, then set 3 times to the threshold value of permissible error, search for again, if can determine that unique angular coordinate is corresponding with it, then is judged to be this angle point and locatees inaccurate.If the search less than, or search plural coordinate, think that then Corner Detection is wrong, reject this angle point.
Concrete, to guarantee that at first the zone of circuit board outside is that color evenly even as far as possible guarantees unanimity.The convenient like this angular coordinate that extracts the substrate edge.The definition horizontal direction is horizontal ordinate, represents with X, and vertical direction is ordinate, represents with Y.Take out the angle point of horizontal ordinate minimum then, (X
Min, Y), and the angle point of ordinate minimum (X, Y
Min).Make k
1=Y-Y
Min, k
2=X-X
Min
(1) if
Or
(W
LBe picture traverse), think that then circuit board is horizontal positioned in the image.The following expression of the process that then converts:
W
X=X
i·K
W
Y=Y
i·K
K: the concrete length of representation unit pixel correspondence on circuit board, unit: (mm/pixel).
(X
i, Y
i) be the coordinate of this angle point.It is emphasized that: coordinate origin is positioned at the position, the upper left corner of substrate, and namely all point coordinate values all are that the image coordinate value deducts substrate upper left corner coordinate figure and obtains.
W
X, W
Y: this angle point is with respect to the distance of the vertical border of circuit board and horizontal boundary.
(2) if
And
Think that then there is certain rotation in the circuit board along continuous straight runs in the image.Work as k
1K
2The time, image as shown in Figure 5, if k
1<k
2The time, image is as shown in Figure 6.
The definition substrate clockwise turns to positive dirction, namely works as k
1K
2The time, θ〉0; Work as k
1<k
2The time θ<0.
It is as follows to shift the coordinate transformation relation onto:
K: the concrete length of representation unit pixel correspondence on circuit board, unit: (mm/pixel)
W
L: the physical length of indication circuit plate, unit (mm)
(X
Ru, Y
Ru): the coordinate in the indication circuit plate upper right corner, unit (pixel)
After the known anglec of rotation, can get following formula according to rotation relationship:
X′
i=X·cosθ+Y·sinθ
Y′
i=-X·sinθ+Y·cosθ
Thereby can release, this is with respect to the distance W on circuit board border
X, W
Y, obtain by following formula:
W
X=X′
i·K
W
Y=Y′
i·K
By above-mentioned steps, can obtain the positional information of angle point on substrate, finished the conversion from the Pixel Information to the length information.Then corner location is judged.At first set a permissible error threshold value, in the angle point ideal coordinates zone that prestores, search for, if can search unique angular coordinate, then be judged to be accurate positioning; If search less than, then set 3 times to the threshold value of permissible error, search for again, if can determine unique angular coordinate, then be judged to be locate inaccurate.If search less than, perhaps search plural coordinate, think that then Corner Detection is wrong, reject angle point.The positional information of final the inaccurate point of restoring to normal position.
Claims (7)
1. method that detects chip mounter components and parts bearing accuracy, it is characterized in that circuit board level is placed on the zone that the shooting telotism obtains image, after obtaining the components and parts angular coordinate by with the comparison of ideal coordinates value, the location of components and parts is divided into accurately and inaccurate two kinds of situations.
2. according to the method for right 1 described detection chip mounter components and parts bearing accuracy, it is characterized in that comprising the steps:
(1) facility environment is set: make the image imaging quality not have tangible barrel distortion or pincushion distortion, if there is significantly distortion in camera lens, then image is corrected; Fix video camera then to the vertical range of circuit board, calculate the physical length of unit picture element correspondence;
(2) image pre-service: the image of camera collection is converted into gray level image;
(3) determine alternative angle point: use Harris Corner Detection operator, gray level image is carried out Corner Detection, threshold value of initialization obtains the Corner Detection result; Add up the angle point number then, compare with desirable angle point number range, if surpass ideal range, according to the stepping amount that pre-sets, increase or reduce threshold value, carry out the Harris Corner Detection again, the angle point number in the image-region is all dropped in the ideal range, or threshold value has reached maximum or the minimum value that allows setting;
(4) determine final angle point: use Susan Corner Detection operator, the alternative angle point that step (3) is determined screens, and removes noise spot, obtains final angle point, and records all angular coordinates;
(5) judge by coordinate system conversion whether the angle point location is accurate: at first, according to judging whether image has obvious rotation, obtain the physical location of angle point on circuit board according to different coordinate transformation formula, compare with the angle point ideal position that prestores then, draw angle point accurate positioning, inaccurate and result that Corner Detection is wrong.
3. according to the method for right 2 described detection chip mounter components and parts bearing accuracies, it is characterized in that in the step (1), surpass 1% distortion if image exists, then use digital image processing techniques that optical system is carried out distortion correction, to guarantee the accurate of lens imaging; Camera vertical height relative and the circuit board placement platform should be fixed, and circuit board horizontal positioned in image.
4. according to the method for right 2 described detection chip mounter components and parts bearing accuracies, it is characterized in that step (3) determines that alternative angle point step comprises:
(3.1) the directional derivative I of computed image
x, I
y, and the product of calculated direction derivative
I
Xy, obtain five matrixes identical with the image size;
(3.2) right then
I
XyCarry out gaussian filtering, the autocorrelation matrix M of each point in the computed image,
(3.3) calculated response function R, and R (x, y)=det (M (x, y))-k*trace (M (x, y))
2And (x, y) maximal value in is designated as R to keep R
Max
(3.4) setting threshold, utilize response function R (x, y)〉t*R
MaxAnd non-maximal value suppresses to make a decision, and obtains some angle points;
(3.5) statistics angle point number, judge in whether the angle point number is between desirable number field or whether the threshold value adjustment has reached the border, if do not satisfy above-mentioned condition, adjust threshold value according to default stepping amount, repeating step (3.4), add up the angle point number again, drop on ideal zone up to detected angle point number, perhaps threshold value arrives adjustable border.
5. according to the method for right 2 described detection chip mounter components and parts bearing accuracies, it is characterized in that step (4) specifically comprises:
(4.1) use circular masterplate to travel through alternative angle point;
(4.2) utilize similarity comparison function c (r, the r of each alternative point
0), calculate its USAN area n (r
0),
(4.3) two threshold value g are set
Min, g
Max, the calculated response function R,
(4.4) calculate the center of gravity and the distance at circular masterplate center in USAN zone, as if satisfy the USAN regional barycenter to the distance at masterplate center greater than certain threshold value and R (r
0) 0 o'clock, judge that then it is angle point.
6. according to the method for right 2 described detection chip mounter components and parts bearing accuracies, it is characterized in that the field color of circuit board outside evenly or as far as possible guarantees consistent.
7. according to the method for right 2 described detection chip mounter components and parts bearing accuracies, it is characterized in that in the step (5), the definition horizontal direction is horizontal ordinate, represents with X, vertical direction is ordinate, represents with Y, at first takes out the angle point (X of horizontal ordinate minimum
Min,Y), then take out angle point (X, the Y of ordinate minimum
Min); Make k
1=Y-Y
Min, k
2=X-X
Min,
Utilize k
1, k
2Size, whether the decision circuitry plate horizontal positioned, if circuit board is horizontal positioned in the image, the process that then converts is as follows:
W
X=X
i·K
W
Y=Y
i·K
Wherein, K: the concrete length of representation unit pixel correspondence on circuit board, unit: mm/pixel;
(X
i, Y
i) refer to the angular coordinate that step (4) is finally determined; Coordinate origin is positioned at the position, the upper left corner of substrate, i.e. all angular coordinate value (X
i, Y
i) all be to deduct substrate upper left corner coordinate figure by image actual coordinate value to obtain; Wherein, the substrate coordinate figure is (X
Min,Y);
W
X, W
Y: be that the final angle point of determining of step (4) is with respect to the distance of the vertical border of circuit board and horizontal boundary;
If circuit board is not horizontal positioned in the image, then define substrate and clockwise turn to positive dirction, namely work as k
1K
2The time, θ〉0; Work as k
1<k
2The time θ<0;
The coordinate transformation relation is as follows:
Wherein, K: the concrete length of representation unit pixel correspondence on circuit board, unit is mm/pixel;
W
L: the physical length of indication circuit plate, unit is mm; (X
Ru, Y
Ru): the coordinate in the indication circuit plate upper right corner;
Behind the known anglec of rotation θ, point (X
i, Y
i) rotate to point (X '
i, Y '
i) can get following formula according to rotation relationship:
X′
i=X·cosθ+Y·sinθ
Y′
i=-X·sinθ+Y·cosθ
Thereby this is with respect to the distance W on circuit board border
X, W
YFor:
W
X=X′
i·K;
WY=Y′
i·K
By above-mentioned steps, obtain the positional information of angle point on substrate, finish the conversion from the Pixel Information to the length information; Then corner location is judged, at first set a permissible error threshold value, in the angle point ideal coordinates zone that prestores, search for, if can search unique angular coordinate, then be judged to be accurate positioning; If search less than, then set the described permissible error threshold value of several times, search for again, if can determine unique angular coordinate, it is inaccurate then to be judged to be the location, if search for less than, perhaps search plural coordinate, think that then Corner Detection is wrong, the rejecting angle point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310209928.6A CN103279956B (en) | 2013-05-30 | 2013-05-30 | A kind of method detecting chip mounter components and parts positioning precision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310209928.6A CN103279956B (en) | 2013-05-30 | 2013-05-30 | A kind of method detecting chip mounter components and parts positioning precision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103279956A true CN103279956A (en) | 2013-09-04 |
CN103279956B CN103279956B (en) | 2016-06-29 |
Family
ID=49062463
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310209928.6A Expired - Fee Related CN103279956B (en) | 2013-05-30 | 2013-05-30 | A kind of method detecting chip mounter components and parts positioning precision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103279956B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103785622A (en) * | 2014-01-28 | 2014-05-14 | 浙江理工大学 | Machine-vision-based part sorting device and method |
CN104463896A (en) * | 2014-12-26 | 2015-03-25 | 武汉大学 | Image corner point detection method and system based on kernel similar region distribution characteristics |
CN104792263A (en) * | 2015-04-20 | 2015-07-22 | 合肥京东方光电科技有限公司 | Method and device for determining to-be-detected area of display mother board |
CN105389805A (en) * | 2015-10-25 | 2016-03-09 | 合肥市航嘉电子技术有限公司 | Optical detection method of component shortage, wrong component and reverse component in plug-in operation |
CN105809657A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Angular point detection method and device |
CN106485749A (en) * | 2016-10-19 | 2017-03-08 | 哈尔滨工业大学 | A kind of rectangular pins element rough localization method based on angle point |
CN107133724A (en) * | 2017-04-18 | 2017-09-05 | 冀月斌 | A kind of chip mounter imports bearing calibration and the system of creation data |
CN109429473A (en) * | 2017-08-28 | 2019-03-05 | 株洲中车时代电气股份有限公司 | Automatic check method and device with polarity electronic component in circuit board |
CN109556515A (en) * | 2018-11-26 | 2019-04-02 | 深圳市万相源科技有限公司 | A kind of systematic error calibration method, system and equipment based on machine vision |
CN111401266A (en) * | 2020-03-19 | 2020-07-10 | 杭州易现先进科技有限公司 | Method, device, computer device and readable storage medium for positioning corner points of drawing book |
CN111754461A (en) * | 2020-05-28 | 2020-10-09 | 江苏理工学院 | Method and device for positioning image character area of semiconductor chip |
CN112784851A (en) * | 2019-11-11 | 2021-05-11 | 珠海格力电器股份有限公司 | Threshold value selection circuit and method, and angular point detection circuit and method |
CN113405451A (en) * | 2020-10-15 | 2021-09-17 | 南京航空航天大学 | Tooth-shaped structure assembling and measuring method based on monocular vision |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH09257414A (en) * | 1996-03-25 | 1997-10-03 | Kobe Steel Ltd | Object position detector |
CN201955592U (en) * | 2010-12-09 | 2011-08-31 | 郑春晓 | Multifunctional visual automatic alignment laminating machine |
-
2013
- 2013-05-30 CN CN201310209928.6A patent/CN103279956B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH09257414A (en) * | 1996-03-25 | 1997-10-03 | Kobe Steel Ltd | Object position detector |
CN201955592U (en) * | 2010-12-09 | 2011-08-31 | 郑春晓 | Multifunctional visual automatic alignment laminating machine |
Non-Patent Citations (5)
Title |
---|
A RYAN等: "Manufacturing an environmentally friendly PCB using existing industrial processes and equipment", 《ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING》 * |
何丰等: "一种改进的图像快速高精度角点检测算法", 《电视技术》 * |
姜涛等: "基于机器视觉的印刷电路板误差校正方法", 《上海交通大学学报》 * |
杨立玲等: "基于改进SUSAN算法的片式元件角度识别", 《半导体技术》 * |
杨立玲等: "基于改进SUSAN算法的片式元件角度识别", 《半导体技术》, vol. 34, no. 7, 31 July 2009 (2009-07-31), pages 637 - 640 * |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103785622A (en) * | 2014-01-28 | 2014-05-14 | 浙江理工大学 | Machine-vision-based part sorting device and method |
CN103785622B (en) * | 2014-01-28 | 2015-12-30 | 浙江理工大学 | Based on the part method for sorting of the part sorting equipment of machine vision |
CN104463896A (en) * | 2014-12-26 | 2015-03-25 | 武汉大学 | Image corner point detection method and system based on kernel similar region distribution characteristics |
CN104463896B (en) * | 2014-12-26 | 2017-04-12 | 武汉大学 | Image corner point detection method and system based on kernel similar region distribution characteristics |
CN105809657A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Angular point detection method and device |
US10210605B2 (en) | 2015-04-20 | 2019-02-19 | Boe Technology Group Co., Ltd. | Method and device for detecting boundary of region on display motherboard |
CN104792263A (en) * | 2015-04-20 | 2015-07-22 | 合肥京东方光电科技有限公司 | Method and device for determining to-be-detected area of display mother board |
CN105389805A (en) * | 2015-10-25 | 2016-03-09 | 合肥市航嘉电子技术有限公司 | Optical detection method of component shortage, wrong component and reverse component in plug-in operation |
CN106485749A (en) * | 2016-10-19 | 2017-03-08 | 哈尔滨工业大学 | A kind of rectangular pins element rough localization method based on angle point |
CN106485749B (en) * | 2016-10-19 | 2019-01-22 | 哈尔滨工业大学 | A kind of rectangular pins element rough localization method based on angle point |
CN107133724A (en) * | 2017-04-18 | 2017-09-05 | 冀月斌 | A kind of chip mounter imports bearing calibration and the system of creation data |
CN109429473A (en) * | 2017-08-28 | 2019-03-05 | 株洲中车时代电气股份有限公司 | Automatic check method and device with polarity electronic component in circuit board |
CN109556515A (en) * | 2018-11-26 | 2019-04-02 | 深圳市万相源科技有限公司 | A kind of systematic error calibration method, system and equipment based on machine vision |
CN109556515B (en) * | 2018-11-26 | 2019-11-12 | 深圳市万相源科技有限公司 | A kind of systematic error calibration method, system and equipment based on machine vision |
CN112784851A (en) * | 2019-11-11 | 2021-05-11 | 珠海格力电器股份有限公司 | Threshold value selection circuit and method, and angular point detection circuit and method |
CN111401266A (en) * | 2020-03-19 | 2020-07-10 | 杭州易现先进科技有限公司 | Method, device, computer device and readable storage medium for positioning corner points of drawing book |
CN111401266B (en) * | 2020-03-19 | 2023-11-03 | 杭州易现先进科技有限公司 | Method, equipment, computer equipment and readable storage medium for positioning picture corner points |
CN111754461A (en) * | 2020-05-28 | 2020-10-09 | 江苏理工学院 | Method and device for positioning image character area of semiconductor chip |
CN111754461B (en) * | 2020-05-28 | 2024-03-01 | 江苏理工学院 | Method and device for positioning image character area of semiconductor chip |
CN113405451A (en) * | 2020-10-15 | 2021-09-17 | 南京航空航天大学 | Tooth-shaped structure assembling and measuring method based on monocular vision |
CN113405451B (en) * | 2020-10-15 | 2022-05-31 | 南京航空航天大学 | Tooth-shaped structure assembling and measuring method based on monocular vision |
Also Published As
Publication number | Publication date |
---|---|
CN103279956B (en) | 2016-06-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103279956A (en) | Method for detecting positional accuracy of components of chip mounter | |
CN107424160B (en) | System and method for finding image central line by visual system | |
US9818026B2 (en) | People counter using TOF camera and counting method thereof | |
CN106529587B (en) | Vision course recognition methods based on object detection | |
WO2017092631A1 (en) | Image distortion correction method for fisheye image, and calibration method for fisheye camera | |
CN106778737B (en) | A kind of license plate antidote, device and a kind of video acquisition device | |
CN112486207A (en) | Unmanned aerial vehicle autonomous landing method based on visual identification | |
CN108007388A (en) | A kind of turntable angle high precision online measuring method based on machine vision | |
CN103235939A (en) | Datum point positioning method based on machine vision | |
CN103795935B (en) | A kind of camera shooting type multi-target orientation method and device based on image rectification | |
US20130113897A1 (en) | Process and arrangement for determining the position of a measuring point in geometrical space | |
CN101002069A (en) | Method of preparing a composite image with non-uniform resolution | |
CN105718931B (en) | System and method for determining clutter in acquired images | |
CN109360794B (en) | Visual detection method and device for secondary printing precision of crystalline silicon photovoltaic solar cell electrode | |
CN106600592A (en) | Track long chord measurement method based on the splicing of continuous frame images | |
CN109035326A (en) | High-precision location technique based on sub-pix image recognition | |
CN113538583A (en) | Method for accurately positioning position of workpiece on machine tool and vision system | |
CN112017249A (en) | Vehicle-mounted camera roll angle obtaining and mounting angle correcting method and device | |
CN105021127A (en) | Surface mounting machine reference camera calibration method | |
CN108288065A (en) | A kind of four-wheel position finder detection method based on image analysis | |
KR101469099B1 (en) | Auto-Camera Calibration Method Based on Human Object Tracking | |
CN115717867A (en) | Bridge deformation measurement method based on airborne double cameras and target tracking | |
CN113607058B (en) | Straight blade size detection method and system based on machine vision | |
CN107767366B (en) | A kind of transmission line of electricity approximating method and device | |
CN105758337A (en) | Method for obtaining included angel between a lens plane and an image sensor plane |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160629 |
|
CF01 | Termination of patent right due to non-payment of annual fee |