CN103630542B - Defect detecting device, defect correction device and defect inspection method - Google Patents
Defect detecting device, defect correction device and defect inspection method Download PDFInfo
- Publication number
- CN103630542B CN103630542B CN201310377121.3A CN201310377121A CN103630542B CN 103630542 B CN103630542 B CN 103630542B CN 201310377121 A CN201310377121 A CN 201310377121A CN 103630542 B CN103630542 B CN 103630542B
- Authority
- CN
- China
- Prior art keywords
- defect
- pixel
- brightness
- substrate
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 0 *C1*CCC1 Chemical compound *C1*CCC1 0.000 description 1
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The defects of of the invention, detection means included：Camera, camera shooting substrate simultaneously export electric signal；And image processing apparatus, in the image processing apparatus input have the electric signal exported by camera.Image processing apparatus uses contrast larger in the respective picture contrast of multiple signals that is obtained from electric signal, portion the defects of to detect substrate.Due to detecting defective part using contrast larger in the respective picture contrast of multiple signals, therefore even for the less defective part of contrast in the image being divided into before each signal between normal portion, also defective part is capable of detecting when, without significantly reducing sensitivity.
Description
Technical field
The present invention relates to defect detecting device, defect correction device and defect inspection method, more particularly to being formed in base
The technology that the defects of repeat patterns on plate is detected.
Background technology
In liquid crystal display (LCD：Liquid Crystal Display) TFT (Thin Film Transistor：It is thin
Film transistor) substrate etc. production line in, use correcting device the defects of being modified to the defects of TFT substrate.As amendment
The size of the defects of object has various.Particularly, it is provided with substrate in the TFT substrate of the circuits such as transistor, also deposits
The situation that size as object the defects of is several μm.Accordingly it is desirable to checked using high-resolution lens.
In addition, the representative as liquid crystal display, with FPD (flat panel display：Flat-panel monitor) it is big
The development of type, Pixel Dimensions are also becoming big.Japanese Patent Laid-Open 2011-203710 publications (patent document 1) disclose
The method that large area defect is tracked while detection resolution is maintained.
The defects of described in the claim 8 of Japanese Patent Laid-Open 2011-203710 publications, correcting device prolonged in defect
In the case of reaching beyond coverage, track extension, the image photographed during to by making tracking carry out fitting so as to
The defects of being included in the image of generation is detected.
In Japanese Patent Laid-Open 2011-203710 publications, carried out premised on it can detect defect exactly each
Kind explanation.However, have as feature the defects of correcting object various.The defects of there is also being not easy to detect.Especially
It is that the detection of the defects of contrast is relatively low turns into problem.
For this problem, Japanese Patent Laid-Open 2000-146537 publications (patent document 2) are proposed using differential
Interference opticses are to ITO (the Indium Tin Oxide in TFT substrate：Tin indium oxide) film the defects of the side detected
Method.However, in the case where loading laser, optical system becomes complicated, so as to turn into expensive structure.
Prior art literature
Patent document
Patent document 1：Japanese Patent Laid-Open 2011-203710 publications
Patent document 2：Japanese Patent Laid-Open 2000-146537 publications
The content of the invention
The technical problems to be solved by the invention
The present invention is to complete in view of the above problems, it is intended that improving the detection spirit for low contrast defect
Sensitivity is without loading special optical system.
Technical scheme used by solution technical problem
In a certain embodiment, defect detecting device includes：Camera, camera shooting substrate simultaneously export electric signal；
And image processing apparatus, in the image processing apparatus input have the electric signal exported by camera.Image processing apparatus uses
Larger contrast in the respective picture contrast of multiple signals obtained from electric signal, portion the defects of to detect substrate.
Invention effect
Due to detecting defective part using contrast larger in the respective picture contrast of multiple signals, thus it is even if right
In the less defective part of contrast in the image being divided into before each signal between normal portion, also it is capable of detecting when to lack
The portion of falling into, without significantly reducing sensitivity.Thus, without the special optical system loaded for detecting defective part.Cause
This, can use simple and cheap optical system.Because optical system turns into simple structure, therefore maintainability can be suppressed
Reduction.
Brief description of the drawings
Fig. 1 is the integrally-built figure for representing defect correction device.
Fig. 2 is the stereogram for representing viewing optical system and the major part of ink coater structure.
Fig. 3 is the figure for representing ink coating action.
Fig. 4 is the functional block diagram of image processing apparatus.
Fig. 5 is the figure (one) for the picture for representing monitor.
Fig. 6 be the picture for representing monitor figure (secondly).
Fig. 7 is the figure for representing to search for the order of defect.
Fig. 8 is the figure for representing searching position.
Fig. 9 is the figure for representing panoramic picture.
Figure 10 is the figure for the coordinate value for representing X-axis workbench and Y-axis workbench.
Figure 11 is the figure for representing comparison processing when pattern spacing is h ＜ w.
Figure 12 is the figure for representing comparison processing when pattern spacing is h >=w.
Figure 13 be represent remove periodic component after image figure.
Figure 14 is the figure for the those suspected defects for representing line segment shape.
Figure 15 is the figure for being denoted as the bounding rectangle that defect candidate extraction goes out.
Figure 16 is the figure for the coordinate for representing bounding rectangle center.
Figure 17 be represent cutting start when cutting position centre coordinate and cutting at the end of cutting position center
The figure of coordinate.
Figure 18 is the figure for representing the cutting position during cutting operation.
Figure 19 is the centre coordinate of the centre coordinate for being coated with circle when representing coating beginning and the coating circle at the end of coating
Figure.
Figure 20 is the figure for the centre coordinate for representing the coating circle in ink coating action process.
Embodiment
Below, embodiments of the present invention are described with reference to.In the following description, same parts are added with identical mark
Number.Their title and function is also identical.Therefore, do not repeat on their detailed description.
[apparatus structure]
Fig. 1 is the integrally-built figure for representing the defects of present embodiment correcting device 100.Defect correction device 100
Including：Head is corrected, the amendment head is by viewing optical system 31, CCD (Charge Coupled Device：Charge-coupled device
Part) camera 32, cutting laser aid 33, ink coater structure 34 and ink solidification light source 35 form；Z axis workbench
36, the Z axis workbench 36 makes the amendment head in direction (the Z axis side vertical with the liquid crystal color filter substrate 5 of amendment object
To) on move；X-axis workbench 37, the X-axis workbench 37 load Z axis workbench 36 so that it is moved in the X-axis direction；Y-axis work
Make platform 38, the Y-axis workbench loads substrate 5 so that it is moved in the Y-axis direction；Control computer 39, the control calculate
Machine 39 controls the action of whole device；Image processing apparatus 6, input has in the image processing apparatus 6 is shot by CCD camera 32
To image and the image is handled；Monitor 40, the display image of monitor 40, the actions menu etc. of device；And
Guidance panel 41, the guidance panel 41 are used to input the instruction from staff to control computer 39.
Viewing optical system 31 observes the surface state of substrate 5, be coated with by ink coater structure 34 after amendment ink
State.Image observed by viewing optical system 31 is converted into including red (r), green (g), blue (b) by CCD camera 32
Electric signal including respective color signal, and be shown on monitor 40.Cutting is with laser aid 33 via observation optical system
System 31 irradiates laser to be removed to the nonuseable part on substrate 5.
Ink coater structure 34 is to caused White Defects coating amendment ink on substrate 5, to be modified.Ink solidification
With light source 35 for example including CO2 lasers, laser is irradiated to the amendment ink after the coating of ink coater structure 34 so that it is solid
Change.
In addition, the apparatus structure is an example, such as it can also use the structure of hereinafter referred to as stand mode：It will load
The Z axis workbench 36 for having the grade of viewing optical system 31 is loaded into X-axis workbench 37, and X-axis workbench 37 then is loaded into Y-axis work
Make platform 38, Z axis workbench 36 is moved up in XY side, as long as making the Z axis work for being mounted with the grade of viewing optical system 31
Platform 36 can relatively move relative to the substrate 5 of amendment object on XY directions, then can use arbitrary structures.
Next, the example of the ink coater structure using multiple pins is illustrated.
Fig. 2 is the stereogram for representing viewing optical system 31 and the major part of ink coater structure 34.Fig. 3 (a)~(c)
Figure when being the major part from Fig. 2 A directions, it is the figure for representing ink coating action., should in Fig. 2 and Fig. 3 (a)~(c)
Defect correction device 100 includes：Different multiple (such as 5) object lens 2 of movable platen 42, multiplying power and for being coated with different face
Multiple (such as 5) coater units 43 of the ink of color.
Movable platen 42 is arranged between the observation lens barrel 31a of viewing optical system 31 lower end and substrate 5 in X-axis
Moved in direction and Y direction.In addition, formed with 5 corresponding with 5 object lens 2 respectively through holes on movable platen 42
42a。
Configure 5 through hole 42a at a predetermined interval in the Y-axis direction.Each object lens 2 are fixed on the following table of movable platen 42
Face so that its optical axis is consistent with corresponding through hole 42a center line.In addition, observation lens barrel 31a optical axis and each object lens
2 optical axis is configured in the Z-direction vertical with X-direction and Y direction.
In addition, 5 coater units 43 are fixed on the lower surface of movable platen 42 at a predetermined interval in the Y-axis direction.5 coatings
Unit 43 is configured to adjacent with 5 object lens 2 respectively.
Using the ink coater structure of multiple pins due to it is known that various technologies than that described above, therefore do not weigh
It is multiple to describe in detail.One example of ink coater structure is shown in Japanese Patent Laid-Open 2009-122259 publications etc..
Defect correction device 100 by using the mechanism for example shown in Fig. 2 using as ink coater structure 34, so as to using multiple
The pin of desired coating diameter carrys out corrective pitting in pin.
Fig. 4 is the functional block diagram of the image processing apparatus 6 of the present invention.Reference picture 4, image processing apparatus 6 include：Image is defeated
Enter portion 61, memory 62 and defects detection portion 65.
Moreover, defects detection portion 65 includes：Image-capturing positions operational part 67, image converter section 68, panoramic picture generation
Portion 69, image comparing section 70, defect extraction unit 71 and defect coordinate operational part 72.
Image processing apparatus 6 reads memory 62 from CCD camera 32 and via image input unit 61 by view data
In.
CCD camera 32 is colored CCD camera, and CCD each pixel performance is three kinds of colors of red, green, blue, will be with CCD
The identical quantity of resolution ratio of camera 32 is a pair of digital independent into memory 62 with (red, green, blue).
Defects detection portion 65 is with reference to the image read in memory 62, to detect defect.The seat of the defects of detected
Scale value is sent to control computer 39.
The defects of control computer 39 is based on receiving coordinate, obtains the movement of X-axis workbench 37 and Y-axis workbench 38
Coordinate, move viewing optical system 31.In addition, in the following description, also can be total by X-axis workbench 37 and Y-axis workbench 38
Referred to as XY worktable.
[defects detection action]
Defects detection portion 65 is handled with order as shown below perform detection.
In placing substrate 5, control computer 39 from the check device being provided separately with defect correction device 100 (not
Diagram) defective locations information is received, the movement to defective locations is completed before detection process.
However, the XY worktable of check device and defect correction device 100 can because machining accuracy, assembly precision difference and
Produce error so that even if viewing optical system 31 is moved into the position shown by check device, as shown in figure 5, also not
Necessarily can be by the picture center of defect location to monitor 40.In addition, according to defect size, as shown in fig. 6, sometimes also can nothing
Method is completely into in picture.
Due to the above situation, in defects detection portion 65, judge whether to be capable of detecting when the general image of defect, and as schemed
Shown in 7, spirally search for around initial picture.Numbering in Fig. 7 represents the order scanned for.That is, in Fig. 7, compile
Number " 1 " scans for equivalent to initial picture, and according to increased order is numbered.
Image-capturing positions operational part 67 calculates searching position.If the pattern spacing of substrate 5 is set into (w, h), will examine
The visual field for looking into the middle object lens 2 used is sized to (Sw, Sh), and the lap between picture is set into V, then as shown in figure 8, when figure
When case spacing is h ＜ w, calculate searching position to cause the distance between upper picture in longitudinal direction for h, in transverse direction between picture away from
From for Sw-V.As h >=w, searching position is calculated to cause laterally for w, longitudinal direction is Sh-V.
In present embodiment, in order to extract the general image of defect, as shown in figure 9, generating one of Fig. 7 to photographing
Individual or multiple images synthesized after panoramic picture, and checked using the panoramic picture generated.As described later, it is right
The respective color signal of red, green, blue and luminance signal generate panoramic picture.
In addition, in defect extraction unit 71, it is assumed that the existing defects in the pattern in center, examined using repeat patterns inspection technique
Survey defect, the repeat patterns inspection technique by two upper and lower or adjacent left and right patterns be set to normal pattern come it is more mutual dense
It is light.
Therefore, the precondition of panoramic picture is：As h ＜ w, the size of longitudinal direction is at least 3 × more than h, horizontal as h >=w
To size be at least 3 × more than w.Here, the life of panoramic picture when showing to check h ＜ w pattern Fig. 7 picture " 1 "
Into example.
As shown in Figure 10, the coordinate value of the XY worktable during image taking of current picture is set to (xc, yc) in advance, will
The coordinate value of XY worktable during the image taking of adjacent two pictures up and down is set to (x1, y1), (x2, y2).In addition,
The coordinate value of XY worktable is shown as (tx, ty) on image all the time.(tx, ty) is generally consistent with the optical axis of object lens 2.In addition, figure
In 10, although the coordinate value of XY worktable and the coordinate value of image refer to same position, because the coordinate of XY worktable is former
Point is different from the origin of coordinates of image, therefore coordinate value is different in itself.
In addition, the resolution ratio of image is set to (M, N) in advance, and by the XY worktable suitable with image pixel
Amount of movement is set to (mx, my).(mx, my) is determined by the object lens 2 used in image taking.
First, the coordinate value of the left upper end of each image, XY worktable at bottom righthand side is obtained.
[the left upper end coordinate value of current search position]
（Xclu, yclu）=（Xc-tx × mx, yc-ty × my）
[the bottom righthand side coordinate value of current search position]
（Xcrb, ycrb）=（Xc+（M-tx）× mx, yc+（N-ty）× my）
[the left upper end coordinate value of side adjacency search position]
（X1lu, y1lu）=（X1-tx × mx, y1-ty × my）
[the bottom righthand side coordinate value of side adjacency search position]
（X1rb, y1rb）=（X1+（M-tx）× mx, y1+（N-ty）× my）
[the left upper end coordinate value of opposite side adjacency search position]
（X2lu, y2lu）=（X2-tx × mx, y2-ty × my）
[the bottom righthand side coordinate value of opposite side adjacency search position]
（X2rb, y2rb）=（X2+（M-tx）× mx, y2+（N-ty）× my）
Next, minimum or maximum coordinate value is obtained from calculated coordinate value.
[min coordinates value]
（Minx, miny）
=（Min（Xclu, x1lu, x2lu）, Min（Yclu, y1lu, y2lu））
[maximum coordinate value among]
（Maxx, maxy）
=（Max（Xcrb, x1rb, x2rb）, Max（Ycrb, y1rb, y2rb））
Here, Min () is the function for the minimum value for returning to the value in bracket.Max () is the function for returning to maximum.
The resolution ratio (Mp, Np) of panoramic picture is obtained by following formula.
（Mp, Np）=（（Maxx-minx）/ mx,（Maxy-miny）/ my）
Prepare Mp × Np of storage panoramic image data array, the position calculated by following formula is set to left upper end, to each
View data on searching position is replicated.The resolution ratio of the view data replicated is M × N.
[the upper left end position of the view data of current search position]
（Xclu-minx）/ mx-tx,（Yclu-miny）/ my-ty
[the upper left end position of the view data of side adjacency search position]
（X1lu-minx）/ mx-tx,（Y1lu-miny）/ my-ty
[the upper left end position of the view data of opposite side adjacency search position]
（X2lu-minx）/ mx-tx,（Y2lu-miny）/ my-ty
According to Fig. 7 search order and utilization side images, i.e. including defective part image and defect periphery image
Multiple images generate panoramic picture, be confirmed whether to detect the general image of defect and terminate at the time of being capable of detecting when
Search.
The judgement of the general image of defect is carried out as described below.On the side up and down of the bounding rectangle of defect
When the length L to connect with picture boundaries is bigger than pre-determined length Lmin, it is judged as connecting, and continues search for.Than Lmin
Hour end processing.
[comparison of view data and defect extraction]
The original digital image data of panoramic picture is generated by image converter section 68.Image converter section 68 is from reading in memory 62
Color image data in extract red (r), green (g), blue (b) respective color signal (colouring information), generate three width images
Data r, g, b.The resolution ratio of each width image is M × N.In addition, generated according to the data of three width images and for example based on following formula
Equivalent to luminance signal (monochrome information) G of brightness of image, data r, g, b, G of generation four width images altogether.
If a certain position on image is set to (x, y), then the data G (x, y) of the position (x, y) on image G is：
G（X, y）
=0.299 × r（X, y）+ 0.587 × g（X, y）+ 0.114 × b（X, y）.
In addition, conversion method is not limited to above formula, such as it can also use the maximum in (red, green, blue).
Panoramic picture generating unit 69 copies to r, g, b, G for being generated the array of data of four Mp × Np panoramic picture
In.As described above, the position calculated by following formula is set into left upper end, the view data on each searching position is replicated.
[the upper left end position of the view data of current search position]
（Xclu-minx）/ mx-tx,（Yclu-miny）/ my-ty
[the upper left end position of the view data of side adjacency search position]
（X1lu-minx）/ mx-tx,（Y1lu-miny）/ my-ty
[the upper left end position of the view data of opposite side adjacency search position]
（X2lu-minx）/ mx-tx,（Y2lu-miny）/ my-ty
Above-mentioned four panoramic image datas generated like that are set to Pr, Pg, Pb, PG in advance.Next, image comparing section
70 remove periodic component from Pr, Pg, Pb, PG view data.
The example of the comparison processing in the picture " 1 " for performing Fig. 7 when pattern spacing is h ＜ w is shown in Figure 11.In advance
A certain image in Pr, Pg, Pb, PG is set to f, and ph is set to by spacing is compared.Here, it is set to the level for panoramic picture
Position x a-b sections are compared processing.That is, it is respective complete in the color signal and luminance signal generated based on repetition interval
In scape view data, to arbitrary check point compared with its peripheral part.
If the view data of the upright position y on a-b sections is set into f (x, y), then using the data as check object,
And s-p (x, y) and s+p (x, y) is obtained, to be used as data f (x, y-ph) and f (x, y+ph) by following formula.
S-p（X, y）=f（X, y）- f（X, y-ph）
S+p（X, y）=f（X, y）- f（X, y+ph）
Check the inspection and judge check point whether than just for being divided into and judging whether check point is more darker than normal portions
The normal brighter inspection in position both carry out.Will storage whether the image of darker judged result is set to dk, will storage whether
The image of brighter judged result is set to br.
On dk, at s-p (x, y)<(that is, check point compares periphery in the case that s+p (x, y) symbol is also consistent when 0
The darker situation in position), will | s-p (x, y) | and | s+p (x, y) | in a less side be set to dk (x, y).
On br, at s-p (x, y)>(that is, check point compares periphery in the case that s+p (x, y) symbol is also consistent when 0
The brighter situation in position), will | s-p (x, y) | and | s+p (x, y) | in a less side be set to br (x, y).
Above-mentioned processing is carried out for image f all (x, y).
The comparison processing in the picture " 1 " of Fig. 7 when pattern spacing is h >=w is performed in Figure 12.By in Pr, Pg, Pb, PG
A certain image be set to f, spacing will be compared and be set to pw.Here, it is set to enter for the upright position y of panoramic picture a-b sections
Row is relatively handled.
If the view data of the horizontal level x on a-b sections is set into f (x, y) in advance, then using the data as inspection pair
As, and s-p (x, y) and s+p (x, y) is obtained, to be used as data f (x-pw, y) and f (x+pw, y) by following formula.
S-p（X, y）=f（X, y）- f（X-pw, y）
S+p（X, y）=f（X, y）- f（X+pw, y）
Check the inspection and judge check point whether than just for being divided into and judging whether check point is more darker than normal portions
The normal brighter inspection in position both carry out.Will storage whether the image of darker judged result is set to dk, will storage whether
The image of brighter judged result is set to br.
On dk, at s-p (x, y)<In the case that s+p (x, y) symbol is also consistent when 0, will | s-p (x, y) | and | s+p
(x, y) | in a less side be set to dk (x, y).
On br, at s-p (x, y)>In the case that s+p (x, y) symbol is also consistent when 0, will | s-p (x, y) | and | s+p
(x, y) | in a less side be set to br (x, y).
Above-mentioned processing is carried out for image f all (x, y).
The image after removing periodic component is shown in Figure 13.By observing accompanying drawing, defective part is only extracted.
In above-mentioned relatively processing, (pw, ph) represents the comparison spacing of data, is calculated by following formula.
Pw=w/mx
Ph=h/my
Because image is discrete data, therefore (pw, ph) rounds numerical value.Thus, when (pw, the ph) of above formula is not integer
Error can be produced, different positions can be strictly compared each other.This turns into doubtful in the generation of the border of pattern
The main reason for defect, if borderline brightness changes greatly, the doubtful of line segment shape for example shown in Figure 14 is produced sometimes
Defect.
In order to not produce the those suspected defects of this line segment shape, also can be adjacent with the pixel used in comparison or neighbouring
Comparison brightness between pixel, when their difference value is more than a certain threshold value just without comparing.
Will be dark to being shown in image dk, br after image Pr, Pg, Pb, PG removing periodic component, relative to normal portion
The image of defect is set to dPr, dPg, dPb, dPG, and the image for showing the defects of brighter is set into bPr, bPg, bPb, bPG.
According to the image dk after image Pr, Pg, Pb, PG removing periodic component, the figure eventually for extraction defect is generated
As pdk.If a certain position on image is set to (x, y), then the data pdk (x, y) at the same position on image pdk is：
Pdk（X, y）
=Max（DkPr（X, y）, dkPg（X, y）, dkPb（X, y）, dkPG（X, y））.
In addition, according to the image br after image Pr, Pg, Pb, PG removing periodic component, generate eventually for extraction defect
Image pbr.If a certain position on image is set to (x, y), then the data pbr (x, y) at the same position on image pbr
For：
Pbr（X, y）
=Max（BrPr（X, y）, brPg（X, y）, brPb（X, y）, brPG（X, y））.
As described above, generation shows the image pdk of the defects of dark relative to normal portion and shows the figure of the defects of brighter
As pbr.That is, any check point and peripheral part by each color signal and each luminance signal in panoramic image data
It is compared, larger contrast is selected from comparative result.Here, the contrast all bigger than other any one contrasts is used
As the contrast between check point and peripheral part.
Although in addition, general in inspection can use the PG for representing brightness of image mostly, also deposited according to object difference
The situation of higher contrast is shown in Pr, Pg, Pb than PG, due to using maximum therein to enter with higher sensitivity
Row detection, therefore used function Max ().
Defect extraction unit 71 is extracted based on image pdk, the pbr generated by image comparing section 70 and using binary conversion treatment
Go out defective part.As an example, by the data pdk (x, y) of image pdk position (x, y) compared with threshold value Tdk, when
Pdk (x, y) value is set to 1 when larger, be set to 0 when smaller, thus generates binary image Bdk.Similarly, by image
The data pbr (x, y) of pbr position (x, y) is set to " 1 " when pbr (x, y) value is larger compared with threshold value Tbr, when
" 0 " is set to when smaller, thus generates binary image Bbr.
For the binary image Bdk and Bbr being generated as above, value is extracted by known mark processing etc.
The Connected component of " 1 ", go out the inside of the bounding rectangle as defect candidate extraction.The image to Figure 13 is shown in Figure 15
Extraction result.Dotted line frame in figure is bounding rectangle.
If the length and width dimensions of calculated bounding rectangle are set into (rw, rh), by minimum extraction size be set to (rwmin,
Rhmin), then the Connected component of following formula will be met as defect.
（Rw+rh）/ 2 >=（Rwmin+rhmin）/ 2
Wherein, above-mentioned decision condition and this need not be defined in, the characteristic of defect can be also taken into account, extract example
Such as meet rw >=rwmin and rh >=rhmin Connected component, meet rw >=rwmin or rh >=rhmin Connected component.
Defect coordinate operational part 72 obtains the coordinate of XY worktable using the bounding rectangle of defect.If by circumscribed rectangular
The left upper end coordinate of shape is set to (rx, ry), then its centre coordinate (cx, cy) is obtained by following formula.
（Cx, cy）=（Rx+rw/2, ry+rh/2）
As shown in figure 16, because the coordinate of the XY worktable of the left upper end opening position of panoramic picture is (minx, miny), and
Image coordinate on the position is (0,0), and the amount of movement of the XY worktable suitable with a pixel of image is (mx, my), because
In addition the coordinate of the XY worktable of spreading square central is obtained by following formula.
（Minx+mx × cx, miny+my × cy）
Moved by control instruction generating unit (not shown) to generate the above-mentioned coordinate calculated to defect coordinate operational part 72
Dynamic move, and same position control section (not shown) is output to, so as to complete the movement to defective locations.
In embodiment above, although having used the respective color signal of red, green, blue from the input of colored CCD camera,
But HSV colour systems can also be used, the HSV colour systems can be converted to by these color signals.
In HSV colour systems, V is referred to as lightness, represents the brightness of image.In rgb value, if minimum value is set into fmin, will
Maximum is set to fmax, then brightness V be expressed as V=fmax.In addition, H is referred to as form and aspect, S is referred to as chroma, is calculated by following formula.
[defect correction]
After defective locations are moved to, staff confirms the defects of monitor 40 is shown, passes through guidance panel 41, key
The operation of disk, mouse carrys out Introduced Malaria condition to be modified operation.
When being modified operation, remove defective part using cutting with laser aid 33 and be coated with using ink
Mechanism 34 is to the operation for being partially filled with ink after removing.
For staff after completing to correct operation, instruction is moved to next defective locations.And then secondary repetition is above-mentioned
Use the processing after the digital independent of CCD camera 32.If to the processing of all defect that is received from check device etc. all
It has been completed that, then a series of operation is accomplished.
[automatic laser irradiation]
The size of the defects of detected is can also correspond to, the irradiation size of adjust automatically cutting laser aid 33, is come
Irradiate laser.Thereby, it is possible to save the trouble of staff's Introduced Malaria condition, so as to which operating efficiency is improved.
Figure 17 be represent cutting start when cutting position centre coordinate and cutting at the end of cutting position center
The figure of coordinate.In Figure 17, (xA,yA) equivalent to (rx, ry).
With reference to the figure, if incision size is set into (Sx,Sy), summit A coordinate is set to (xA,yA), then when cutting starts
Open center coordinate (xst,yst) represented by following formula.
If in addition, summit D coordinate is set to (xD,yD), then the open center coordinate (x at the end of cuttinged,yed) under
Formula represents.
If the minimum overlay amount of each cutting position is set to Rx, the cutting spacing of horizontal direction is set to px, then horizontal direction
Cutting position quantity nxRepresented by following formula.
Wherein, px≠0···（F3）
px=Sx×(1.0-Rx)
Here, ceiling () is the function of " returning more than the integer in specified number, minimum ".
If the minimum overlay amount of each cutting position is set to Ry, the cutting spacing of vertical direction is set to py, then vertical direction
Cutting position quantity nyRepresented by following formula.
Wherein, py≠0···(F4)
py=Sy×(1.0-Ry)
Figure 18 is the figure for representing the cutting position during cutting operation.
If the cutting spacing of horizontal direction and vertical direction is set to pxAnd py, then cutting position (xij,yij) under
Formula represents.
xij=xst+px×i yij=yst+py×j···（F5）
pxAnd pyRepresented respectively by following formula.
Wherein, nxP when=1x=0
···（F6）
Wherein, nyP when=1y=0
[automatic ink coating]
In addition, can also calculate ink application place automatically, ink is coated with using ink coater structure 34.With laser phase
Together, the trouble of staff's Introduced Malaria condition can be saved, so as to which operating efficiency is improved.
Figure 19 is the centre coordinate of the centre coordinate for being coated with circle when representing coating beginning and the coating circle at the end of coating
Figure.In Figure 19, (xA,yA) equivalent to (rx, ry).
With reference to the figure, if coating diameter of a circle is set into D, summit A coordinate is set to (xA,yA), then when coating starts
It is coated with the centre coordinate (x of circlest,yst) represented by following formula.
If in addition, summit D coordinate is set to (xD,yD), then centre coordinate (the x of the coating circle at the end of being coated withed,yed)
Represented by following formula.
If coating spacing is set to p, the coating circle number n of horizontal directionxRepresented by following formula.
Wherein, p ≠ 0（E3）
Here, ceiling () is the function of " returning more than the integer in specified number, minimum ".
If it is set to p, the coating line number n of vertical direction by the interval that spacing is ink coating is coated withyRepresented by following formula.
Wherein, p ≠ 0（E4）
Figure 20 is the figure for the centre coordinate for representing the coating circle in ink coating action process.
If the coating spacing of horizontal direction and vertical direction is set to pxAnd py, then application place (xij,yij) under
Formula represents.
xij=xst+px×i yij=yst+py×j···（E5）
pxAnd pyRepresented respectively by following formula.
Wherein, nxP when=1x=0
···（F6）
Wherein, nyP when=1y=0
It will be understood that it is restricted content that embodiment of disclosure, which is all example in all respects and is not,.This
For the scope of invention not as shown by described above, but as shown by claim, reply thinks what is be equal with claim
In having altered in implication and scope is all contained in.
Claims (10)
1. a kind of defect detecting device, including：
Camera, camera shooting substrate simultaneously export electric signal；And
Image processing apparatus, in the image processing apparatus input have the electric signal exported by the camera,
Described image processing unit uses the respective image pair of multiple signals obtained from the electric signal from the camera
The larger contrast than in degree, portion the defects of to detect the substrate,
Described image processing unit includes：
Image comparing section, the image comparing section compare the brightness of the pixel on arbitrary check object position in described image, with
The regulation determined by the pattern spacing of the substrate and the ratio between the size of the pixel has been separated from the check object position
The position of spacing is the brightness of the pixel in peripheral part；And
Defect extraction unit, the defect extraction unit use the comparative result of the brightness obtained using described image comparing section, come
The defective part is proposed,
Described image comparing section is configured to：Pixel in the brightness of pixel in the peripheral part and the peripheral part
Adjacent or neighbouring pixel brightness between difference value when being more than defined threshold, do not perform the comparison.
2. defect detecting device as claimed in claim 1, it is characterised in that
Described image processing unit generates panoramic picture to each signal of the multiple signal, according to multiple images, and uses
Larger contrast in the contrast of the multiple respective panoramic picture of signal, portion the defects of to detect the substrate.
3. defect detecting device as claimed in claim 1, it is characterised in that
The multiple signal includes form and aspect, chroma, the respective signal of lightness.
4. a kind of defect correction device, including：
Observe the optical system of substrate；
Camera, the camera shoot the substrate and export electric signal；And
Image processing apparatus, in the image processing apparatus input have the electric signal exported by the camera, and at the image
Reason device uses pair larger in the respective picture contrast of multiple signals obtained from the electric signal from the camera
Than degree, the coordinate in portion the defects of to calculate the substrate；
Described image processing unit includes：
Image comparing section, the image comparing section compare the brightness of the pixel on arbitrary check object position in described image, with
The regulation determined by the pattern spacing of the substrate and the ratio between the size of the pixel has been separated from the check object position
The position of spacing is the brightness of the pixel in peripheral part；And
Defect extraction unit, the defect extraction unit use the comparative result of the brightness obtained using described image comparing section, come
The defective part is proposed,
Described image comparing section is configured to：Pixel in the brightness of pixel in the peripheral part and the peripheral part
Adjacent or neighbouring pixel brightness between difference value when being more than defined threshold, do not perform the comparison,
The defect correction device also includes correction mechanism, and the correction mechanism includes moving the optical system of the observation substrate
Workbench to the coordinate calculated, remove the laser aid of the defective part and to removing described lack on the substrate
Fall into portion after be partially filled with the coating mechanism of ink at least some.
5. defect correction device as claimed in claim 4, it is characterised in that
The defect correction device includes the laser aid and the coating mechanism,
According to the position of the defective part, the irradiation area of laser and the application place of ink are calculated, removes described lack automatically
The portion of falling into, and ink is filled automatically.
6. a kind of defect inspection method, including：
The step of shooting substrate and exporting electric signal；And
Use contrast larger in the respective picture contrast of multiple signals obtained from the electric signal, the detection base
The step of the defects of plate portion,
The step of the defects of detecting substrate portion, includes：
Comparison step, in the comparison step, compare pixel in described image on arbitrary check object position brightness, with from
Rise between the regulation for having separated and having been determined by the pattern spacing of the substrate and the ratio between the size of the pixel check object position
Away from position be the pixel in peripheral part brightness；
Extraction step, in the extraction step, using the comparative result of the brightness obtained by the comparison step, to propose
State defective part；And
Non-executing step, in the non-executing step, the brightness of the pixel in the peripheral part, with the peripheral part
When difference value between the brightness of the adjacent or neighbouring pixel of pixel is more than defined threshold, do not perform in the comparison step
The brightness ratio compared with.
7. defect inspection method as claimed in claim 6, it is characterised in that
The step of also including each signal to the multiple signal, generating panoramic picture according to multiple images,
The step of the defects of detecting substrate portion comprising using the respective panoramic picture of the multiple signal contrast in compared with
Big contrast, detect the step of the defects of substrate portion.
8. defect inspection method as claimed in claim 6, it is characterised in that
The multiple signal includes form and aspect, chroma, the respective signal of lightness.
9. defect inspection method as claimed in claim 6, it is characterised in that
The length that also includes being connected according to the border of the rectangle and picture external with the defective part, judge whether to detect it is whole
The step of individual defective part.
10. defect inspection method as claimed in claim 6, it is characterised in that
Also include based on the rectangular position external with defect and size, calculate laser-irradiated domain and ink application place
Step.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2012186345 | 2012-08-27 | ||
| JP2012-186345 | 2012-08-27 | ||
| JP2013-107126 | 2013-05-21 | ||
| JP2013107126A JP6184746B2 (en) | 2012-08-27 | 2013-05-21 | Defect detection apparatus, defect correction apparatus, and defect detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103630542A CN103630542A (en) | 2014-03-12 |
| CN103630542B true CN103630542B (en) | 2018-03-20 |
Family
ID=50211817
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310377121.3A Expired - Fee Related CN103630542B (en) | 2012-08-27 | 2013-08-26 | Defect detecting device, defect correction device and defect inspection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103630542B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106814479B (en) * | 2017-01-11 | 2019-07-16 | 昆山国显光电有限公司 | A kind of offset compensating method of panel defect location, apparatus and system |
| CN109507815B (en) * | 2018-12-05 | 2022-06-21 | 武汉精立电子技术有限公司 | Method for rapidly positioning defect position in backlight detection of liquid crystal display screen |
| CN109808024B (en) * | 2019-03-27 | 2023-06-23 | 南京林业大学 | Shaving board surface flatness process adjusting system and method based on two-dimensional detection |
| CN115053258A (en) * | 2020-11-30 | 2022-09-13 | 京东方科技集团股份有限公司 | Display panel detection method, device and system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2163965A1 (en) * | 1993-05-28 | 1994-12-08 | Marie Rosalie Dalziel | An automatic inspection apparatus |
| FR2846425B1 (en) * | 2002-10-25 | 2006-04-28 | Bsn Glasspack | METHOD AND DEVICE FOR DETECTING SURFACE DEFECTS PRESENTED BY THE OUTER WALL OF A TRANSPARENT OR TRANSLUCENT OBJECT |
| JP4416827B1 (en) * | 2008-09-10 | 2010-02-17 | シャープ株式会社 | Evaluation apparatus, calibration method, calibration program, and recording medium |
| CN102023164B (en) * | 2009-09-23 | 2015-09-16 | 法国圣-戈班玻璃公司 | For detecting the apparatus and method of the local defect of transparent plate |
| CN101819162A (en) * | 2010-05-13 | 2010-09-01 | 山东大学 | Empty bottle wall defect detection method and device |
-
2013
- 2013-08-26 CN CN201310377121.3A patent/CN103630542B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN103630542A (en) | 2014-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102562020B1 (en) | Substrate inspection method, substrate processing system, and computer storage medium | |
| KR102121089B1 (en) | Systems, devices and methods for the quality assessment of oled stack films | |
| JP5409677B2 (en) | IMAGE CREATION METHOD, SUBSTRATE INSPECTION METHOD, IMAGE CREATION METHOD, OR RECORDING MEDIUM RECORDING PROGRAM FOR EXECUTING SUBSTRATE INSPECTION METHOD | |
| US20140078296A1 (en) | Three-dimensional measuring apparatus | |
| CN103630542B (en) | Defect detecting device, defect correction device and defect inspection method | |
| TW201418665A (en) | Method for inspecting measurement object | |
| JP6233824B1 (en) | Image inspection apparatus, production system, image inspection method, program, and storage medium | |
| CN110853105A (en) | Checkerboard image, method and device for RGB sub-pixel simultaneous positioning and application | |
| KR101261016B1 (en) | method and apparatus for automatic optical inspection of flat panel substrate | |
| JP2009115566A (en) | Apparatus for determining detect position of panel | |
| KR20140070464A (en) | Method for inspection of defects on substrate, apparatus for inspection of defects on substrate and computer-readable recording medium | |
| JP4987323B2 (en) | Color filter defect correcting apparatus and color filter defect correcting method | |
| CN114529500A (en) | Defect inspection method for display substrate | |
| KR101068356B1 (en) | Method for inspecting defect of the Pixels in display panel device by image | |
| KR20140068543A (en) | Inspectiing device of display panel and inspecting method of the same | |
| JP4880380B2 (en) | Substrate defect correcting apparatus and substrate defect correcting method | |
| JP6184746B2 (en) | Defect detection apparatus, defect correction apparatus, and defect detection method | |
| JP2012039096A (en) | Device and method for determining presence/absence of component | |
| JP4975472B2 (en) | Color filter defect correcting apparatus and color filter defect correcting method | |
| JP2003243466A (en) | Equipment and method for inspecting pattern | |
| KR101665764B1 (en) | Drawing apparatus, substrate processing system and drawing method | |
| TWI755460B (en) | Appearance inspection method | |
| JP2007309703A (en) | Inspection method of pixel | |
| JP2008151634A (en) | Inspecting method for electro-optical device | |
| WO2013151191A1 (en) | Method and apparatus for automatic optical inspection of flat panel substrates |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180320 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |