CN102288619B - Method and system for detecting defects of three-dimensional optical film - Google Patents

Method and system for detecting defects of three-dimensional optical film Download PDF

Info

Publication number
CN102288619B
CN102288619B CN2011101842573A CN201110184257A CN102288619B CN 102288619 B CN102288619 B CN 102288619B CN 2011101842573 A CN2011101842573 A CN 2011101842573A CN 201110184257 A CN201110184257 A CN 201110184257A CN 102288619 B CN102288619 B CN 102288619B
Authority
CN
China
Prior art keywords
training
defect detection
image
zone
detection parameter
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.)
Active
Application number
CN2011101842573A
Other languages
Chinese (zh)
Other versions
CN102288619A (en
Inventor
赵新民
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
BenQ Materials Corp
Original Assignee
BenQ Materials Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by BenQ Materials Corp filed Critical BenQ Materials Corp
Priority to CN2011101842573A priority Critical patent/CN102288619B/en
Publication of CN102288619A publication Critical patent/CN102288619A/en
Application granted granted Critical
Publication of CN102288619B publication Critical patent/CN102288619B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention provides a method and a system for detecting the defects of a three-dimensional optical film. The method comprises the following steps of: receiving an image to be detected which corresponds to the three-dimensional optical film; according to at least one area in the image to be detected, acquiring a plurality of first projection amounts, in a first direction, of the area; according to the first projection amounts, which are acquired in the second step, of the area, acquiring a first defect detection parameter which corresponds to the area; according to the first defect detection parameter which corresponds to the area and is acquired in the third step, judging whether the area in the image to be detected has the defects; and outputting a judgment result which is obtained in the previous step. By the invention, the defects of the three-dimensional optical film having the textural features can be detected.

Description

The flaw detection method of three-dimensional optical film and system
Technical field
The invention relates to a kind of automatic optical detection method (Automated Optical Inspection, AOI), refer to especially flaw (defect) detection method and the system of a kind of three-dimensional optical film (3D optical film).
Background technology
Existing blooming Defect Detection technology is not have a product of texture (texture) characteristic for the surface to detect mostly.See also Figure 1A and Figure 1B, the polaroid (polarized film) of texture features of not having take the surface is example, according to the characteristic of image 11,12 on color space (color space) corresponding to polaroid difference, can directly judge whichever and have flaw; Furthermore, these small image 11 corresponding polaroids of color variations are the polaroid of indefectible (defect-free), and these image 12 corresponding polaroids that color variations is larger are tool polaroid defective.
Yet the three-dimensional optical film belongs to the product that the surface has structural texture, so can't utilize the characteristic of its corresponding image on color space to carry out Defect Detection.
Summary of the invention
Therefore, the present invention's purpose is namely at the flaw detection method that a kind of three-dimensional optical film is provided.
So the flaw detection method of three-dimensional optical film of the present invention is to utilize processing unit to cooperate image acquisition unit to carry out, this image acquisition unit is in order to the to be detected image of acquisition corresponding to three-dimensional optical film to be detected; The method comprises the following step: (A) receive this image to be detected; (B) according at least one zone in this image to be detected, try to achieve a plurality of first projection amount of this zone on first direction; These first projection amount in this zone of (C) trying to achieve according to step (B) are tried to achieve the first Defect Detection parameter corresponding to this zone; This first Defect Detection parameter corresponding to this zone of (D) trying to achieve according to step (C) judges whether this zone in this image to be detected has flaw; Reach (E) judged result of output step (D).
According to the flaw detection method of three-dimensional optical film of the present invention, step (B) comprises following substep: (b-1) should carry out binary conversion treatment by image to be detected, to obtain this image to be detected of binaryzation; And (b-2) according to this zone in this image to be detected of binaryzation, try to achieve these a plurality of first projection amount in this zone.Further, step (b-2) comprises following substep: (b-2-1) according to the suitableeest area size of trying to achieve in advance, be plurality of regions with this Image Segmentation to be detected of binaryzation; And (b-2-2) for each zone, according to a plurality of bi-values corresponding to a plurality of pixels difference in this zone, try to achieve these a plurality of first projection amount in this zone; Wherein, these a plurality of first projection amount in this zone that step (C) is tried to achieve according to step (B) are tried to achieve a plurality of the first Defect Detection parameters that correspond respectively to this zone; Wherein, these a plurality of first Defect Detection parameters that correspond respectively to this zone that step (D) is tried to achieve according to step (C) judge whether this zone in this image to be detected has flaw.Further, also in order to the first training image of acquisition corresponding to flawless three-dimensional optical film, the method also comprises step (A) the following step before to this image acquisition unit: (F) receive this first training image; (G) this first training image is carried out binary conversion treatment, to obtain this first training image of binaryzation; (H) according to this first training image of binaryzation, try to achieve at least one horizontal texture cycle, and at least one vertical texture cycle; And this horizontal texture cycle and this vertical texture cycle of (I) trying to achieve according to step (H), determine this suitableeest area size.
According to the flaw detection method of three-dimensional optical film of the present invention, step (C) is calculated the first standard deviation according to these a plurality of first projection amount in this zone, with as this first Defect Detection parameter corresponding to this zone.Further, step (B) is also tried to achieve a plurality of second projection amount of this zone on second direction, and wherein, this second direction is perpendicular to this first direction; Wherein, step (C) is also calculated the second standard deviation according to these a plurality of second projection amount in this zone, with as the second Defect Detection parameter corresponding to this zone; Wherein, this first Defect Detection parameter and this second Defect Detection parameter corresponding to this zone that step (D) is tried to achieve according to step (C) judge whether this zone in this image to be detected has flaw.Further, also in order to the second training image of acquisition corresponding to tool three-dimensional optical film defective, the method also comprises step (A) the following step before to this image acquisition unit: (J) receive this second training image; (K) this second training image is carried out binary conversion treatment, to obtain this second training image of binaryzation; (L) according to a plurality of training zone in this second training image of binaryzation, try to achieve a plurality of first training projection amount of each training zone on this first direction, and a plurality of the second training projection amount on this second direction; Each training zone of (M) trying to achieve according to step (L) should a plurality of first training projection amount and should a plurality of the second training projection amount, calculate corresponding to each and train the first regional training standard poor and the second training standard is poor; And these a plurality of first training standards that correspond respectively to this training zone of (N) trying to achieve according to step (M) are poor, and to reach these a plurality of second training standards poor, try to achieve the first Defect Detection parameter higher limit, the first Defect Detection parameter lower limit value, the second Defect Detection parameter higher limit, and the second Defect Detection parameter lower limit value.
Flaw detection method according to three-dimensional optical film of the present invention, step (N) utilizes following formula to try to achieve this first Defect Detection parameter higher limit, this first Defect Detection parameter lower limit value, this second Defect Detection parameter higher limit, and this second Defect Detection parameter lower limit value:
HP_thr u＝HP_Std avg+k×HP_Std std；
Hp_thr l＝Hp_Std avg-k×Hp_Std std；
VP_thr u=VP_Std Avg+ k * Vp_Std StdAnd
Vp_thr l＝Vp_Std avg-k×Vp_Std std；
Wherein, Hp_thr uRepresent this first Defect Detection parameter higher limit, Hp_Std AvgRepresent these a plurality of the first training standards poor mean value, Hp_Std StdRepresent these a plurality of the first training standards poor standard deviation, Hp_thr 1Show this first Defect Detection parameter lower limit value, Vp_thr uRepresent this second Defect Detection parameter higher limit, Vp_Std AvgRepresent these a plurality of the second training standards poor mean value, Vp_Std StdRepresent these a plurality of the second training standards poor standard deviation, Vp_thr lRepresent this second Defect Detection parameter lower limit value, k is a default constant.
Flaw detection method according to three-dimensional optical film of the present invention, this the first Defect Detection parameter and this second Defect Detection parameter corresponding to this zone that step (D) is tried to achieve according to step (C), and cooperate this first Defect Detection parameter higher limit, this first Defect Detection parameter lower limit value, this second Defect Detection parameter higher limit try to achieve in advance, and this second Defect Detection parameter lower limit value, whether have flaw to judge this zone in this image to be detected.
Another object of the present invention is namely providing a kind of Defect Detection system of three-dimensional optical film.
So the Defect Detection system of three-dimensional optical film of the present invention comprises image acquisition unit and processing unit.This image acquisition unit is in order to the to be detected image of acquisition corresponding to three-dimensional optical film to be detected.This processing unit is in order to carry out: receive this image to be detected; According at least one zone in this image to be detected, try to achieve a plurality of first projection amount of this zone on first direction; According to these first projection amount in this zone, try to achieve the first Defect Detection parameter corresponding to this zone; And according to this first Defect Detection parameter corresponding to this zone, judge whether this zone in this image to be detected has flaw, and output judged result.
Effect of the present invention is: by this first Defect Detection parameter of trying to achieve according to these a plurality of first projection amount in this zone in the image to be detected, can carry out Defect Detection to the three-dimensional optical film with texture features.
Description of drawings
Figure 1A and Figure 1B are Photograph image figure, and a flawless polaroid is described, and a tool polaroid defective, and both distinguish corresponding image;
Fig. 2 is a calcspar, and a preferred embodiment of the Defect Detection system of three-dimensional optical film of the present invention is described;
Fig. 3 A and Fig. 3 B are Photograph image figure, and one first training image corresponding to a flawless three-dimensional optical film is described, reach one second training image corresponding to a tool three-dimensional optical film defective;
Fig. 4 is a process flow diagram, illustrates in the system initialisation phase of a preferred embodiment of the flaw detection method of three-dimensional optical film of the present invention, is used for trying to achieve the step of suitable area size;
Fig. 5 A, Fig. 5 B and Fig. 5 C are synoptic diagram, and a binary example image, first curve map and second curve map are described;
Fig. 6 is a process flow diagram, explanation is in this system initialisation phase, be used for trying to achieve first a Defect Detection parameter higher limit, first a Defect Detection parameter lower limit value, second a Defect Detection parameter higher limit, and the step of second a Defect Detection parameter lower limit value; And
Fig. 7 is a process flow diagram, and the included step of system's detection-phase of this preferred embodiment of the flaw detection method of three-dimensional optical film of the present invention is described.
Embodiment
For making purpose of the present invention, structure, feature and function thereof there are further understanding, hereby cooperate embodiment to be described in detail as follows.
Consult Fig. 2, the preferred embodiment of the Defect Detection system 2 of three-dimensional optical film of the present invention comprises optical lens 21, image acquisition unit 22, and processing unit 23.Wherein, this image acquisition unit 22 is in order to the image of acquisition corresponding to a three-dimensional optical film under this optical lens 21; This processing unit 23 is in order to carry out Defect Detection according to this image corresponding to this three-dimensional optical film.
Below cooperate a preferred embodiment of the flaw detection method of three-dimensional optical film of the present invention, the step that this processing unit 23 is performed is described.Wherein, the flaw detection method of this three-dimensional optical film is divided into system initialisation phase, and system's detection-phase.
[system initialisation phase]
At first, consult Fig. 2～Fig. 5 C, these image acquisition unit 22 acquisitions are corresponding to first training (training) image 31 of flawless three-dimensional optical film, and wherein, this first training image 31 has structural texture features as shown in Figure 3A.
Then, this processing unit 23 is carried out following steps, to obtain an area size the suitableeest (size).
In step 41, this processing unit 23 receives this first training image 31.
In step 42, this processing unit 23 carries out binary conversion treatment (thresholding) with this first training image 31, to obtain this first training image 31 of binaryzation, wherein, this of binaryzation the first training image 31 is a binary (binary) image.
In step 43, this processing unit 23 is tried to achieve at least one horizontal texture cycle according to this first training image 31 of binaryzation, and at least one vertical texture cycle.
This horizontal texture cycle and this vertical texture cycle cooperate a binary example image 51 shown in Fig. 5 A to further specify as rear; Wherein, this binary example image 51 comprises that its bi-values is a plurality of pixels (pixel) 511 of 0, and bi-values is a plurality of pixels 512 of 1.In this preferred embodiment, this processing unit 23 is according to the result of calculation of formula (1)～(2), to obtain this horizontal texture cycle and this vertical texture cycle.
Make the size of this binary example image 51 represent with m * n; And make each pixel 511 in this binary example image 51,512 two-dimentional coordinate with (x, y) expression wherein, the two-dimentional coordinate of the pixel 511 in the upper left corner is (1,1) in this binary example image 51, the two-dimentional coordinate of the pixel 511 of last cell is (m, n); P (x, y) represents its two-dimentional coordinate and is the pixel 511 of (x, y), 512 bi-values.
SUM H _ diff ( i ) = Σ x = 1 m - i Σ y = 1 n | P ( x + i , y ) - P ( x , y ) | - - - ( 1 )
SUM V _ diff ( j ) = Σ x = 1 m Σ y = 1 n - j | P ( x , y + j ) - P ( x , y ) | - - - ( 2 )
Wherein, i=1,2, K, m-1, j=1,2, K, n-1.
For convenience of explanation, the result of calculation of formula (1)～(2) is illustrated with first curve Figure 52 (such as Fig. 5 B) and second curve Figure 53 (such as Fig. 5 C) respectively; Figure 52 can learn SUM by this first curve H_diff(4)=SUM H_diff(8)=SUM H_diff(12)=0, its corresponding physical significance is: this binary example image 51 repetition period of texture on horizontal direction is 4,8,12; Similarly, can learn SUM by this second curve Figure 53 V_diff(4)=SUM V_diff(8)=SUM V_diff(12)=0, its corresponding physical significance is: this binary example image 51 repetition period of texture on vertical direction is 4,8,12.Make the horizontal texture cycle represent with H_period, the vertical texture cycle represents with V_period, then in this example, obtains three horizontal texture cycles and three vertical texture cycles; That is, H_period=4,8,12, V_period=4,8,12.
More broad sense is sayed it, as long as find out respectively the trough among this first curve Figure 52 and this second curve Figure 53, can obtain horizontal texture cycle (H_period) and vertical texture cycle (V_period); In other words, according to the result of calculation of formula (1)～(2), cooperate default first threshold value (with thr 1The expression) and second threshold value (with thr 2Expression), can obtain H_period and V_period; Wherein, SUM H_diff(H_period) and thr 1, and SUM V_diff(V_period) and thr 2Relational expression further expression such as formula (3)～(4).
SUM H_diff(H_period)＜thr 1 (3)
SUM V_diff(V_period)＜thr 2 (4)
In step 44, this processing unit 23 determines this suitableeest area size according to this horizontal texture cycle and this vertical texture cycle.Make this suitableeest area size represent with w * h; In this preferred embodiment, be this horizontal texture cycle with minimum, and this minimum vertical texture cycle is decided this suitableeest area size; That is, w=min (H_period), h=min (V_period).
Then, consult Fig. 2, Fig. 3 B and Fig. 6, these image acquisition unit 22 acquisitions are corresponding to the second training image 32 of tool three-dimensional optical film defective, this processing unit 23 is then carried out following steps, to obtain the first Defect Detection parameter higher limit, the first Defect Detection parameter lower limit value, the second Defect Detection parameter higher limit, reach the second Defect Detection parameter lower limit value.
In step 61, this processing unit 23 receives this second training image 32.
In step 62, this processing unit 23 carries out binary conversion treatment with this second training image 32, to obtain this second training image 32 of binaryzation.
In step 63, this processing unit 23 according to this suitableeest area size of trying to achieve in advance (w * h), it is regional that this second training image 32 of binaryzation is divided into a plurality of training; Wherein, the size in each training zone equals this suitableeest area size.
In step 64, this processing unit 23 is tried to achieve a plurality of first training projection amount of each training zone on first direction according to these training zones in this second training image 32 of binaryzation, and a plurality of the second training projection amount on second direction.In this preferred embodiment, this first direction is horizontal direction, and this second direction is vertical direction, and this second direction is perpendicular to this first direction; These first and second training projection amount in each training zone utilizes formula (5)～(6) to calculate.
Hp ( y ) = Σ x = 1 w P ( x , y ) - - - ( 5 )
Vp ( x ) = Σ y = 1 h P ( x , y ) - - - ( 6 )
Wherein, for each training zone, Hp (y) represents these the first training projection amount in this training zone, and Vp (x) represents these the second training projection amount in this training zone; Y=1,2, K, h, x=1,2, K, w; P (x, y) representative is in this training zone, and its two-dimentional coordinate is the bi-values of the pixel of (x, y).
In step 65, this processing unit 23 is according to these first training projection amount and these the second training projection amount in each training zone, calculates poor corresponding to first training standard in each training zone and the second training standard is poor; It calculates shown in formula (7)～(8).
Hp _ Std = 1 h Σ y = 1 h ( Hp ( y ) - Hp avg ) 2 - - - ( 7 )
Vp _ Std = 1 w Σ x = 1 w ( Vp ( x ) - Vp avg ) 2 - - - ( 8 )
Wherein, for each training zone, the Hp_Std representative is poor corresponding to this first training standard in this training zone, and the Vp_Std representative is poor corresponding to this second training standard in this training zone; Hp AvgRepresent the mean value of these the first training projection amount in this training zone, Vp AvgRepresent the mean value of these the second training projection amount in this training zone.
In step 66, these first training standards that these processing unit 23 bases correspond respectively to these training zones are poor and these second training standards are poor, try to achieve this first Defect Detection parameter upper limit value and lower limit value, and this second Defect Detection parameter upper limit value and lower limit value; It calculates shown in formula (9)～(12).
Hp_thr u＝Hp_Std avg+k×Hp_Std std (9)
Hp_thr l＝HP_Std avg-k×HP_Std std (10)
Vp_thr u＝Vp_Std avg+k×Vp_Std std (11)
Vp_thr l＝Vp_Std avg-k×Vp_Std std (12)
Wherein, Hp_thr uRepresent this first Defect Detection parameter higher limit, Hp_Std AvgRepresent the poor mean value of these the first training standards, Hp_Std StdRepresent the poor standard deviation of these the first training standards, Hp_thr lRepresent this first Defect Detection parameter lower limit value, Vp_thr uRepresent this second Defect Detection parameter higher limit, Vp_Std AvgRepresent the poor mean value of these the second training standards, Vp_Std StdRepresent the poor standard deviation of these the second training standards, Vp_thr lRepresent this second Defect Detection parameter lower limit value, k is a default constant.In this preferred embodiment, k=3.
[system's detection-phase]
Consult Fig. 2 and Fig. 7, after having carried out system initialisation phase, this suitableeest area size that can will try to achieve in advance (w * h), this first Defect Detection parameter upper limit value and lower limit value (Hp_thr u, Hp_thr l), and this second Defect Detection parameter upper limit value and lower limit value (Vp_thr u, Vp_thr l) be applied to the Defect Detection of three-dimensional optical film.
This image acquisition unit 22 continues acquisition corresponding to the image to be detected of three-dimensional optical film to be detected.
In step 71, this processing unit 23 receives image to be detected.
In step 72, this processing unit 23 should carry out binary conversion treatment by image to be detected, to obtain this image to be detected of binaryzation.
In step 73, (w * h) is plurality of regions with this Image Segmentation to be detected of binaryzation to this processing unit 23 according to this suitableeest area size; Wherein, the size in each zone equals this suitableeest area size.
In step 74, this processing unit 23 is according to these zones in this image to be detected of binaryzation, try to achieve a plurality of first projection amount of each zone on this first direction (horizontal direction), and a plurality of the second projection amount on this second direction (vertical direction); Its compute classes is similar to formula (5)～(6), so repeat no more.
In step 75, this processing unit 23 is according to these first projection amount in each zone, and these second projection amount, calculates the first Defect Detection parameter corresponding to each zone (with H dExpression), reach the second Defect Detection parameter (with V dExpression); In this preferred embodiment, this processing unit 23 calculates the first standard deviation as this first Defect Detection parameter corresponding to this zone according to these first projection amount in each zone, and calculate the second standard deviation as corresponding to this second Defect Detection parameter in this zone according to these second projection amount in each zone, its compute classes is similar to formula (7)～(8), so repeat no more.
In step 76, this processing unit 23 is according to this first Defect Detection parameter and this second Defect Detection parameter corresponding to each zone, and cooperates this first Defect Detection parameter upper limit value and lower limit value (Hp_thr u, Hp_thr 1), and this second Defect Detection parameter upper limit value and lower limit value (Vp_thr u, Vp_thr l), whether have flaw to judge each zone in this image to be detected.Wherein, for each zone, if its first Defect Detection parameter and the second Defect Detection parameter both wherein any one does not meet relational expression (11)～(12), then represent this zone and have flaw; Otherwise it is indefectible representing this zone.
Hp_thr l＜H d＜Hp_thr u (11)
Vp_thr l＜V d＜Vp_thr u (12)
In step 77, the judged result of this processing unit 23 output steps 76, that is, and about these zones tool testing result defective whether; Then, get back to step 71 and continue reception corresponding to the image next to be detected of next (next) three-dimensional optical film to be detected.
It is worth mentioning that, by the judged result of step 76, can learn the ratio of defect areas in the three-dimensional optical film to be detected; Moreover, because the decorating position of this optical lens 21 of the Defect Detection system 2 of this three-dimensional optical film is known, so this image to be detected also is known with the geometrical correspondence of this three-dimensional optical film to be detected, therefore, by these regional testing results in this image to be detected, can corresponding learn the relevant position that flaw occurs in this three-dimensional optical film to be detected.
In sum, by these first and second Defect Detection parameters that correspond respectively to these zones in this image to be detected of receiving corresponding to this three-dimensional optical film to be detected, can carry out Defect Detection to the three-dimensional optical film with texture features; Further, cooperate this suitableeest area size, this first Defect Detection parameter upper limit value and lower limit value of trying to achieve in advance, and this second Defect Detection parameter upper limit value and lower limit value, can carry out full automatic Defect Detection to the three-dimensional optical film with texture features, so really can reach purpose of the present invention.
The present invention is described by above-mentioned related embodiment, yet above-described embodiment is only for implementing example of the present invention.Must be pointed out that the embodiment that has disclosed does not limit the scope of the invention.On the contrary, the change of doing without departing from the spirit and scope of the present invention and retouching all belong to scope of patent protection of the present invention.

Claims (5)

1. the flaw detection method of a three-dimensional optical film, utilize processing unit to cooperate image acquisition unit to carry out, this image acquisition unit is in order to the to be detected image of acquisition corresponding to three-dimensional optical film to be detected, also train image in order to acquisition corresponding to the first training image of flawless three-dimensional optical film with corresponding to second of tool three-dimensional optical film defective, it is characterized in that the method comprises the following step:
(F) receive this first training image;
(G) this first training image is carried out binary conversion treatment, to obtain this first training image of binaryzation;
(H) according to this first training image of binaryzation, try to achieve at least one horizontal texture cycle, and at least one vertical texture cycle;
This horizontal texture cycle and this vertical texture cycle of (I) trying to achieve according to step (H), determine the suitableeest area size;
(J) receive this second training image;
(K) this second training image is carried out binary conversion treatment, to obtain this second training image of binaryzation;
(L) according to this suitableeest area size, be that a plurality of training are regional with this second training Image Segmentation of binaryzation; According to this second training image of binaryzation should be a plurality of training regional, try to achieve a plurality of first training projection amount of each training zone on first direction, and a plurality of the second training projection amount on second direction, wherein, this second direction is perpendicular to this first direction;
Each training zone of (M) trying to achieve according to step (L) should a plurality of first training projection amount and should a plurality of the second training projection amount, calculate corresponding to each and train the first regional training standard poor and the second training standard is poor;
This a plurality of first training standards that correspond respectively to this training zone of (N) trying to achieve according to step (M) are poor, and to reach these a plurality of second training standards poor, try to achieve the first Defect Detection parameter higher limit, the first Defect Detection parameter lower limit value, the second Defect Detection parameter higher limit, and the second Defect Detection parameter lower limit value;
(A) receive this image to be detected;
(B) should carry out binary conversion treatment by image to be detected, to obtain this image to be detected of binaryzation; According to this suitableeest area size, be plurality of regions with this Image Segmentation to be detected of binaryzation; According at least one zone in this image to be detected, try to achieve a plurality of first projection amount of this zone on this first direction, and a plurality of the second projection amount on this second direction;
These a plurality of first projection amount in this zone of (C) trying to achieve according to step (B) and these a plurality of second projection amount are tried to achieve the first Defect Detection parameter and the second Defect Detection parameter corresponding to this zone;
This first Defect Detection parameter and this second Defect Detection parameter corresponding to this zone of (D) trying to achieve according to step (C), and cooperate this first Defect Detection parameter higher limit, this first Defect Detection parameter lower limit value, this second Defect Detection parameter higher limit and this second Defect Detection parameter lower limit value, judge whether this zone in this image to be detected has flaw; And
(E) judged result of output step (D).
2. the flaw detection method of three-dimensional optical film as claimed in claim 1 is characterized in that, step (B) comprises following substep:
For each zone, according to a plurality of bi-values corresponding to a plurality of pixels difference in this zone, try to achieve these a plurality of first projection amount and these a plurality of second projection amount in this zone;
Wherein, step (C) reaches these a plurality of second projection amount according to these a plurality of first projection amount in this zone that step (B) is tried to achieve, and tries to achieve a plurality of the first Defect Detection parameters and a plurality of the second Defect Detection parameter that correspond respectively to this zone;
Wherein, these a plurality of first Defect Detection parameters that correspond respectively to this zone and these a plurality of second Defect Detection parameters that step (D) is tried to achieve according to step (C) judge whether this zone in this image to be detected has flaw.
3. the flaw detection method of three-dimensional optical film as claimed in claim 1 is characterized in that, step (C) is calculated the first standard deviation according to these a plurality of first projection amount in this zone, with as this first Defect Detection parameter corresponding to this zone; Step (C) is also calculated the second standard deviation according to these a plurality of second projection amount in this zone, with as the second Defect Detection parameter corresponding to this zone.
4. the flaw detection method of three-dimensional optical film as claimed in claim 1, it is characterized in that, step (N) utilizes following formula to try to achieve this first Defect Detection parameter higher limit, this first Defect Detection parameter lower limit value, this second Defect Detection parameter higher limit, and this second Defect Detection parameter lower limit value:
Hp_thr u=Hp_Std avg+k×Hp_Std std；
Hp_thr l=Hp_Std avg-k×Hp_Std std；
Vp_thr u=Vp_Std Avg+ k * Vp_Std StdAnd
Vp_thr l=Vp_Std avg-k×Vp_Std std；
Wherein, Hp_thr uRepresent this first Defect Detection parameter higher limit, Hp_Std AvgRepresent these a plurality of the first training standards poor mean value, Hp_Std StdRepresent these a plurality of the first training standards poor standard deviation, Hp_thr lRepresent this first Defect Detection parameter lower limit value, Vp_thr uRepresent this second Defect Detection parameter higher limit, Vp_Std AvgRepresent these a plurality of the second training standards poor mean value, Vp_Std StdRepresent these a plurality of the second training standards poor standard deviation, Vp_thr lRepresent this second Defect Detection parameter lower limit value, k is a default constant.
5. the Defect Detection system of a three-dimensional optical film, it is characterized in that comprising image acquisition unit and processing unit, this image acquisition unit is in order to the to be detected image of acquisition corresponding to three-dimensional optical film to be detected, also train image in order to acquisition corresponding to the first training image of flawless three-dimensional optical film with corresponding to second of tool three-dimensional optical film defective, this processing unit is in order to carry out:
Receive this first training image;
This first training image is carried out binary conversion treatment, to obtain this first training image of binaryzation;
According to this first training image of binaryzation, try to achieve at least one horizontal texture cycle, and at least one vertical texture cycle;
According to this horizontal texture cycle of trying to achieve and this vertical texture cycle, determine the suitableeest area size;
Receive this second training image;
This second training image is carried out binary conversion treatment, to obtain this second training image of binaryzation;
According to this suitableeest area size, be that a plurality of training are regional with this second training Image Segmentation of binaryzation;
According to this second training image of binaryzation should be a plurality of training regional, try to achieve a plurality of first training projection amount of each training zone on first direction, and a plurality of the second training projection amount on second direction, wherein, this second direction is perpendicular to this first direction;
According to each training zone of trying to achieve should a plurality of first training projection amount and should a plurality of the second training projection amount, calculate corresponding to each and train the first regional training standard poor and the second training standard is poor;
Poor to reach these a plurality of second training standards poor according to this a plurality of first training standards that correspond respectively to this training zone of trying to achieve, try to achieve the first Defect Detection parameter higher limit, the first Defect Detection parameter lower limit value, the second Defect Detection parameter higher limit, and the second Defect Detection parameter lower limit value;
Receive this image to be detected;
Should carry out binary conversion treatment by image to be detected, to obtain this image to be detected of binaryzation;
According to this suitableeest area size, be plurality of regions with this Image Segmentation to be detected of binaryzation;
According at least one zone in this image to be detected, try to achieve a plurality of first projection amount of this zone on this first direction, and a plurality of the second projection amount on this second direction;
According to these a plurality of first projection amount and these a plurality of second projection amount in this zone, try to achieve the first Defect Detection parameter and the second Defect Detection parameter corresponding to this zone; And
According to this first Defect Detection parameter and this second Defect Detection parameter corresponding to this zone, and cooperate this first Defect Detection parameter higher limit, this first Defect Detection parameter lower limit value, this second Defect Detection parameter higher limit and this second Defect Detection parameter lower limit value, judge whether this zone in this image to be detected has flaw, and the output judged result.
CN2011101842573A 2011-07-01 2011-07-01 Method and system for detecting defects of three-dimensional optical film Active CN102288619B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2011101842573A CN102288619B (en) 2011-07-01 2011-07-01 Method and system for detecting defects of three-dimensional optical film

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2011101842573A CN102288619B (en) 2011-07-01 2011-07-01 Method and system for detecting defects of three-dimensional optical film

Publications (2)

Publication Number Publication Date
CN102288619A CN102288619A (en) 2011-12-21
CN102288619B true CN102288619B (en) 2013-04-24

Family

ID=45335214

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2011101842573A Active CN102288619B (en) 2011-07-01 2011-07-01 Method and system for detecting defects of three-dimensional optical film

Country Status (1)

Country Link
CN (1) CN102288619B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101697071B1 (en) * 2014-04-18 2017-01-17 동우 화인켐 주식회사 Method for discriminating defect of polarizing plate
TWI611901B (en) * 2016-01-27 2018-01-21 高瞻資訊股份有限公司 Inspecting method of inspecting status of mold
TWI608369B (en) 2016-11-23 2017-12-11 財團法人工業技術研究院 Classification method, classification module and computer program product
CN108459023A (en) * 2018-03-27 2018-08-28 松下电子部品（江门）有限公司 Biradical pseudo-capacitance appearance images detection method
CN114782416B (en) * 2022-06-16 2023-09-05 济宁爱丝制衣股份有限公司 Textile quality detection method and system based on image recognition

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1737545A (en) * 2004-08-06 2006-02-22 株式会社巴川制纸所 Test parts of optic membrane and test method for optic membrane
CN101339140A (en) * 2007-07-02 2009-01-07 台达电子工业股份有限公司 Object flaw detection device and method, detection condition adjustment method
CN101510258A (en) * 2009-01-16 2009-08-19 北京中星微电子有限公司 Certificate verification method, system and certificate verification terminal
CN102012376A (en) * 2009-09-04 2011-04-13 索尼公司 Defect inspection apparatus and defect inspection method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050259854A1 (en) * 2004-05-21 2005-11-24 University Of Chicago Method for detection of abnormalities in three-dimensional imaging data

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1737545A (en) * 2004-08-06 2006-02-22 株式会社巴川制纸所 Test parts of optic membrane and test method for optic membrane
CN101339140A (en) * 2007-07-02 2009-01-07 台达电子工业股份有限公司 Object flaw detection device and method, detection condition adjustment method
CN101510258A (en) * 2009-01-16 2009-08-19 北京中星微电子有限公司 Certificate verification method, system and certificate verification terminal
CN102012376A (en) * 2009-09-04 2011-04-13 索尼公司 Defect inspection apparatus and defect inspection method

Also Published As

Publication number Publication date
CN102288619A (en) 2011-12-21

Similar Documents

Publication Publication Date Title
CN102288619B (en) Method and system for detecting defects of three-dimensional optical film
CN102175700B (en) Method for detecting welding seam segmentation and defects of digital X-ray images
CN100587717C (en) Medical large transfusion machine vision on-line detection method
CN104978748B (en) A kind of liquid crystal display defect inspection method based on local pixel value
CN109900711A (en) Workpiece, defect detection method based on machine vision
CN103198476B (en) Image detection method of thick line type cross ring mark
CN109752392B (en) PCB defect type detection system and method
CN104574393A (en) Three-dimensional pavement crack image generation system and method
CN108956420B (en) Concrete pore feature extraction method and system
CN113393464B (en) Three-dimensional detection method for plate glass defects
EP3857874B1 (en) Hybrid depth processing
CN110503638B (en) Spiral adhesive quality online detection method
CN113160161B (en) Method and device for detecting defects at edge of target
CN106327464A (en) Edge detection method
CN116165216B (en) Liquid crystal display micro scratch flaw 3D detection method, system and computing equipment
CN103914829B (en) Method for detecting edge of noisy image
CN108871185B (en) Method, device and equipment for detecting parts and computer readable storage medium
CN116026843A (en) Battery appearance defect detection method, device and detection system
CN110189375A (en) A kind of images steganalysis method based on monocular vision measurement
CN101447077A (en) Edge detection method of color textile texture image oriented to textile industry
CN106709952A (en) Automatic calibration method of display screen
CN108492306A (en) A kind of X-type Angular Point Extracting Method based on image outline
CN103630070A (en) Detection method for image detector and image detector
CN106526651B (en) The method for building up and system of a kind of detector crystal position table
CN104537627A (en) Depth image post-processing method

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