JP3985891B2 - Image defect detection apparatus and image defect detection method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image defect detection apparatus, and more particularly to an image defect detection apparatus that sets a threshold value for detecting a defect in an image displayed on a display device and detects an image defect with high accuracy.
[0002]
[Prior art]
A display device that requires uniform transmission characteristics, such as a display device using a liquid crystal panel, inspects its display image for defects in its production process. Such defect inspection is mainly performed by visual inspection by an operator. Since the visual inspection is easily affected by environmental differences and individual differences, it is difficult to obtain an appropriate inspection result.
[0003]
Against this background, the present inventors specialize in an image defect detection device that automatically detects defects in a display image in order to automate the defect inspection for detecting defects in the display image from visual inspection to automatic inspection. This is proposed in Japanese Patent Application No. 2000-107015. This image defect detection apparatus digitizes pixel data representing pixels constituting an image, calculates a comparison value based on the digitized pixel data, and based on the calculated comparison value and a preset threshold value Automatically detect defects in the displayed image.
[0004]
[Problems to be solved by the invention]
In order to automatically detect a defect in a display image by such an image defect detection device, a threshold value for detecting a defect in the display image must be set in advance. The threshold value that should be set in advance is determined by a skilled person viewing a large amount of image data collected at the production site.
[0005]
However, it takes a lot of time and labor to collect a large amount of image data at the production site, and it also takes a lot of time to set a threshold by visually checking a large amount of collected image data. And labor are necessary.
[0006]
The present invention has been made to solve such a problem, and an object of the present invention is to provide an image defect detection apparatus and an image defect detection capable of automatically setting a threshold for detecting a defect in a display image. It is to provide a method.
[0007]
[Means for Solving the Problems]
An image defect detection apparatus according to the present invention includes an imaging unit that captures an image composed of a plurality of pixels, N detection areas for detecting defects in the image (N is an integer of 2 or more), Area setting means for setting N reference areas respectively corresponding to the N detection areas in the image; Said A plurality of pixel display data for respectively displaying a plurality of pixels in each detection region; Said Corresponding to each detection area Said Based on a plurality of pixel display data for displaying each of a plurality of pixels in the reference region, Said A comparison value calculating means for calculating a comparison value for each detection region; a rearranging means for rearranging the N comparison values calculated by the comparison value calculating means in the order of the values; and The difference between the J-th comparison value and the (J + w) -th comparison value from the smallest comparison value among the N comparison values rearranged (J and w are 1 ≦ J ≦ N− w, an integer satisfying 1 ≦ w ≦ N), and a threshold value for setting a threshold value for determining the presence / absence of a defect in the image based on a predetermined threshold setting reference value Setting means; and defect detection means for detecting defects in the image based on the threshold value and the N comparison values; In the threshold value setting means, the difference between the J-th comparison value and the (J + w) -th comparison value from the smaller value of the comparison value is larger than the predetermined threshold setting reference value. When the (J + w) th comparison value is set as the threshold value And the above-mentioned object is achieved.
[0008]
The sorting means may sort the N comparison values in ascending order.
[0010]
When the difference between the J-th comparison value and the (J + w) -th comparison value is not greater than the predetermined threshold setting reference value, the rearranging unit is configured to input data input by a user. The threshold value may be set based on
[0011]
The image defect detection device detects defects of a plurality of types of images, and the region setting unit sets the N detection regions and the N reference regions for each of the defects of the plurality of types of images, The comparison value calculation unit calculates the N comparison values for each defect of the plurality of types of images, and the rearrangement unit rearranges the N comparison values for each defect of the plurality of types of images. The threshold value setting means sets the threshold value for each of the plurality of types of image defects, and the defect detection means sets the threshold value set for each of the plurality of types of image defects. The image defect may be detected for each of the plurality of types of image defects based on the N comparison values calculated for each of the plurality of types of image defects.
[0012]
The plurality of types of image defects may include point defects, spot defects, line defects, and unevenness defects.
[0013]
The image has color components of a hue system (red, green, blue, etc.), the comparison value calculation means calculates the N comparison values for each color component, and the rearrangement means The N comparison values are rearranged for each color component, the threshold value setting means sets the threshold value for each color component, and the defect detection means detects a defect in the image for each color component. May be.
[0014]
An image defect detection method according to the present invention includes an imaging step of capturing an image composed of a plurality of pixels, N detection areas for detecting defects in the image (N is an integer of 2 or more), An area setting step for setting N reference areas respectively corresponding to the N detection areas; Said A plurality of pixel display data for respectively displaying a plurality of pixels in each detection target region; Said Corresponding to each detection target area Said Based on a plurality of pixel display data for displaying each of a plurality of pixels in the reference region, Said A comparison value calculation step for calculating a comparison value for each detection region; and a rearrangement step for rearranging the N comparison values calculated by the comparison value calculation step in the order of the values of the N comparison values. The difference between the J-th comparison value and the (J + w) -th comparison value from the smaller of the comparison value values among the N comparison values sorted in the sorting step (J and w are 1 ≦ J ≦ N−w, integers satisfying 1 ≦ w ≦ N), and a threshold value for determining the presence or absence of defects in the image based on a predetermined threshold setting reference value Based on the threshold value setting step for setting the threshold value, the threshold value set in the threshold value setting step, and the N comparison values calculated in the comparison value calculation step. Includes defect detection process to detect In the threshold value setting step, the difference between the J-th comparison value and the (J + w) -th comparison value from the smaller value of the comparison value is determined based on the predetermined threshold value setting reference value. Is also larger, the (J + w) th comparison value is set as the threshold value. And the above-mentioned object is achieved.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
An image defect detection apparatus according to this embodiment will be described. The image defect detection apparatus according to the present embodiment automatically sets a threshold value for detecting defects in an image displayed by a display apparatus that requires uniform transmission characteristics such as a display screen of a liquid crystal panel. Set to.
[0016]
FIG. 1 is a block diagram of an image defect detection apparatus 100 according to the present embodiment. The image defect detection device 100 detects a defect in an image displayed on the display screen of the liquid crystal panel. The image defect detection apparatus 100 includes an imaging unit 3 having a lens 1 and a camera 2. The imaging unit 3 captures an image displayed on the display screen of the liquid crystal panel 25 with the camera 2 through the lens 1.
[0017]
An AD converter 4 is connected to the imaging unit 3. The AD converter 4 performs analog-digital (AD) conversion on image data representing an image captured by the imaging unit 3. The output of the AD converter 4 is given to the memory 5, and the memory 5 stores the image data AD-converted by the AD converter 4.
[0018]
The image data stored in the memory 5 is separated for each color component of red (R), green (G) and blue (B) by the RGB separator 6. The image data separated by the RGB separator 6 is stored in the memory 7 for each of R, G, and B components. Here, although R, G and B are separated by the RGB separator 6, they may be separated into other hue systems.
[0019]
The image defect detection device 100 includes an area setting unit 17. The area setting unit 17 includes N detection areas (N is an integer of 2 or more) for detecting defects in the image captured by the imaging unit 3, and N references corresponding to the N detection areas, respectively. The area is set as the imaging area of the image captured by the imaging unit 3.
[0020]
The image defect detection apparatus 100 is provided with an inspection unit 8. FIG. 2 is a block diagram of the inspection unit 8. The inspection unit 8 includes an area storage unit 18, a comparison value calculation unit 19, and a defect detection unit 20. The area storage unit 18 includes position data representing the positions of the N detection areas set by the area setting unit 17 and position data representing the positions of the N reference areas respectively corresponding to the N detection areas. Stored. The comparison value calculation unit 19 includes a plurality of pixel data in each detection region stored in the region storage unit 18 among a plurality of pixel data constituting the image data stored in the memory 7 and a reference corresponding to each detection region. A comparison value is calculated for each detection region based on the pixel data in the region. The defect detection unit 20 detects defects in the image captured by the imaging unit 3 based on the N comparison values calculated by the comparison value calculation unit 19 and a threshold value that is automatically set as will be described later. Is detected.
[0021]
The image defect detection apparatus 100 includes a rearrangement unit 10. The rearrangement unit 10 rearranges the N comparison values calculated by the comparison value calculation unit 19 in the order of the values of the N comparison values.
[0022]
The image defect detection apparatus 100 is provided with a threshold setting unit 12. The threshold setting unit 12 automatically sets a threshold for determining the presence or absence of a defect in the image captured by the imaging unit 3. The threshold setting unit 12 includes a distribution determination unit 11. The distribution determination unit 11 calculates the difference between the J-th comparison value and the (J + w) -th comparison value from the smallest comparison value among the N comparison values sorted by the sorting unit 10 (J and w is an integer satisfying 1 ≦ J ≦ N−w and 1 ≦ w ≦ N, respectively, and a distribution of N comparison values rearranged based on a predetermined threshold setting reference value When the state is determined and it is determined that the state is a predetermined distribution state, the (J + w) th comparison value is set as a threshold value. The threshold value setting unit 12 is provided with a threshold value register 13, and the threshold value register 13 stores the threshold value set by the distribution determining unit 11.
[0023]
The image defect detection apparatus 100 includes a result output unit 15 and a threshold value input unit 16. The result output unit 15 displays the distribution state of the N comparison values when the distribution determination unit 11 determines that the distribution state of the N comparison values is not the predetermined distribution state. The threshold value input unit 16 determines the threshold value when the distribution state of the N comparison values is determined not to be a predetermined distribution state and the distribution state of the N comparison values is displayed on the result output unit 15. Operated to input. The image defect detection apparatus 100 is provided with a central processing unit (CPU) 14 that controls the overall operation of the image defect detection apparatus 100.
[0024]
FIG. 3 is an explanatory diagram of image defects detected by the image defect detection apparatus 100. The image defect detection apparatus 100 can detect a plurality of types of defects. The plurality of types of defects include a point defect 21, a spot defect 22, a line defect 23, and a mura defect 24. The point defect 21 is a defect in which the value of pixel display data for displaying one pixel is extremely larger (or smaller) than the value of pixel display data representing the surrounding pixels. The spot defect 22 is a value of pixel display data for displaying a plurality of pixels adjacent in a planar shape along the horizontal direction and the vertical direction from the value of pixel display data for displaying the surrounding pixels. Although it is not as extreme as the point defect 21, it is a considerably large (or small) defect. The line defect 23, like the spot defect 22, is not as extreme as the point defect 21, but has a value that is significantly larger (or smaller) than the value of the pixel display data for displaying the surrounding pixels. It is assumed that a plurality of pixels represented by data are adjacent to each other in a line along a predetermined direction. In the example illustrated in FIG. 3, a line defect 23 in which a plurality of pixels are linearly adjacent along the horizontal direction and a line defect 23 where the pixels are linearly adjacent along the vertical direction are illustrated. The unevenness defect 24 is a defect in which a plurality of pixels whose difference from the pixel display data value for displaying the surrounding pixels is smaller than the spot defect 22 are adjacent over a wider range than the spot defect 22. To do.
[0025]
FIG. 4 is a flowchart illustrating an image defect detection method performed by the image defect detection apparatus 100 according to the embodiment.
[0026]
First, N detection areas for detecting image defects and N reference areas respectively corresponding to the N detection areas are set as imaging areas by operating the area setting unit 17, and the inspection unit 8 Is stored in the area storage unit 18 provided in.
[0027]
N detection areas and N reference areas are set for each of R, G, and B color components. The N detection areas and the N reference areas set for each of the R, G, and B color components are further set according to the type of defect detected by the image defect detection apparatus 100.
[0028]
FIG. 5 is a diagram for explaining a method of setting the detection area 26 and the reference area 27 for detecting the point defect 21 as the imaging area. First, the detection area 26 is set to the imaging area 28. The detection area 26 is set to a matrix size of 3 pixels × 3 pixels. Although the point defect 21 is usually a defect of one pixel, since the image data is dulled by AD conversion or the like in the AD converter 4, the defect of one pixel may affect the surrounding pixels. Therefore, the detection area 26 is preferably set to a matrix size such as 3 × 3 pixels larger than one pixel in consideration of such dullness of the image data. Next, a reference area 27 of 3 pixels × 3 pixels is set so as to be adjacent to the detection area 26. In the example illustrated in FIG. 5, an example adjacent to the left side of the detection region 26 is illustrated.
[0029]
In this way, N detection areas 26 for detecting the point defects 21 are set so as to cover the entire imaging area 28, and N reference areas adjacent to the N detection areas 26 are set. .
[0030]
The detection area and the reference area for detecting the spot defect 22 are set by a matrix size larger than the matrix size of the detection area 26 and the reference area 27 for detecting the point defect 21 described above.
[0031]
FIG. 6 is a diagram for explaining a method of setting the detection area 26 </ b> A and the reference area 27 </ b> A for detecting the line defect 23 in the imaging area 28. For example, in order to detect the line defect 23 along the vertical direction, a detection area 26A and a reference area 27A having an elongated matrix size along the vertical direction are set. The matrix size of the detection area 26A and the reference area 27A is, for example, 2 pixels (horizontal direction) × 50 pixels (vertical direction). In order to detect the line defect 23 along the horizontal direction, a detection area and a reference area having an elongated matrix size along the horizontal direction may be set.
[0032]
FIG. 7 is a diagram illustrating a method for setting the detection area 26 </ b> B and the reference area 27 </ b> B for detecting the mura defect 24 in the imaging area 28. The detection area 26B and the reference area 27B for detecting the mura defect 24 are respectively set by a matrix size larger than the matrix size of the detection area and the reference area for detecting the spot defect 22 described above. The matrix size of the detection area 26B and the reference area 27B is, for example, 30 pixels × 30 pixels. In the mura defect 24, since the pixel display data gradually changes over a plurality of pixels, it is preferable to set the reference region 27B with an appropriate distance from the detection region 26B.
[0033]
As described above, the detection area and the reference area are set as the imaging area according to the type of defect detected by the image defect apparatus 100 (S31).
[0034]
Next, the comparison value calculation unit 19 displays a plurality of pixel display data for displaying a plurality of pixels in each detection region and a plurality of pixels in a reference region corresponding to each detection target region, respectively. The comparison value is calculated for each detection region based on the plurality of pixel display data.
[0035]
For example, with respect to the detection area 26 for detecting the point defect 21 and the reference area 27, a total value of 9 pixel display data for displaying 3 × 3 = 9 pixels in the detection area 26, respectively. Then, the absolute value of the difference from the total value of nine pixel display data for displaying 3 × 3 = 9 pixels in the reference area 27 is calculated as a comparison value (S32).
[0036]
For example, when the value of pixel display data for displaying one pixel is a value between 0 and 255 represented by 8 bits, and the ideal value of the pixel display data is 128 represented by 8 bits If the total value of 3 × 3 = 9 pixel display data in the detection area 26 is 1150 and the total value of 3 × 3 = 9 pixel display data in the corresponding reference area 27 is 1100, The comparison value 50 is calculated as 1150-1100 = 50.
[0037]
Similarly to the point defect 21, the total value of the pixel display data in the detection area and the reference area 27 are also detected for the detection area and the reference area for detecting the spot defect 22, the line defect 23, and the unevenness defect 24. The absolute value of the difference from the total value of the pixel display data is calculated as a comparison value.
[0038]
Next, it is determined whether comparison values have been calculated for all of the set N detection areas (S33). When it is determined that comparison values have not been calculated for all of the set N detection areas 26 (NO in S33), addresses for designating detection areas 26 and reference areas 27 for which comparison values have not yet been calculated are specified. Generate (S34), return to S32, and calculate a comparison value for the detection area 26 and the reference area 27 specified by the address generated in S34.
[0039]
When it is determined that comparison values have been calculated for all N detection regions 26 (YES in S33), comparison value calculation unit 19 compares N comparison values corresponding to the calculated N detection regions. The value is stored in the value storage unit 9 (S35). In this way, the comparison value calculation unit 19 calculates N comparison values based on the N detection areas 26 and the N reference areas 27 corresponding to the N detection areas 26, respectively.
[0040]
Then, it is determined whether or not N comparison values have been calculated for the image data for all the R, G, and B color components (S36). If it is determined that N comparison values have not been calculated for the image data of all color components (NO in S36), the process returns to S32 and the comparison values for the color component image data for which the comparison values have not been calculated yet. Is calculated.
[0041]
When it is determined that N comparison values have been calculated for image data of all color components (YES in S36), all types of defects such as point defect 21, spot defect 22, line defect 23, and mura defect 24 are detected. It is determined whether or not N comparison values have been calculated (S37). When it is determined that N comparison values have not been calculated for all types of defects (NO in S37), the process returns to S31, and detection areas and reference areas are set for the types of defects that have not yet been calculated. .
[0042]
When it is determined that N comparison values have been calculated for all types of defects (YES in S37), rearrangement unit 10 converts each of N comparison values stored in comparison value storage unit 9 into each color component. The data is rearranged in ascending order for each defect type and each defect type (S38).
[0043]
FIG. 8 is a graph showing the comparison value distribution curve 61 plotted so as to connect the values of the N comparison values rearranged by the rearrangement unit 10. In FIG. 8, the horizontal axis indicates the rank of each comparison value rearranged in ascending order, and the vertical axis indicates the value (absolute value) of each comparison value. The example shown in FIG. 8 is a comparison value calculated based on the N comparison values that are calculated based on the detection region 26 and the reference region 27 for detecting the point defect 21 in the R color component. A distribution curve is shown.
[0044]
If the transmission characteristics of the liquid crystal panel 25 are completely uniform over the entire display screen, the pixel display data in the detection area and the pixel display data in the reference area are equal. Should be zero. However, as can be seen from the fact that the transmitted light tends to be weaker in the peripheral part of the display screen than in the central part, there is a difference in transmission characteristics in different regions on the display screen. Therefore, even if the comparison value is calculated with the inspection region 26 and the reference region 27 adjacent to each other as shown in FIG. 6, the value of the comparison value does not become zero but shows a certain value.
[0045]
When N comparison values are rearranged in ascending order, the comparison values are normally distributed as shown by the comparison value distribution curve 61. In the comparative value distribution curve 61, the value of the comparative value is from the first rank to the intermediate rank representing the (N / 2) th rank, which is ½ of the maximum rank, even when approaching the intermediate rank. Little increase. From the intermediate rank to the maximum rank representing the Nth comparison value, the value of the comparison value greatly increases as the distance from the intermediate rank approaches the maximum rank.
[0046]
FIG. 9 is a flowchart showing a detailed operation of the threshold setting unit 12 in the image defect detection method according to the embodiment. First, an intermediate rank that is ½ of the maximum rank is set in the rank variable J that represents the rank of the sorted comparison values (S51).
[0047]
Then, based on (Expression 1) shown below, a difference d between the J-th comparison value and the (J + w) -th comparison value is obtained (S52).
[0048]
d = H (J + w) −H (J) (Formula 1),
here,
H (J): a function indicating the value of the comparison value of the Jth rank,
J: Rank variable indicating the rank of the comparison value,
w: rank range,
d: difference between the J-th order comparison value and the (J + w) -th comparison value,
The rank J and the rank width w are integers satisfying 1 ≦ J ≦ N−w and 1 ≦ w ≦ N, respectively.
[0049]
(Expression 1) corresponds to the calculation for obtaining the slope of the comparison value distribution curve 61. The rank width w is determined in advance according to the total number N of comparison values. For example, when the total number N of comparison values is 1000 and the intermediate rank is 500, the rank width w is set to 5.
[0050]
Then, it is determined whether or not the difference d obtained in S52 is larger than a predetermined threshold setting reference value S (S53). When it is determined that the difference d is larger than a predetermined threshold value setting reference value S (YES in S53), the value H (J + w) of the (J + w) th comparison value is set as the threshold value. (S57).
[0051]
When it is determined that the difference d is not greater than the predetermined threshold setting reference value S (NO in S53), 1 is added to the rank variable J (S54). Then, it is determined whether or not the value of the rank variable J is greater than or equal to the maximum rank (S55). When the value of rank variable J is smaller than the maximum rank (NO in S55), the process returns to S52 to obtain difference d.
[0052]
When it is determined that the value of rank variable J is greater than or equal to the maximum rank (YES in S55), the rearranged N comparison values are separated from the middle rank and are maximum as shown in comparison value distribution curve 61. It is judged that it is not distributed so as to increase greatly as it approaches the ranking. In this case, since the threshold value cannot be automatically set, a user setting screen for the user to set the threshold value is output to the result output unit 15 (S56).
[0053]
For example, the threshold setting reference value S is set to be 10 when the total number N of comparison values is 1000 and the maximum value of (H (J + w) −H (J)) is about 50. To do.
[0054]
FIG. 10 is a graph showing a comparative value distribution curve 61A that is not distributed so that the value of the comparative value increases greatly even if the maximum ranking is moved away from the intermediate ranking. Similar to FIG. 6, the horizontal axis indicates the order of each comparison value rearranged, and the vertical axis indicates the value (absolute value) of each comparison value.
[0055]
In the comparative value distribution curve 61A, even if the intermediate ranking approaches the maximum ranking, the value of the comparative value does not increase as much as the comparative value distribution curve 61 shown in FIG. 6, so the difference d is a predetermined threshold value. It will not be larger than the setting reference value S. For this reason, the threshold value cannot be set automatically.
[0056]
FIG. 11 is an explanatory diagram of a user setting screen for the user to set a threshold value when the threshold value cannot be set automatically. In the example shown in FIG. 11, a graph showing the comparison value distribution curve 61A is displayed on the left half of the screen, and on the right half of the screen is a sample corresponding to N that is the number of comparison values rearranged. The minimum value, maximum value, and intermediate value of the number of N comparison values can be displayed. Below the graph showing the comparison value distribution curve 61A, the values of the comparison values of any rank are displayed. It can be done. In this case, you may display the image imaged by the imaging part 3 simultaneously. Based on the information displayed on the screen as described above, the user can input an arbitrary threshold value.
[0057]
When the comparative value distribution curve 61A shown in FIG. 10 is obtained, the liquid crystal panel 25 to be inspected may be a non-defective product having no image defect, but the liquid crystal panel 25 is a defective product having an image defect. Even in such a case, depending on the type of defect to be inspected, for example, since the setting of the matrix size of the detection region 26 and the reference region 27 is inappropriate, the comparison value (difference) between the detection region 26 and the reference region 27 is Since the image defect is not a value that appropriately represents the image defect, there is a risk that a comparison value distribution curve 61A that is not distributed so that the value of the comparison value increases greatly despite the presence of the image defect in the liquid crystal panel 25 may be obtained. There is. In order to enable correct detection of image defects even in such a case, when the comparison value distribution curve 61A shown in FIG. 10 is obtained, the user setting screen shown in FIG. By operating the value input unit 16, the user can set a threshold value.
[0058]
The threshold value can be set similarly for point defects in the G and B color components.
[0059]
Since the matrix size and the like of the detection area and the reference area differ depending on the type of defect, the calculated comparison value differs depending on the type of defect. For example, the comparison value for detecting a spot defect is about 10 times larger than the comparison value for detecting a point defect. In this case, the comparison value for detecting the spot defect is normalized and processed by 1/10 times, and the comparison calculated based on the detection area and the reference area for detecting the point defect as shown in FIG. 8 or FIG. It is displayed as a graph having a comparative value with the same value as the value graph. Similarly, the comparison value for detecting the line defect and the mura defect is multiplied by a weighted coefficient, normalized, processed, and displayed as a graph as shown in FIG. 8 or FIG.
[0060]
As described above, according to the present embodiment, among the N comparison values rearranged by the rearrangement unit 10, the J-th comparison value and the (J + w) -th comparison value from the smaller comparison value value (J and w are integers satisfying 1 ≦ J ≦ N−w and 1 ≦ w ≦ N, respectively) and a predetermined threshold setting reference value S, Since the threshold value setting unit 12 for setting the threshold value for determining the presence or absence is provided, the threshold value for detecting a defect in the display image can be automatically set.
[0061]
Further, since the threshold value setting unit 12 sets a threshold value for each of a plurality of types of image defects, the presence or absence of defects can be optimally determined according to the type of image defect.
[0062]
Furthermore, since the threshold value setting unit 12 sets a threshold value for each color component, the presence / absence of a defect can be optimally determined according to each color component.
[0063]
Furthermore, since the comparison value is normalized according to the type of image defect, the threshold values can be easily compared and analyzed, and the user can easily set the threshold value.
[0064]
Further, the threshold value setting unit 12 has a difference between the J-th comparison value and the (J + w) -th comparison value from the smaller comparison value, which is larger than a predetermined threshold setting reference value. In this case, since the (J + w) th comparison value is set as the threshold value, the threshold value can be easily set.
[0065]
Further, the threshold value setting unit 12 is inputted by the user when the difference between the Jth comparison value and the (J + w) th comparison value is not larger than a predetermined threshold setting reference value. Since the threshold value is set based on the obtained data, it can be easily verified whether the comparison value is distributed based on the distribution model that is assumed in advance, and based on the distribution model that is assumed in advance. The user can quickly set a threshold value.
[0066]
In addition, although the example which detects the defect of the image displayed on the display screen of a liquid crystal panel was shown, this invention is not limited to this. The present invention can also be applied to inspection of products that require uniformity in output. For example, for inspection of defects in images displayed by elements that require uniformity in output, such as solid-state imaging devices, inspection of uniformity of coating conditions, and inspection of defects in images displayed on CRTs The present invention can be applied.
[0067]
Further, although an example in which the AD converter 4 is connected to the camera 2 is shown, the AD converter 4 may be built in the camera 2.
[0068]
Further, an example in which image data is separated into R, G, and B color components by the RGB separator 6 after AD conversion of the image data by the AD converter 4 is shown. However, the image data is separated into R, G, and B colors. After separating each component, AD conversion may be performed.
[0069]
【The invention's effect】
As described above, according to the present invention, it is possible to provide an image defect detection apparatus and an image defect detection method capable of automatically setting a threshold value for detecting a defect in a display image.
[Brief description of the drawings]
FIG. 1 is a block diagram of an image defect detection apparatus according to an embodiment.
FIG. 2 is a block diagram of an inspection unit constituting the image defect detection device according to the embodiment.
FIG. 3 is an explanatory diagram of image defects detected by the image defect detection device according to the embodiment.
FIG. 4 is a flowchart showing an image defect detection method according to the embodiment.
FIG. 5 is a diagram for explaining a method of setting a detection area and a reference area for detecting point defects as an imaging area according to the embodiment;
FIG. 6 is a diagram for explaining a method of setting a detection area and a reference area for detecting a line defect as an imaging area according to the embodiment;
FIG. 7 is a diagram for explaining a method of setting a detection region and a reference region for detecting a mura defect as an imaging region according to an embodiment;
FIG. 8 is a graph showing an example of the distribution of comparison values rearranged by the rearrangement unit of the image defect detection device according to the embodiment;
FIG. 9 is a flowchart showing a detailed operation of the threshold setting unit in the image defect detection method according to the embodiment.
FIG. 10 is a graph showing another example of the distribution of comparison values rearranged by the rearrangement unit of the image defect detection device according to the embodiment;
FIG. 11 is an explanatory diagram of a user setting screen of the image defect detection device according to the embodiment;
[Explanation of symbols]
3 Imaging unit
10 Sorting part
12 Threshold setting section
17 Area setting section
19 Comparison value calculator
20 Defect detection unit
d Difference
S Reference value for threshold setting

Claims (7)

Imaging means for imaging an image composed of a plurality of pixels;
Area setting means for setting N detection areas for detecting defects in the image (N is an integer of 2 or more) and N reference areas respectively corresponding to the N detection areas. When,
A plurality of pixel display data for a plurality of pixels to display each of the respective detection region, the to a plurality of pixel display data for a plurality of pixels to display each of the reference field corresponding to each detection region based on a comparison value calculation means for calculating each comparative value to each of said detection regions,
Rearrangement means for rearranging the N comparison values calculated by the comparison value calculation means in the order of the values;
The difference between the Jth comparison value and the (J + w) th comparison value from the smaller of the comparison value values among the N comparison values rearranged by the rearrangement means (J and w are 1 ≦ J ≦ N−w, integers satisfying 1 ≦ w ≦ N), and a threshold value for determining the presence / absence of a defect in the image based on a predetermined threshold setting reference value Threshold setting means to set;
A defect detection means for detecting a defect in the image based on the threshold value and the N comparison values;
In the threshold value setting means, the difference between the J-th comparison value and the (J + w) -th comparison value from the smaller value of the comparison value is larger than the predetermined threshold setting reference value. In this case, the image defect detection apparatus is characterized in that the (J + w) th comparison value is set as the threshold value .   The image defect detection apparatus according to claim 1, wherein the rearranging unit rearranges the N comparison values in ascending order.   When the difference between the J-th comparison value and the (J + w) -th comparison value is not greater than the predetermined threshold setting reference value, the rearranging unit is configured to input data input by a user. The image defect detection device according to claim 1, wherein the threshold value is set based on the threshold value. The image defect detection device detects defects of a plurality of types of images,
The area setting means sets the N detection areas and the N reference areas for each defect of the plurality of types of images,
The comparison value calculation means calculates the N comparison values for each defect of the plurality of types of images,
The rearranging means rearranges the N comparison values for each defect of the plurality of types of images,
The threshold setting means sets the threshold for each defect of the plurality of types of images,
The defect detection means detects defects in the image based on a threshold value set for each defect of the plurality of types of images and N comparison values calculated for the defects of the plurality of types of images. The image defect detection apparatus according to claim 1, wherein the defect is detected for each defect of the plurality of types of images. The image defect detection device according to claim 4 , wherein the plurality of types of image defects include a point defect, a spot defect, a line defect, and a mura defect. The image has a hue-based (red, green, blue, etc.) color component,
The comparison value calculation means calculates the N comparison values for each color component,
The rearranging means rearranges the N comparison values for each color component,
The threshold value setting means sets the threshold value for each color component,
The image defect detection device according to claim 1, wherein the defect detection unit detects a defect of the image for each color component. An imaging step of imaging an image composed of a plurality of pixels;
An area setting step for setting N detection areas for detecting defects in the image (N is an integer of 2 or more), and N reference areas respectively corresponding to the N detection areas;
A plurality of pixel display data for displaying said plurality of pixel display data for a plurality of pixels to display each of the respective detection target region, a plurality of pixels of the reference field corresponding to each detection target region, respectively based on the bets, the comparison value calculation step for calculating each comparative value to each of said detection regions,
A rearrangement step of rearranging the N comparison values calculated by the comparison value calculation step in the order of the values of the N comparison values;
The difference between the J-th comparison value and the (J + w) -th comparison value from the smaller of the N comparison values rearranged by the rearrangement step (J and w are 1 ≦ J ≦ N−w, integers satisfying 1 ≦ w ≦ N), and a threshold value for determining the presence / absence of a defect in the image based on a predetermined threshold setting reference value A threshold setting process to be set;
A defect detection step of detecting a defect of the image based on the threshold value set by the threshold value setting step and the N comparison values calculated by the comparison value calculation step ;
In the threshold value setting step, the difference between the J-th comparison value and the (J + w) -th comparison value from the smallest value of the comparison value is larger than the predetermined threshold value setting reference value. In this case, the image defect detection method is characterized in that the (J + w) th comparison value is set as the threshold value .

Images