JP2008232639A - Stain defect detection method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a stain defect detection method and a device capable of detecting highly accurately a stain defect, even when an illuminance difference of the stain defect caused by a voltage change has a noise level. <P>SOLUTION: Imaged images in each different inspection voltage value are acquired (imaged image acquisition process S2), and each difference between every two picked up images is taken successively, and a differential image is generated (differential image generation process S4). Then, each difference between every two differential images is taken successively relative to the differential images, and a variation image is generated (variation image generation process S5). Thereafter, absolute values of each value of each variation image are integrated, and an integrated variation image is generated (integrated variation image generation process S6), and a stain defect is detected based on the integrated variation image (stain defect detection process S7-S11). Therefore, even when the illuminance difference of the stain defect caused by the voltage change has the noise level, the stain defect can be detected highly accurately. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a spot defect detection method and apparatus for automatically and accurately detecting a spot defect in an inspection process of various products such as an inspection process in manufacturing a display device such as a liquid crystal panel and a projector as an application product thereof.

In the inspection of LCD panels such as TFT panels, surface defects called spots and unevenness are difficult to detect automatically with an inspection apparatus because of their indefinite shape and low contrast. For this reason, the inspection is still carried out by the inspector's visual inspection, and automation of the inspection is urgently required to reduce the manufacturing cost.
Note that a spot or uneven defect is a state in which an area on a display screen has a difference in luminance from other areas, and is a state in which there is a bright part or a dark part in a certain range compared to the surrounding area. Usually, a case where the defect area is relatively small is often referred to as a spot defect, and a case where the defect area is relatively large is often referred to as a mura defect. However, since there is no strict definition, in the present invention, surface defects such as a spot defect and a mura defect are collectively referred to as a spot defect.

Conventionally, a method using a difference image (difference image data) has been proposed as a method for automating the detection of a spot defect (see, for example, Patent Document 1).
In the spot defect detection method described in Patent Document 1, the LCD panel is driven with each inspection voltage value having a different voltage value, and the display screen is imaged to obtain a captured image for each inspection voltage value. Next, in the captured image, a difference is created for each of the two captured images sequentially. Subsequently, clusters (white spots or black spots) having different brightness from the surroundings are extracted as defect candidate clusters from the captured image.
In the defect candidate cluster, a value (luminance difference) that is more than the threshold value than the average value of the difference image is detected as a spot defect (specific part).

JP 2003-194669 A

  However, in the detection method described in Patent Document 1, in the case of a defect whose luminance difference due to voltage change is close to the noise level, the value of the defect (luminance difference) is almost equal to the average value and is smaller than the threshold value. The defect candidate is not determined as a defect and the defect cannot be detected. Further, when the threshold value is lowered, there is a problem that normal pixels are also detected as defects. Due to these problems, in the detection method described in Patent Document 1, it is difficult to detect a defect with high accuracy when a luminance difference caused by a voltage change is a defect close to a noise level.

  The object of the present invention has been made in view of the above-described problems, and is capable of detecting a spot defect with high accuracy even if the brightness difference of the spot defect due to a voltage change is a noise level. It is an object to provide a method and apparatus.

  The spot defect detection method of the present invention drives a display device with each inspection voltage value set to a different voltage value, captures a display image of the display device, and captures a captured image for each inspection voltage value. A captured image acquisition step to acquire, a difference image creation step of creating a difference image by sequentially taking a difference for each of the two captured images in each captured image, and each of the two difference images in each difference image A change amount image creating step for creating a change amount image by sequentially taking a difference, an accumulated change amount image creating step for creating an accumulated change amount image by accumulating absolute values of the values of the respective change amount images, and the accumulated change And a spot defect detection step of detecting a spot defect based on the quantity image.

  The spot defect detection apparatus of the present invention drives a display device with each inspection voltage value set to a different voltage value, captures a display image of the display device, and captures a captured image for each inspection voltage value. A captured image acquisition unit to acquire, a difference image creation unit that sequentially creates a difference image for each of the two captured images in each captured image, and a difference image creation unit that generates a difference image for each of the two difference images in each of the difference images. Change amount image creation means for creating a change amount image by sequentially taking the difference, integrated change amount image creation means for adding up absolute values of the values of the respective change amount images to create an integrated change amount image, and the integrated change A spot defect detecting means for detecting a spot defect based on the quantity image is provided.

  Here, the value of each pixel of each created captured image is a luminance value, and thus the value of each pixel of each difference image is a difference in luminance value, that is, a luminance difference. The value of each pixel in the change amount image is a difference in luminance difference, that is, a change amount. The value of each pixel of the integrated change amount image is an integrated value of the absolute value of the change amount. In addition, the detection of a spot defect extracts a defect candidate based on an integrated variation | change_quantity image, Based on the said defect candidate, the suitable captured image in which the said defect candidate appears clearly from a captured image is used as an inspection image. Selection may be made based on this inspection image.

According to the spot defect detection method and apparatus of the present invention, even if the luminance difference of a defective pixel due to a voltage change is a noise level, the absolute value of the luminance difference difference (change amount) is integrated and integrated for each pixel. Since the change amount image is created, the difference between the value of the background pixel (normal part pixel) and the value of the defective pixel (defective part pixel) can be clearly obtained in the integrated change amount image. Therefore, a threshold for extracting defective pixels (defect candidates) can be set appropriately and easily. Therefore, a spot defect can be detected easily and with high accuracy.
In addition, since the integrated change amount image is created by integrating the absolute value of the difference in luminance for each pixel, the integrated change amount image includes the luminance value of each pixel at each inspection voltage value. The change (luminance difference) is reflected. Therefore, a defect that appears only in a specific voltage value can be extracted based on the integrated change image. Therefore, also from this point, the present invention can perform the spot defect with high accuracy.

  In the spot defect detection method of the present invention, in the captured image acquisition step, each captured image is stored in a storage unit, and in the change amount image creation step, each change amount image is stored in the storage unit, The stain defect detection step includes a defect candidate extraction step of extracting a defect candidate based on the integrated change amount image, and a change in which an absolute value of a pixel value included in the defect candidate is maximized from the change amount image. An inspection image selection step of extracting a quantity image and selecting, as an inspection image, a picked-up image in which the defect candidate appears clearly from the picked-up image based on the change amount image; and a spot defect based on the inspection image It is preferable to include a spot defect detection processing step for detecting.

  The spot defect detection device of the present invention includes a storage unit, the captured image acquisition unit stores the captured image in the storage unit, and the change amount image creation unit stores the change amount image. And the spot defect detecting means includes a defect candidate extracting means for extracting a defect candidate based on the integrated change amount image, and an absolute value of a pixel included in the defect candidate from the change amount image. An inspection image selecting means for extracting a variation image having a maximum value and selecting, as an inspection image, a captured image in which the defect candidate appears clearly from the captured image based on the variation image; It is preferable to include a spot defect detection processing unit that detects a spot defect based on the image.

  According to the spot defect detection method and apparatus of the present invention, a defect candidate is extracted from an accumulated change amount image, a captured image in which the defect candidate appears clearly from the captured image is used as an inspection image, and the defect candidate is clearly defined. Since the spot defect is detected based on the inspection image appearing on the spot, the spot defect can be reliably detected with high accuracy.

  In the spot defect detection method of the present invention, in the inspection image selection step, a change amount image in which an absolute value of a pixel value included in the defect candidate is maximized is extracted, and the captured image is based on the change amount image. A captured image in which an absolute value of a difference between an average value of pixels included in the defect candidate and an average value of pixels around the defect candidate is maximized is selected as the inspection image. preferable.

  In the spot defect detection apparatus of the present invention, the inspection image selection unit extracts a change amount image in which the absolute value of a pixel value included in the defect candidate is maximized, and the captured image is based on the change amount image. A captured image in which an absolute value of a difference between an average value of pixels included in the defect candidate and an average value of pixels around the defect candidate is maximized is selected as the inspection image. preferable.

  According to the spot defect detection method and apparatus of the present invention, a picked-up image in which defect candidates clearly appear can be selected as an inspection image, so that a spot defect can be detected more reliably and with high accuracy. it can.

In the spot defect detection method and apparatus of the present invention, it is preferable that the voltage values for inspection are set at equal intervals.
According to the present invention, in the case of inspecting a display device in which the luminance value increases or becomes darker in proportion to the applied voltage value, the luminance difference is obtained by appropriately providing a width between the inspection voltage values. The amount of change or the amount of change may not be calculated properly, and the luminance difference or amount of change may take a large value depending on the width before and after the inspection voltage value. However, in the present invention, since the inspection voltage values are set at equal intervals, there is no such thing, and the brightness difference and the amount of change can be calculated appropriately, so that the spot defect is more reliably and accurately detected. Can be detected.

An embodiment of the present invention will be described with reference to the drawings.
[1. Structure of spot defect detection apparatus]
FIG. 1 is a block diagram showing a configuration of a spot defect detection apparatus according to an embodiment of the present invention.
The spot defect inspection apparatus is an apparatus that performs an appearance inspection of a liquid crystal panel (liquid crystal light valve) 1 as a display device to be inspected and detects a spot defect of the liquid crystal panel 1. In the present embodiment, the liquid crystal panel 1 has a minimum transmittance (forms a black display) when no voltage is applied, and the transmittance increases in proportion to the application of the voltage so that light can pass through. It is configured. Note that the liquid crystal panel 1 may be configured in a normally white mode in which the transmittance is maximum (a white display is formed) when no voltage is applied.

  As shown in FIG. 1, this spot defect inspection apparatus includes a projector 2 as an image projection apparatus, a pattern generator 3 as a pattern generation apparatus, a screen 4, and a charge-coupled device (CCD) camera 5 as an imaging means. And a computer device 6 as image processing means.

A liquid crystal panel 1 can be set to the projector 2 from the outside. The pattern generator 3 can display inspection images of various patterns on the liquid crystal panel 1, and in this embodiment, the liquid crystal panel 1 displays inspection images (solid display) having the same gradation in the surface. Display.
The CCD camera 5 includes a CCD having a resolution equal to or higher than the resolution of the liquid crystal panel 1 and photographs an inspection image projected on the screen 4. The computer device 6 controls the pattern generator 3 and the CCD camera 5 to detect a spot defect in the liquid crystal panel 1. The computer device 6 is connected to an input means 7 including a keyboard and a display device 8.

  The computer device 6 includes an inspection voltage value calculation unit 601, a voltage signal output unit 602, a captured image acquisition unit 603, a target voltage value arrival determination unit 604, a difference image generation unit 605, a change amount image generation unit 606, and an integrated change amount image. A creation unit 607, a binarized image creation unit 608, a labeling unit 609, a defect candidate presence / absence judgment unit 610, an inspection image selection unit 611, a spot defect detection processing unit 612, and a storage unit 613 are provided. Note that the storage means 613 is constituted by, for example, a semiconductor memory or a hard disk device.

The test voltage value calculation unit 601 of the computer device 6 calculates each test voltage value based on the numerical value input from the input unit 7 and stores the test voltage value in the storage unit 613.
The voltage signal output means 602 performs a voltage signal output step S1 for sequentially outputting a voltage signal for driving the liquid crystal panel 1 with each inspection voltage value to the pattern generator 3. The pattern generator 3 drives the liquid crystal panel 1 with each inspection voltage value according to the voltage signal output from the voltage signal output means 602, and causes the liquid crystal panel 1 to display an inspection image for each inspection voltage value.

  The captured image acquisition unit 603 causes the CCD camera 5 to capture an inspection image projected on the screen 4 every time the liquid crystal panel 1 is driven with each inspection voltage value, and to inspect the inspection image for each inspection voltage value. The captured image acquisition step S2 for storing the captured image in the storage unit 613 is performed. Note that the value of each pixel of each captured image is a luminance value.

The difference image creation unit 605 performs a difference image creation step S <b> 4 in which a difference image is sequentially created for each two captured images to create a difference image and the difference image is stored in the storage unit 613. In addition, the value of each pixel of each created difference image is a difference in luminance value, that is, a luminance difference.
The change amount image creating unit 606 creates a change amount image by sequentially taking a difference for every two difference images in the difference image, and performs a change amount image creating step S5 for storing the change amount image in the storage unit 613. . Note that the value of each pixel of the created change amount image is a difference in luminance difference, that is, a change amount.
The integrated change amount image creating means 607 performs an integrated change amount image creating step S6 in which the absolute values of the values of the respective change amount images are integrated to create one integrated change amount image.

The binarized image creation unit 608 performs a binarized image creation step S7 for binarizing the integrated change amount image and creating a binarized image.
The labeling unit 609 performs a labeling step S8 that assigns an identification number to each cluster (pixel group) in the binarized image.

The defect candidate presence / absence determination means 610 performs a defect candidate presence / absence determination step S9 for determining whether or not a defect candidate exists in the cluster. The defect candidate extraction unit of the present invention includes the binarized image creation unit 608, the labeling unit 609, and the defect candidate presence / absence determination unit 610.
The inspection image selection unit 611 performs an inspection image selection step S10 of selecting, as an inspection image, a captured image in which a cluster determined to be a defect candidate is clearly displayed in the defect candidate presence / absence determination unit 610.

  The spot defect detection processing means 612 performs a spot defect detection processing step S11 for detecting a spot defect using the selected inspection image. In addition, since the detection of the spot defect in the spot defect detection processing unit 612 can be performed by a known method disclosed in, for example, Japanese Patent Application Laid-Open No. 2006-258713, the description thereof is omitted here. The binarized image creating unit 608, the labeling unit 609, the defect candidate presence / absence determining unit 610, the inspection image selecting unit 611, and the spot defect detecting unit 612 constitute the spot defect detecting unit of the present invention.

[2. Operation of the spot defect detection device)
Next, the operation of the spot defect detection device in this embodiment will be described.
First, the inspector sets the liquid crystal panel 1 to be inspected in the projector 2 and operates the input means 7 to input the minimum voltage value, the maximum voltage value, and the number of steps. In this embodiment, the minimum voltage value is input as 0.0 V, the maximum voltage value is 5.0 V, and the number of steps is input as 20.

  When these values are input, the inspection voltage value calculation means 601 of the computer device 6 calculates the step value by dividing the difference between the input minimum voltage value and the maximum voltage value by the input number of steps. To do. Then, the test voltage value calculation means 601 stores each value (including the minimum voltage) obtained by sequentially adding the step value to the minimum voltage value until the minimum voltage value reaches the maximum voltage value as the test voltage value. The information is stored in the means 613.

  In the present embodiment, the minimum voltage value is 0.0V, the maximum voltage value is 5.0V, the number of steps is 20, and the step value is 0.25V. Therefore, in the present embodiment, each value (0.0V, 0.25V, 0.50V ... 4.75V, 5) obtained by adding 0.25V to 0.0V until 0.0V becomes 5.0V. .0V) is stored in the storage means 613 as an inspection voltage value.

  Then, the inspector drives the liquid crystal panel 1 with the minimum voltage value or the maximum voltage value, and the CCD camera 5 performs the inspection image with a low gradation or the inspection with a high gradation with the set values such as the set aperture and shutter speed. Test whether the image can be captured accurately. If the CCD camera 5 cannot accurately capture the inspection image, the set value is adjusted to an appropriate value. This is because when the setting values such as the aperture and the shutter speed are not properly adjusted, the liquid crystal panel 1 is driven at the minimum voltage value or the maximum voltage value, and the CCD is displayed when an inspection image with low or high gradation is displayed. This is because the camera 5 cannot accurately capture the inspection image, and the captured image may become black or white.

  Then, after setting the inspection voltage value by the inspector and adjusting the setting value of the CCD camera 5, the computer device 6 controls the pattern generator 3 so that each inspection voltage value (0.0V, 0.25V, (4.75 V, 5.0 V), the liquid crystal panel 1 is sequentially driven, and the inspection images sequentially displayed on the liquid crystal panel 1 are projected onto the screen 4 by the projector 2. Thereby, in this embodiment, the gradation of the inspection image displayed on the liquid crystal panel 1 is sequentially increased from all black to all white.

Next, the computer device 6 causes the CCD camera 5 to pick up inspection images sequentially projected on the screen 4 and obtains a picked-up image of the inspection image for each inspection voltage value. Subsequently, the computer device 6 selects a captured image suitable for detection of a spot defect as an inspection image from among the captured images acquired for each voltage value for inspection.
Then, the computer device 6 detects a spot defect based on the selected inspection image, and displays the spot defect on the display device 8.

[3. Spot detection process by computer device]
FIG. 2 is a flowchart for explaining the operation of the spot defect detection apparatus of the present embodiment, FIG. 3 is a diagram for explaining a process of creating an integrated change amount image from a captured image, and FIG. 4 is for each inspection. It is a figure which shows the value of the background pixel (pixel of a normal part) and the defective pixel (pixel of a defective part) of each image in a voltage value. Note that the operation shown in FIG. 2 is realized by a program executed on the computer device 6.
FIG. 2 is a flowchart of a spot defect detection process performed by the computer device 6 after each inspection voltage value (0.0 V to 5.0 V) is set and the setting value of the CCD camera 5 is adjusted. Based on

  First, the voltage signal output unit 602 reads each inspection voltage value (0.0 V to 5.0 V) from the storage unit 613 and drives the liquid crystal panel 1 with the minimum voltage value for the inspection voltage (0.0 V). Voltage signal is output to the pattern generator 3 (voltage signal output step S1).

  Next, the captured image acquisition unit 603 causes the CCD camera 5 to capture the inspection image projected on the screen 4 and stores the captured image G (1) (see FIG. 3) in the storage unit 613 (captured image acquisition). Step S2). Note that a captured image (imaging data) captured by the CCD camera 5 is converted into digital data of 4096 gradations (12 bits) by an A / D converter (not shown) and stored in the storage unit 613.

  Subsequently, as shown in FIG. 2, the target voltage value arrival determination means 604 determines that the voltage signal currently being output is not a voltage signal for driving the liquid crystal panel 1 with the maximum voltage value that is the target voltage value. The process is returned to the voltage signal output process S1 (NO in the target voltage value attainment determination process S3). Then, the voltage signal output means 602 outputs a voltage signal for driving the liquid crystal panel 1 to the pattern generator 3 with a voltage value for inspection (0.25V) which is higher by one step value (0.25V). (Voltage signal output step S1).

In the same manner as described above, the captured image acquisition unit 603 controls the CCD camera 5 to store the captured image G (2) at the inspection voltage value in the storage unit 613 (captured image acquisition step S2).
Thus, until the target voltage value arrival determination means 604 determines that the currently output voltage signal is a voltage signal for driving the liquid crystal panel 1 with the maximum voltage value that is the target value, the voltage signal The output process S1 and the captured image acquisition process S2 are repeated, and the captured images G (n) (G1, G2,..., G21) at the respective inspection voltage values (0.0V to 5.0V) (see FIG. 3; Also, in order to make G (1), T (1), U (1), etc. easy to understand, they may be expressed as G1, T1, U1, etc.).

  When the target voltage value arrival determination means 604 determines that the voltage signal currently output to the pattern generator 3 is a voltage signal for driving the liquid crystal panel 1 with the maximum voltage value which is the target value, the process is the next difference. The process proceeds to image creation step S4 (YES in target voltage value attainment determination step S3). Note that the value of each pixel of each captured image G (n) is a luminance value (see background luminance and defect luminance in FIG. 4).

  In the difference image creation step S4, as shown in FIG. 3, the difference image creation means 605 has two captured images ((G1, G2), (G2, G3),... (G20) whose test voltage values are different by one step value. , G21)) to obtain a difference image T (j) (T1, T2,..., T20) by sequentially taking the difference and store the difference image T (j) in the storage means 613 (difference image creation step S4). . In addition, the value of each pixel of each created difference image T (j) is a luminance value difference, that is, a luminance difference (see the difference between the background and the background before and after the difference in FIG. 4).

  Subsequently, as shown in FIG. 3, the change amount image creating means 606 has two captured images ((G1, G2), (G2, G3),... (G19, G20) whose inspection voltage values are different by one step value. ) Created from the difference image (T1, T2,..., T19), the captured image (G2, G3..., G20) having a higher voltage value of the two captured images, and the captured image having a higher voltage value. The difference image (T2, T3,..., T20) created from the captured image (G3, G4,..., G21) whose voltage value is one step higher is taken as a difference image U (k) (U1, U2. , U19) and the change amount image U (k) is stored in the storage means 613 (change amount image generation step S5). That is, the change amount image creating means 606 has three captured images ((G1, G2, G3), (G2, G3, G4),..., (G19, G20, G21) each having a different step voltage value. One change amount image U (k) is created for each pixel of the created change amount image U (k), which is the difference in luminance difference, that is, the change amount (the background change amount in FIG. 4). And defect variation).

  Next, as shown in FIG. 3, the integrated change amount image creating means 607 integrates the absolute value of the pixel value in each change amount image U (k) for each pixel (background absolute value integration in FIG. 4). Further, one integrated variation image H is created (integrated variation image creating step S6).

  FIG. 5 is a diagram showing changes in values of background pixels (normal pixels) and defective pixels in each captured image G (n) (vertical axis: luminance value, horizontal axis: identification value n of captured image G), FIG. 6 is a diagram showing changes in values of background pixels and defective pixels in each difference image T (j) (vertical axis: luminance difference, horizontal axis: identification value j of differential image).

  As shown in FIGS. 5 and 6, when the luminance difference of the defect caused by the voltage change is close to the noise level, the value difference between the background pixel and the defective pixel becomes very small. For this reason, it turns out that it is difficult to extract a defect candidate using a threshold value from difference image T (j).

  FIG. 7 is a diagram showing changes in values of background pixels and defective pixels in each variation image U (k) (vertical axis: variation, horizontal axis: identification value k of variation image U), and FIG. The figure which shows the change of the value of a background pixel and a defective pixel by the degree of integration | accumulation of the variation | change_quantity image U (k) in the variation | change_quantity image H (a vertical axis | shaft: the integrated value of the absolute value of variation | change_quantity, a horizontal axis: Identification value k).

  However, in this embodiment, as shown in FIG. 8, for each pixel, the absolute value of the change amount of the pixel is integrated to create one integrated change amount image H. Even if the difference from the luminance difference of the pixel is small, the difference between the background pixel and the defective pixel can be clearly obtained in the integrated change amount image H. For this reason, the threshold value can be easily and appropriately set in the binarized image creating step S7 of the next step. In addition, since the change (brightness difference) of the luminance value of each pixel at each inspection voltage value is reflected in the integrated change amount image H, the specific change is clearly based on the integrated change amount image H. It is possible to extract defects that appear as follows.

After the accumulated change amount image creating step S6, the binarized image creating unit 608 sets 1 (white) as a pixel having a value equal to or greater than a predetermined threshold (for example, 33) in the accumulated change amount image H (see FIG. 9). Then, the pixel below the threshold is set to 0 (black), and a binarized image (see FIG. 10) is created (binarized image creation step S7).
After the binarized image creation step S7, as shown in FIG. 11, the labeling means 609 assigns identification numbers 0 to 3 to the clusters R0 to R3 in the binarized image (labeling step S8).

  Then, the defect candidate presence / absence determining means 610 determines and extracts clusters R2 and R3 having the number of pixels included in the cluster equal to or greater than a predetermined threshold (for example, 5) as defect candidates among these clusters R0 to R3. (YES in defect candidate presence / absence determination step S9 in FIG. 2). Here, if the defect candidate presence / absence determining means 610 determines that there is no defect candidate in the clusters R0 to R3, it is assumed that the liquid crystal panel 1 to be inspected has no defect, and a spot defect is detected. The process ends (NO in defect candidate presence / absence determination step S9 in FIG. 2). The defect candidate extraction means of the present invention includes these binarized image creation step S7, labeling step S8, and defect candidate presence / absence determination step S9.

FIG. 12 is a diagram illustrating a process of extracting the change amount image U (10) in which the absolute value of the defect candidate value is maximum from the change amount image U (k).
Next, the inspection image selection unit 611 has the maximum absolute value (change amount) of the pixels included in the defect candidate clusters R2 and R3 from the change amount image U (k) (for example, the defect change amount in FIG. 4). The change amount image U (10) that becomes 8) is extracted.
At this time, as shown in FIG. 11, the pixels surrounding the centroid coordinates of the defect candidates indicated by the crosses (the positions of the centroid coordinates come within the pixels without checking all the pixels included in the defect candidate cluster. In such a case, the inspection image selection means 611 may be configured to investigate only the pixel).

  Then, the inspection image selection unit 611 uses the defect candidate cluster R2 among the three captured images G (10), G (11), and G (12) used to create the variation image U (10). , R3, a captured image in which the absolute value of the difference between the average value of the pixel values (luminance values) included in the defect candidates and the average value of the pixels around the defect candidate clusters R2 and R3 is maximized. The inspection image that appears most clearly is selected (inspection image selection step S10).

  Then, the spot defect detection processing unit 612 detects the spot defect using the inspection image by a known method disclosed in Japanese Patent Laid-Open No. 2006-258713 (Stain defect detection processing step S11). The detected spot defect is displayed on the display device 8 so that the inspector can easily grasp it. It should be noted that the spot defect detection step of the present invention includes the binarization processing image creation step S7, the labeling step S8, the defect candidate presence / absence determination step S9, the inspection image selection step S10, and the spot defect detection processing step S11. Yes.

[4. Effects according to this embodiment
According to this embodiment, there are the following effects.
(1) In this embodiment, for each pixel, the absolute value of the value (change amount) of the pixel in each change amount image U (k) is integrated to create one integrated change amount image H. Even if the luminance difference of the defective pixel due to the change is a noise level, the difference between the background pixel and the defective pixel can be clearly obtained in the integrated change amount image H. For this reason, a threshold value can be set easily and appropriately, and defect candidates can be easily and appropriately extracted from the integrated change amount image H. In addition, since the change (brightness difference) of the luminance value of each pixel at each inspection voltage value is reflected in the integrated change amount image H, based on the integrated change amount image H, it is strongly applied only to a specific voltage value. Appearing defects can be extracted. Therefore, also from this point, the present invention can detect a spot defect with high accuracy.

(2) The degree of appearance of the spot defect varies depending on each inspection voltage value, that is, each captured image G (n). For this reason, when detecting a spot defect based on the captured image G (n), it is preferable to detect the spot defect based on a suitable captured image in which the spot defect appears clearly.
In this regard, in the present embodiment, the change amount image U (10) in which the absolute value of the pixel values (change amounts) included in the defect candidate clusters R2 and R3 is maximized from the change amount image U (k). To extract. Then, among the three captured images G (10), G (11), and G (12) used to create the variation image U (10), the defect candidate clusters R2 and R3 are the most clearly defined. Since the picked-up image that appears and the spot defect is detected based on the picked-up image, the spot defect can be detected with higher accuracy.

(3) When inspecting the liquid crystal panel 1 in which the luminance value increases in proportion to the applied voltage value, if a width between the inspection voltage values is appropriately provided, the luminance difference and the amount of change cannot be calculated appropriately. Depending on the width before and after the inspection voltage value, the luminance difference or the amount of change may take a large value. In this respect, in the present embodiment, the inspection voltage values are set at equal intervals, so that there is no such thing, and the brightness difference and the amount of change can be surely and appropriately calculated, and the spot defect is surely highly accurate. Can be detected.

[5. Modification of Embodiment]
As mentioned above, although this embodiment was shown concretely, various improvement and deformation | transformation are possible in the range which is not restricted to the said embodiment and does not deviate from the summary of this invention.
For example, in the embodiment, of the three captured images, the difference between the average value of the pixel values (luminance values) included in the defect candidate cluster and the average value of the pixels around the defect candidate cluster Although the captured image having the maximum absolute value is selected as the inspection image, the method is not limited to this as long as a captured image in which a cluster of defect candidates appears clearly can be selected.

In the embodiment, each inspection voltage value is set based on each numerical value (minimum voltage value, maximum voltage value, and number of steps) input by the inspector, but a predetermined inspection voltage value is set in advance. May be.
In the said embodiment, although the captured image was acquired from the order with the low voltage value, the captured image may be acquired from the order with the high voltage value. Further, if captured images with all the inspection voltage values are acquired, the order of acquisition may be random and may be appropriate.

In the embodiment, the difference image is created by subtracting a low voltage value captured image from a high voltage value captured image. However, the difference image pulls a high voltage value captured image from a low voltage value captured image. It may be created.
In the embodiment, the liquid crystal panel 1 is configured such that the luminance value increases in proportion to the applied voltage value. For this reason, the voltage values for inspection applied to the liquid crystal panel are set at equal intervals. However, in the case of a display device in which the applied voltage value and the luminance value are not proportional, the inspection voltage value is set to an equal interval if the interval between the luminance values before and after the normal pixel is constant. It does not have to be.

  Contrary to the above-described embodiment, the change amount image is a difference image created from two captured images having a voltage value that differs by one step value. It may be created by subtracting from a difference image created from a captured image whose voltage value is one step value higher than that of a higher captured image.

  The detection target of the spot defect is not limited to the liquid crystal light valve using the TFT element as in the above embodiment. That is, the present invention is a display panel component such as a liquid crystal panel using other diode elements, a plasma display, an organic EL display, a DMD (direct mirror device) panel, and a display device / product inspection using them. Needless to say, even when used for these, they are not excluded from the scope of the present invention.

Furthermore, the present invention is not limited to the inspection of various display devices. For example, when there are spot-like scratches on a printed matter, a case of a household electrical appliance, a car body, etc., an image with these spots is obtained by imaging these spots. If it is obtained, the stain can be detected, so that it can also be applied to stain inspection of the surface and layer of various products and printed matter.
In short, the present invention can be used to detect a brightness spot defect or a color spot defect since it can be detected if it is a spot defect having a brightness difference from the surroundings.

The figure which shows the structure of the spot defect detection apparatus which concerns on one Embodiment of this invention. The flowchart for demonstrating operation | movement of the spot defect detection apparatus in the said embodiment. The figure for demonstrating the process which produces an integrated variation | change_quantity image from the captured image in the said embodiment. The figure which shows the value of the background pixel and defective pixel of each image in each voltage value for a test | inspection in the said embodiment. The figure which shows the change of the value of the background pixel and defective pixel in each captured image in the said embodiment. The figure which shows the change of the value of the background pixel and defective pixel in each difference image. The figure which shows the change of the value of the background pixel and defect pixel in each variation | change_quantity image in the said embodiment. The figure which shows the change of the value of a background pixel and a defective pixel by the degree of integration | accumulation of the variation | change_quantity image in the integrated variation | change_quantity image in the said embodiment. The figure which shows the integrated variation | change_quantity image in the said embodiment. The figure which shows the binarization process image in the said embodiment. The figure which shows the labeled cluster defect in the said embodiment. The figure which shows the process of extracting the variation | change_quantity image from which the absolute value of the value of a defect candidate becomes the maximum from the variation | change_quantity image in the said embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Liquid crystal panel (display device), 5 ... CCD camera (imaging means), 603 ... Captured image acquisition means, 605 ... Difference image creation means, 606 ... Change amount image creation means, 607 ... Integrated change amount image creation means, 611 ... inspection image selection means, 612 ... stain defect detection processing means, 613 ... storage means, S2 ... captured image acquisition step, S4 ... difference image creation step, S5 ... change amount image creation step, S6 ... integrated change amount image creation step, S10: Inspection image selection step, S11: Blemish defect detection processing step.

Claims (8)

Driving the display device with each inspection voltage value set to a different voltage value, capturing a display image of the display device, and acquiring a captured image for each inspection voltage value; and
In each captured image, a difference image creation step of creating a difference image by sequentially taking a difference for each of the two captured images;
In each difference image, a change amount image creating step of taking a difference sequentially for each of the two difference images to create a change amount image;
An accumulated change amount image creating step of creating an accumulated change amount image by accumulating absolute values of the values of the respective change amount images;
A spot defect detection method comprising: a spot defect detection step of detecting a spot defect based on the integrated change amount image. In the spot defect detection method according to claim 1,
In the captured image acquisition step, each captured image is stored in a storage unit;
In the change amount image creating step, each change amount image is stored in the storage means,
The spot defect detection step includes
A defect candidate extraction step for extracting defect candidates based on the integrated change amount image;
A change amount image in which the absolute value of the pixel value included in the defect candidate is maximized is extracted from the change amount image, and the defect candidate is clearly identified from the captured image based on the change amount image. Inspection image selection process for selecting the captured image that appears as an inspection image;
A spot defect detection method comprising: a spot defect detection processing step of detecting a spot defect based on the inspection image. In the spot defect detection method according to claim 2,
In the inspection image selection step, a change amount image in which the absolute value of the pixel value included in the defect candidate is maximized is extracted, and is included in the defect candidate from the captured image based on the change amount image. A spot defect detection method, wherein a picked-up image having a maximum absolute value of a difference between an average value of pixel values and an average value of pixels around the defect candidate is selected as the inspection image. In the spot defect detection method according to any one of claims 1 to 3,
A spot defect detection method characterized in that each inspection voltage value is set at equal intervals. Driving the display device with each inspection voltage value set to a different voltage value, capturing a display image of the display device, and acquiring a captured image for each inspection voltage value; and
In each captured image, difference image creation means for creating a difference image by sequentially taking a difference for each of the two captured images;
In each of the difference images, a change amount image creating unit that sequentially creates a change amount image for each of the two difference images;
Integrated change amount image creating means for integrating the absolute values of the values of the respective change amount images to create an integrated change amount image;
A spot defect detection device comprising spot defect detection means for detecting a spot defect based on the integrated change amount image. In the spot defect detection device according to claim 5,
A storage means,
The captured image acquisition means stores the captured images in the storage means,
In the change amount image creating means, each change amount image is stored in the storage means,
The spot defect detecting means is
Defect candidate extraction means for extracting defect candidates based on the integrated variation image;
A change amount image in which the absolute value of the pixel value included in the defect candidate is maximized is extracted from the change amount image, and the defect candidate is clearly identified from the captured image based on the change amount image. Inspection image selection means for selecting a captured image appearing as an inspection image;
A spot defect detection apparatus comprising spot defect detection processing means for detecting a spot defect based on the inspection image. In the spot defect detection device according to claim 6,
The inspection image selection unit extracts a change amount image in which the absolute value of the pixel value included in the defect candidate is maximized, and is included in the defect candidate from the captured image based on the change amount image. A spot defect detection apparatus, wherein a picked-up image having a maximum absolute value of a difference between an average value of pixel values and an average value of pixels around the defect candidate is selected as the inspection image. In the spot defect detection device according to any one of claims 5 to 7,
A spot defect detection apparatus characterized in that each inspection voltage value is set at equal intervals.