JP2008003063A - Shading correction method, defect detection method, and defect detector and control method program thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a shading correction method for improving precision for removing shading in a display defect detection step of a display device and detecting display defects precisely, and to provide a defect detection method, a defect detector, and a control method program of the defect detector. <P>SOLUTION: The defect detector determines area information on whether a partial image in a captured image is an image at an intermediate section or that at a peripheral section (ST21), determines whether a luminance value for each pixel in the partial image is an application pixel in a prescribed luminance range (ST22), and applies a smoothing filter selected based on the area information to the application pixel of the partial image (ST23) from a smoothing filter for images at the intermediate section corresponding to the images at the intermediate section and a smoothing filter for images at the peripheral section corresponding to the images at the peripheral section, thus calculating the background luminance value of a target pixel in the partial image for generating a background image. By calculating the difference between the background image and the captured image, shading can be removed appropriately from the captured image. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a shading correction method and a defect detection method for removing shading in order to accurately detect display defects of an image display device in an inspection process in manufacturing an image display device such as a liquid crystal panel or a projector as an application product thereof. The present invention relates to a defect detection apparatus and a control method program thereof.

Conventionally, in a display appearance inspection of a liquid crystal panel (image display device) such as a TFT (Thin Film Transistor), display defects (for example, spots, unevenness, streaks, etc.) appearing in an image displayed on the liquid crystal panel are detected by image processing. In this case, an image display device or its display image is captured by an imaging device such as a CCD camera to acquire image data, and display defects are detected based on the image data.
The image data acquired by the CCD camera includes shading caused by uneven illumination of a light beam irradiated on the liquid crystal panel, a decrease in the amount of peripheral light of the image due to a lens, and the like. For this reason, in order to improve the display defect detection accuracy, it is necessary to remove components (shading) other than the defect components in the image data (shading correction).
The following techniques have been proposed as methods for performing shading correction using image processing (see, for example, Patent Documents 1 and 2).

In the technique described in Patent Document 1, background image data created in advance from image data (inspection image data) acquired by capturing an image displayed on a liquid crystal panel to be inspected with a CCD camera. The shading correction which takes the difference is adopted. Here, the background image data is obtained by taking a plurality of image data by capturing an image displayed on a liquid crystal panel that is not an inspection target with as few defects as possible, and averaging the plurality of image data.
On the other hand, the technique described in Patent Document 2 employs shading correction that takes a difference between inspection image data and a shading correction pattern (background image data) generated from the inspection image data. Here, a method is used in which a shading gain is obtained, a shading correction pattern is generated by an appropriate calculation, and shading correction is performed based on the shading correction pattern.

JP 2003-168103 A JP 2004-272077 A

In the technique described in Patent Document 1, the background image data is acquired by capturing an image displayed on a liquid crystal panel different from the inspection target. When the inspection image data is acquired and the installation position of the liquid crystal panel Is slightly displaced, the position of the corresponding pixel is slightly displaced. For this reason, when the difference between the inspection image data and the background image data is taken, a difference cannot be obtained between the corresponding pixels, and shading cannot be appropriately removed from the inspection image data. That is, there is a problem that display defects cannot be detected with high accuracy.
On the other hand, in the technique described in Patent Document 2, a shading gain between the image central portion and the image peripheral portion is obtained, and the luminance difference between the image central portion and the image peripheral portion is obtained by multiplying the gain by the luminance value of the image peripheral portion. However, if there is a display defect at the periphery of the image, the background image data including this defect component will be generated. The defect component is removed from the inspection image data that is generated and obtained the difference from the background image data. For this reason, even in the technique described in Patent Document 2, display defects cannot be detected with high accuracy, and in particular, there is a problem that the accuracy of shading correction in the peripheral portion of the image is lowered.

  An object of the present invention is to provide a shading correction method, a defect detection method, a defect detection apparatus, and a control method program thereof that can improve the accuracy of shading removal in a display defect detection process of a display device and can detect display defects with high accuracy. is there.

  The shading correction method of the present invention removes shading from the captured image by taking the difference between the captured image acquired by capturing the display image of the display device to be inspected and the background image generated from the captured image. In the shading correction method, the partial image having a predetermined size including an arbitrary pixel of interest in the captured image is an intermediate image composed of only the image of the inspection target portion of the display device, or the inspection target An area information determination step for determining area information as to whether or not the image is a peripheral image including an image of a non-inspection portion outside the outside, and a luminance value for each pixel in the partial image is within a predetermined luminance range An applied pixel determination step for determining whether or not there is an applied pixel that has been determined to be within a predetermined luminance range; an intermediate image smoothing filter corresponding to the intermediate image; and A smoothing filter is selected based on the area information from at least two types of smoothing filters prepared in advance including a peripheral image smoothing filter corresponding to a side image, and the selected smoothing filter is applied to the partial image. A background luminance value calculation step of calculating a background luminance value of the pixel of interest by applying to the pixel, and each of the plurality of pixels of the inspection target portion as the pixel of interest, the area information determination step, the application pixel determination step, and the A background luminance value calculating step is executed to generate the background image.

In the present invention, a background image is generated from the captured image, and the pixel position of the display device is shifted between the captured image and the background image by removing the shading by taking the difference between the captured image and the background image. The difference between the pixels corresponding to each other can be obtained, and the shading curve can be matched by eliminating the influence of variations in illumination conditions, etc., and shading can be appropriately removed from the captured image. it can. Therefore, it is possible to accurately detect a display defect having a minute contrast.
In addition, it is determined whether the partial image is an intermediate image or a peripheral image by the area information determination process, and the peripheral image is applied to the peripheral image by applying a smooth filter for the peripheral image. It is possible to execute shading correction that appropriately evaluates the influence of the inspection non-target portion outside.
In addition, the smoothing filter is applied only to the applicable pixels by excluding pixels that are outside the predetermined luminance range, that is, pixels that may have display defects, and non-inspection areas that are non-transmission parts or non-light-emitting parts. By doing so, it is possible to prevent the defect component from being removed from the difference image obtained by taking the difference between the background image and the captured image without including the defect component in the background image. Here, when applying the smoothing filter only to the application pixel in the partial image, the total value is obtained by multiplying each application pixel and the coefficient of the smoothing filter, and then the coefficient value of the smoothing filter corresponding to the application pixel is calculated. By dividing the total value by the sum, the background luminance value of the pixel of interest can be calculated appropriately.
Therefore, it is possible to improve the shading removal accuracy in the peripheral portion of the inspection target portion or a portion where there is a possibility of display defect, and it is possible to improve the display defect detection accuracy.

  In the present invention, in the area information determination step, the partial image is a peripheral image at any position of the upper side, the lower side, the left vertical side, and the right vertical side of the inspection target part, or the partial image is Determining the area information of the peripheral image at any one of the four corners of the upper left corner, the upper right corner, the lower left corner, and the lower right corner of the inspection target portion, and the peripheral image smoothing filter includes the upper side A first peripheral image smoothing filter corresponding to the peripheral image of the upper and lower sides, a second peripheral image smoothing filter corresponding to the peripheral image of the left vertical side and the right vertical side, and the four corners At least three types including a third peripheral image smoothing filter corresponding to the peripheral image of the image, and in the background luminance value calculation step, the application pixel of the peripheral image on the upper side and the lower side is prepared. The first peripheral image smoothing Applying a filter, applying a second peripheral image smoothing filter to the application pixel of the peripheral image of the left vertical side and right vertical side, and applying a third peripheral to the application pixel of the peripheral image of the four corners It is preferable to apply a partial image smoothing filter to calculate the background luminance value of the pixel of interest in each peripheral image.

  According to such a configuration, the partial image is any of the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, the upper right corner, the lower left corner, and the lower right corner of the inspection target part. And the first to third peripheral image smoothing filters are selected and applied according to the determined position of the partial image, so that the shading removal accuracy in the peripheral portion of the inspection target portion is applied. Can be further enhanced.

In the present invention, the first peripheral image smoothing filter emphasizes the pixel of interest and the pixels arranged on the left and right of the pixel of interest, and weakens the pixels farther upward and downward from the pixels arranged on the left and right. The second peripheral image smoothing filter emphasizes the pixel of interest and the pixels arranged above and below the pixel of interest, and the left and right sides of the pixels arranged above and below are configured. The third peripheral image smoothing filter is configured to emphasize the target pixel, and the pixel farther from the target pixel in the radial direction is configured to have a weak tone process. It is preferable that the filter array is configured.
Here, the enhancement process means a process of increasing the weighting of the luminance value of the target pixel, that is, a process of applying a large coefficient, and the weak adjustment process is a filter value (coefficient) smaller than the filter value to be enhanced. By applying, it means a process of weakening the weighting of the luminance value of the target pixel.

According to such a configuration, the first peripheral image smoothing filter applied to the peripheral image of the upper side and the lower side in the inspection target portion is the pixel of interest and the pixels aligned on the left and right of the pixel of interest, that is, the upper side In addition, the applied pixels along the boundary line between the inspection target part and the inspection non-target part in the lower part and the lower part are processed to increase the weighting of the luminance value, and the applied pixel farther from the boundary line is weighted with the luminance value. Will be processed to weaken. Therefore, the background luminance value at the boundary portion between the inspection target portion and the inspection non-target portion having a large luminance change can be calculated with high accuracy, and the accuracy of shading removal using the generated background image can be further improved. it can.
Further, in the second peripheral image smoothing filter, in the same manner as in the first peripheral image smoothing filter, the pixel of interest and the pixels aligned above and below the pixel of interest, that is, the left vertical side portion and the right vertical side portion are inspected. The background luminance value at the boundary part (left and right vertical sides of the inspection target part) is accurately processed by increasing the weighting of the luminance value for the applied pixels along the boundary line between the inspection part and the non-inspection part. It can be calculated well.
Further, in the third peripheral image smoothing filter, in substantially the same manner as the first and second peripheral image smoothing filters, in the target pixel, that is, the upper left corner, the upper right corner, the lower left corner, and the lower right corner. The applied pixel at the position where the boundary line between the inspection target part and the non-inspection part intersects is processed to increase the weighting of the luminance value, and the luminance value weighting is applied to the applied pixel farther from the boundary line intersection position. By performing the process of weakening the brightness, the background luminance value at the intersection position (four corners of the inspection target portion) can be calculated with high accuracy.

  In the present invention, in the area information determination step, the partial image is a peripheral image at any position of the upper side, the lower side, the left vertical side, and the right vertical side of the inspection target part, or the part. Determining the area information of the peripheral image at any one of the four corners of the upper left corner, the upper right corner, the lower left corner and the lower right corner of the inspection target part; and the peripheral image smoothing filter, An upper side image smoothing filter corresponding to the peripheral image of the upper side part, a lower side image smoothing filter corresponding to the peripheral part image of the lower side part, and a left vertical side corresponding to the peripheral part image of the left vertical side part A smooth filter for a side image, a smooth filter for a right vertical side image corresponding to a peripheral image of the right vertical side, a smooth filter for an upper left corner image corresponding to a peripheral image of the upper left corner, Upper right corner image corresponding to the peripheral image of the upper right corner A smoothing filter, a lower left corner image smoothing filter corresponding to the lower left corner peripheral image, and a lower right corner image smoothing filter corresponding to the lower right corner peripheral image. Eight types are prepared in advance, and in the background luminance value calculating step, the upper side image smoothing filter is applied to the application pixel of the peripheral image of the upper side part, and the application pixel of the peripheral image of the lower side part is applied. Applying the smoothing filter for the lower side image, applying the smoothing filter for the left vertical side image to the application pixel of the peripheral image of the left vertical side, and applying the peripheral image of the right vertical side Applying the right vertical side image smoothing filter to a pixel, applying the upper left corner image smoothing filter to the application pixel of the upper left corner peripheral image, and correcting the upper right corner peripheral image The upper right corner image plane is applied to the application pixel. Applying a filter, applying the lower left corner image smoothing filter to the application pixels of the lower left corner peripheral image, and applying the lower right corner image to the application pixels of the lower right corner peripheral image The background luminance value of the pixel of interest in each peripheral image may be calculated by applying a smoothing filter for the image.

  At this time, in the present invention, the smoothing filter for the upper side image emphasizes the pixel of interest and the pixels arranged on the left and right of the pixel of interest, and the weakening process is performed on a pixel farther downward from the pixels arranged on the left and right. The lower side image smoothing filter emphasizes the pixel of interest and the pixels of the pixel of interest that are arranged to the left and right of the pixel of interest. The left vertical side portion is configured to have a filter array that performs a weak tone process on a pixel farther upward from each pixel arranged on the left and right sides, and excludes a lower pixel than each pixel arranged on the left and right sides. The image smoothing filter emphasizes the pixel of interest and the pixels arranged above and below the pixel of interest, and performs a weakening process on a pixel farther to the right side from the pixels arranged above and below, and to the left of the pixels arranged above and below Exclude pixels The right vertical side image smoothing filter emphasizes the pixel of interest and the pixels arranged above and below the pixel of interest, and is a pixel separated to the left from the pixels arranged above and below The upper left corner image smoothing filter emphasizes the target pixel, and performs the target adjustment. The upper right corner image is configured to have a filter array that performs a weak tone process on a pixel that is lower in the radial direction on the lower side and on the right side from the pixel, and that excludes an upper pixel and a left pixel from the target pixel. The smoothing filter for emphasizing the target pixel performs a weak tone process on a pixel that is lower in the radial direction on the lower side and the left side from the target pixel, and performs an exclusion process on the upper pixel and the right pixel from the target pixel. Filter arrangement The lower left corner image smoothing filter performs enhancement processing on the target pixel, performs weak tone processing on a pixel that is separated from the target pixel in the upper side and right radial direction, The lower right corner image smoothing filter is configured to emphasize the target pixel, and the upper and left radial directions from the target pixel. The lower right corner image smoothing filter is configured to exclude the lower pixel and the left pixel. It is preferable that the pixel is further configured to have a filter array that performs a weak tone process on a pixel farther away from the target pixel and excludes a lower pixel and a right pixel from the target pixel.

According to such a configuration, the background luminance value of the pixel of interest in each peripheral image is changed using the eight types of peripheral image smoothing filters instead of the first to third peripheral image smoothing filters described above. By calculating, the calculation accuracy of the background luminance value can be further improved.
Here, for example, in the upper side image smoothing filter, excluding the pixel of interest and the pixels above the pixels arranged on the left and right of the pixel of interest is a portion corresponding to the inspection non-target portion above the upper side of the inspection target portion. Means the process of setting the luminance value of the image to 0 (zero). Accordingly, it is possible to exclude the influence of the inspection non-target portion from the generated background image.

In the present invention, in the area information determination step, an upper side determination pattern filter, a lower side determination pattern filter, a left vertical side determination pattern filter, a right vertical side determination pattern filter, and an upper left corner determination prepared in advance are prepared. 8 types of pattern filters including a pattern filter for determination, an upper right corner determination pattern filter, a lower left corner determination pattern filter, and a lower right corner determination pattern filter, and each pattern filter is applied to the partial image. It is preferable to determine the area information.
According to such a configuration, it corresponds to each position of the upper side, lower side, left vertical side, right vertical side, upper left corner, upper right corner, lower left corner, and lower right corner of the inspection target part. By determining the area information using the eight types of pattern filters, it is possible to accurately determine which position of the peripheral image is the image of the inspection target portion. It can be executed properly.

  On the other hand, the defect detection method of the present invention is a defect detection method for detecting a display defect of the display device based on a captured image obtained by capturing a display image of a display device to be inspected, and capturing the display device Then, the captured image is acquired by acquiring the captured image, the background image generating step of generating the background image from the captured image, and the difference between the captured image and the background image is taken to remove the shading. A shading correction step for generating a difference image; and a defect detection step for detecting a display defect of the display device based on the difference image, wherein the background image generation step includes selecting any pixel of interest in the captured image. The partial image having a predetermined size is an intermediate image composed only of the image of the inspection target portion of the display device, or is not inspected outside the inspection target portion. An area information determination step for determining area information as to whether the image is a peripheral image including an image of an elephant, and whether or not the luminance value for each pixel in the partial image is within a predetermined luminance range. An applied pixel determining step for recording an applied pixel determined to be within a predetermined luminance range; an intermediate image smoothing filter corresponding to the intermediate image; and a peripheral image smoothing filter corresponding to the peripheral image A smoothing filter is selected based on the area information from at least two types of smoothing filters prepared in advance, and the selected smoothing filter is applied to the application pixel of the partial image to obtain the background luminance value of the target pixel. A background luminance value calculation step for calculating, and each of the plurality of pixels of the inspection target portion as a pixel of interest, the area information determination step, the application pixel determination step, and And generating a background image by executing the background luminance value calculation step.

  The defect detection apparatus according to the present invention is a defect detection apparatus that detects a display defect of the display device based on a captured image obtained by capturing a display image of a display device to be inspected, and includes a luminous flux applied to the display device. A light source device that irradiates light, an image pickup device that picks up an image of a light beam through the display device, and a control device that drives and controls the display device and the image pickup device. The control device drives the display device to perform inspection. An image display control unit that displays an image, an image data acquisition unit that causes the imaging device to capture an inspection image displayed on the display device and acquires captured image data, and generates background image data from the captured image data A background image data generation unit, and shading from the captured image data by taking a difference between the captured image data and the background image data The background image data, comprising: a shading correction unit that generates the difference image data that has been removed; a defect detection unit that detects a display defect of the display device based on the difference image data; and a data storage unit that stores various data. In the generation unit, the partial image having a predetermined size including any target pixel in the captured image data is an intermediate image configured only with an image of the inspection target portion of the display device, or the inspection target portion An area information determination unit that determines area information as to whether the image is a peripheral image including an image of an outer non-inspection portion, and a luminance value for each pixel in the partial image data is within a predetermined luminance range An applied pixel determination unit that records an applied pixel determined to be within a predetermined luminance range by determining whether or not there is a smoothing pixel read from the data storage unit based on the area information And a background luminance value calculation unit for calculating a background luminance value of the pixel of interest by applying the data to the application pixels of the partial image data, and the data storage unit includes an intermediate unit corresponding to the intermediate unit image data At least two types of smoothing filter data including image smoothing filter data and peripheral image smoothing filter data corresponding to the peripheral image data are stored.

According to the defect detection method and the defect detection apparatus of the present invention as described above, in the defect detection process of the display device, in the same manner as the above-described shading correction method, the peripheral part of the inspection target part and the part that may have display defects The shading removal accuracy can be increased, and the display defect detection accuracy can be improved.
Further, the data storage unit includes the first to third peripheral image smoothing filters, the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, and the upper right corner of the inspection target unit. 8 types of smoothing filters corresponding to the positions of the lower left corner and the lower right corner, or the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, and the upper right corner of the part to be inspected If 8 types of pattern filter data corresponding to the positions of the lower left corner and the lower right corner are stored, these smoothing filters and pattern filters need not be calculated by calculation, and stored. It can be applied to each partial image simply by reading out a smoothing filter and a pattern filter, and smoothing and pattern matching processing can be executed quickly.

  On the other hand, the defect detection apparatus control method program according to the present invention drives and controls any one of the defect detection apparatuses, and the control apparatus constituting the defect detection apparatus can receive any attention in the captured image data. A peripheral image in which a partial image having a predetermined size including pixels is an intermediate image composed only of an image of an inspection target portion of the display device, or includes an image of an inspection non-target portion outside the inspection target portion Determination of area information as to whether it is a partial image, determination of whether or not the luminance value of each pixel in the partial image data is within a predetermined luminance range, and the application pixel determined to be within the predetermined luminance range From the recording, at least two types of smoothing filters including the intermediate part image smoothing filter corresponding to the intermediate part image data and the peripheral part image smoothing filter corresponding to the peripheral part image data Selection of a smoothing filter based on the information, calculation of a background luminance value of the pixel of interest by applying the selected smoothing filter to the application pixel of the partial image data, and a plurality of pixels of the inspection target part, respectively And generating the background image by calculating a background luminance value as the pixel of interest.

  According to the present invention, in the same manner as the above-described shading correction method, defect detection method, and defect detection apparatus, in the defect detection process of the display device, the shading removal accuracy in the peripheral portion of the inspection target portion or a portion that may have a display defect And the display defect detection accuracy can be improved. If a computer equipped with a CPU or the like is used as a control device and the computer executes the control method program, the processing speed of each process can be increased and the entire defect detection process can be executed quickly. .

[First embodiment]
A first embodiment of the present invention will be described with reference to the drawings.
[Configuration of defect detection device]
FIG. 1 is a diagram illustrating a configuration of the defect detection apparatus 1.
The defect detection apparatus 1 is an apparatus that detects a display defect of the liquid crystal panel 10 by performing a display appearance inspection of the liquid crystal panel 10 as an image display device to be inspected. As shown in FIG. 1, the defect detection device 1 includes an optical system 2, a screen 3, a CCD camera 4 as an imaging device, a panel control device 5, and a control device 6.

  Here, the liquid crystal panel 10 to be inspected is a transmissive liquid crystal panel, and has, for example, a configuration in which liquid crystal molecules are hermetically sealed between a TFT substrate and a counter substrate. The image display unit 11 transmits light and the outer peripheral frame unit 12 is provided so as to surround the outer periphery of the image display unit 11 and does not transmit light flux. The liquid crystal panel 10 is electrically connected to the panel control device 5 in a state where the liquid crystal panel 10 is installed in a panel installation unit (not shown) in the defect detection device 1, for example, and the panel control device 5 connects the TFT substrate and the counter substrate. By applying a voltage having a predetermined voltage value (including 0 V) between them, the arrangement state of the liquid crystal molecules is changed, and a predetermined optical image is formed by transmitting or blocking the incident light beam. In the present embodiment, the liquid crystal panel 10 is configured in a normally black mode in which a black display is performed by blocking an incident light beam when no voltage is applied (voltage value is 0 V). Further, the liquid crystal panel 10 may be configured in a normally white mode in which white light is displayed by transmitting all incident light beams in a state where no voltage is applied.

The optical system 2 is an optical system that irradiates the liquid crystal panel 10 with a light beam emitted from a light source and enlarges and projects the light beam via the liquid crystal panel 10 toward the screen 3. As shown in FIG. 1, the optical system 2 includes a light source device 21, a condensing lens 22 that condenses the light emitted from the light source device 21 and irradiates the liquid crystal panel 10, and the liquid crystal panel 10. And a projection lens 23 for enlarging and projecting the optical image toward the screen 3.
Among these, the light source device 21 includes a light source lamp that performs discharge light emission and a reflector that reflects a light beam emitted from the light source lamp, although not specifically illustrated. The light beam emitted from the light source lamp is emitted toward the condenser lens 22 by the reflector.
The light source device 21 is not limited to the discharge light source device, and various solid light emitting elements such as an LED (Light Emitting Diode) element, a laser diode, an organic EL (Electro Luminescence) element, and a silicon light emitting element are adopted. Also good.

  The screen 3 is configured as a reflective screen that reflects and projects an optical image (display image) enlarged and projected by the projection lens 23. The screen 3 is not limited to a reflective screen, and may be configured as a transmissive screen that transmits and projects an incident optical image.

Under the control of the control device 6, the CCD camera 4 images the projection surface of the screen 3 and outputs a signal corresponding to the captured image to the control device 6. Although not specifically shown, this CCD camera 4 is a CCD that is an area sensor, a condensing lens that condenses the light and irradiates the CCD, and under the control of the control device 6, A shutter or the like that can change the light reception time is provided.
The CCD of the CCD camera 4 preferably has a resolution higher than that of the liquid crystal panel 10.
The panel control device 5 applies a voltage having a predetermined voltage value to the liquid crystal panel 10 under the control of the control device 6 to form a predetermined optical image on the liquid crystal panel 10.

FIG. 2 is a block diagram illustrating a schematic configuration of the control device 6.
The control device 6 is composed of, for example, a computer including a CPU (Central Processing Unit) that reads and executes a predetermined program, and controls the entire defect detection device 1. As illustrated in FIG. 2, the control device 6 includes a control unit 61 and a memory 62.
The control unit 61 is a part that executes predetermined processing (defect detection processing) in accordance with a control program stored in the memory 62, and includes an image display control unit 611, an image data acquisition unit 612, a background image generation unit 613, and the like. , A shading correction unit 614, a defect detection unit 615, and the like.

The image display control unit 611 reads voltage value information related to the voltage value applied to the liquid crystal panel 10 stored in the memory 62, and a predetermined control signal for driving the liquid crystal panel 10 with an inspection voltage value based on the voltage value information. Is output to the panel control device 5. The panel control device 5 drives the liquid crystal panel 10 with the inspection voltage value according to the control signal output from the image display control unit 611, and causes the liquid crystal panel 10 to form an inspection image.
In the present embodiment, the inspection voltage value is a voltage that drives the liquid crystal panel 10 to transmit a light beam with a transmittance of 50% to form an inspection image (a gray image having an intermediate gradation value). Is set to a value.

When an inspection image is formed on the liquid crystal panel 10, the image data acquisition unit 612 outputs a predetermined control signal to the CCD camera 4 to display the projection surface (including the inspection image) of the screen 3 on the CCD camera 4. Let's take an image. The image data acquisition unit 612 receives an electrical signal output from the CCD camera 4 and converts it into a computer-readable signal (digital signal), and includes information on the pixel value (luminance value) for each pixel. Inspection image data (captured image data) P is acquired. Then, the image data acquisition unit 612 stores the acquired inspection image data P in the memory 62.
As shown in FIG. 3, the acquired inspection image data P includes an inspection target part P1 corresponding to the image display unit 11 of the liquid crystal panel 10, and an inspection non-target part P2 corresponding to the outer peripheral frame part 12 of the liquid crystal panel 10. Is stored as image data.

The background image generation unit 613 generates background image data P ′ (see FIG. 12) from the inspection image data P acquired by the image data acquisition unit 612 and stored in the memory 62. The background image generation unit 613 includes an area information determination unit 613A, a filter application pixel determination unit 613B, and a background luminance value calculation unit 613C.
The area information determination unit 613A determines whether the partial image data Q having a predetermined size including any target pixel in the inspection image data P is the intermediate image data Q1 configured only by the inspection target part P1, or the inspection non-target part Area information as to whether the image data is the peripheral image data Q2 configured to include P2 is determined and stored in the memory 62. Here, the partial image data Q is, for example, image data composed of 81 pixels of 9 × 9, and the central pixel is the pixel of interest. The area information is determined by pattern matching using a pattern filter described later.
The filter application pixel determination unit 613B determines whether or not the luminance value for each pixel in the partial image data Q is within a predetermined luminance range, and stores it in the memory 62.
The background luminance value calculation unit 613C calculates the background luminance value of the pixel of interest by applying smoothing filter data (described later) read from the memory 62 based on the area information to the partial image data Q.
Then, the background image generation unit 613 calculates a background luminance value using all the pixels of the inspection target unit P1 in the inspection image data P as the target pixels, and generates and generates background image data P ′ based on the background luminance value. The background image data P ′ is stored in the memory 62.

The shading correction unit 614 calculates a difference between the background image data P ′ and the inspection image data P for each corresponding pixel (shading correction). Then, the shading correction unit 614 performs shading correction, resulting in uneven illumination of the light beam emitted from the light source device 21 or a decrease in the peripheral light amount of the image by the lenses 22 and 23 constituting the optical system 2. Difference image data R from which shading (components other than display defect components) has been removed is generated. The shading correction unit 614 includes a difference image data generation unit 614A.
The difference image data generation unit 614A generates a difference image data R by calculating a difference between the pixel values for each pixel corresponding to each other in the background image data P ′ and the inspection image data P stored in the memory 62. The difference image data R is stored in the memory 62.

The defect detection unit 615 reads the difference image data R stored in the memory 62, performs a general defect enhancement process for enhancing display defects based on the difference image data R, and displays spots, unevenness, streaks, and the like. A pixel position corresponding to the defect is specified.
The memory 62 stores pattern filter data 621 and smoothing filter data 622 in addition to a predetermined control program, information for executing processing by the control unit 61, data processed by the control unit 61, and the like. To do.

  The pattern filter data 621 stores eight types (pattern A to pattern H) of pattern filters as shown in FIG. These pattern filters PT1 to PT8 are composed of 81 coefficients of 9 × 9 corresponding to the partial image data Q, and the upper side, the lower side, the left vertical side, the right vertical side, and the upper side of the inspection target part P1. It matches the partial image data Q at any one of the left corner, upper right corner, lower left corner, and lower right corner, and indicates whether the partial image data Q is the peripheral image data Q2 at any position. The information determination unit 613A determines. That is, when each pattern filter PT1 to PT8 is multiplied by the corresponding pixel value of the partial image data Q, the total value when the matching pattern filters PT1 to PT8 are multiplied is the other pattern filter PT1 to PT8. By being larger than the multiplied value, it is possible to determine at which position the peripheral image data Q2 is. On the other hand, if the total value obtained by multiplying the pattern filters PT1 to PT8 and the corresponding pixel values of the partial image data Q is smaller than a predetermined threshold, the area information determination unit 613A determines the partial image data. Q is determined as intermediate image data Q1.

In the upper side determination pattern filter PT1, the coefficients from the top to the fourth line are all set to negative values (-1/36), and the coefficients from the top to the fifth line are all set to 0 (zero). To 6th to 9th lines, all the coefficients are set to a positive value (1/36).
Contrary to the upper side judgment pattern filter PT1, the lower side judgment pattern filter PT2 is set such that all the coefficients from the top to the fourth line are set to positive values (1/36), and the fifth line from the top. Are set to 0 (zero), and each coefficient from the 6th line to the 9th line from the top has a filter array in which all the coefficients are set to negative values (-1/36).
In the left vertical side determination pattern filter PT3, the coefficients from the left to the fourth column are all set to negative values (-1/36), and the coefficients from the left to the fifth column are all set to 0 (zero). Each of the coefficients from the 6th column to the 9th column from the left has a filter array in which all the coefficients are set to a positive value (1/36).
In contrast to the left vertical side determination pattern filter PT3, the right vertical side determination pattern filter PT4 has all the coefficients in the fourth column from the right set to negative values (-1/36), and from the right It is configured with a filter array in which all the coefficients in the fifth column are set to 0 (zero) and all the coefficients from the sixth column to the ninth column from the right are all set to positive values (1/36). Yes.

In the upper left corner determination pattern filter PT5, each coefficient from the top to the fourth line and each coefficient from the left to the fourth column are all set to negative values (-1/56), and from the bottom to the fourth line In addition, the filter arrangement is such that the coefficients from the right to the fourth column (lower right coefficient) are all set to positive values (1/16) and the remaining coefficients are set to 0 (zero). .
The upper right corner determination pattern filter PT6 is bilaterally symmetric with the upper left corner determination pattern filter PT5, and all the coefficients from the top to the fourth row and the coefficients from the right to the fourth column are all negative values ( -1/56), the coefficients from the bottom to the fourth row and from the left to the fourth column are all set to positive values (1/16), and the remaining coefficients are 0 (zero) ) To the filter array set.
The lower left corner determination pattern filter PT7 is configured vertically symmetrically with the upper left corner determination pattern filter PT5, and all the coefficients from the bottom to the fourth row and the coefficients from the left to the fourth column are all negative values (− 1/56), the coefficients from the top to the fourth line and from the right to the fourth column (upper right coefficients) are all set to positive values (1/16), and the remaining coefficients are 0 (zero) The filter arrangement is set to
The lower right corner determination pattern filter PT8 is configured symmetrically with the lower left corner determination pattern filter PT7, and all the coefficients from the bottom to the fourth row and the coefficients from the right to the fourth column are all negative values ( -1/56), the coefficients from the top to the 4th row and from the left to the 4th column are all set to positive values (1/16), and the remaining coefficients are 0 (zero) ) To the filter array set.

  On the other hand, the smoothing filter data 622 stores nine types of smoothing filters as shown in FIGS. Among the nine types of smoothing filters F1 to F9, the smoothing filters F1 to F8 are the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, and the upper right determined by the pattern filters PT1 to PT8. Smoothing is performed on the peripheral image data Q2 at each of the corner, lower left corner, and lower right corner, and the smoothing filter F9 performs smoothing on the intermediate image data Q1. . The smoothing filters F1 to F9 are composed of 9 × 9 81 coefficients corresponding to the partial image data Q, and each coefficient of the smoothing filters F1 to F9 corresponds to the background luminance value calculation unit 613C. It is multiplied by each pixel of the partial image data Q. The smoothing filter F9 is configured to have a filter array in which all the 9 × 9 coefficients are set to the same coefficient value (1.0), and smoothing is performed without weighting each pixel of the intermediate image data Q1. .

  Then, in the upper side image smoothing filter F1 corresponding to the peripheral image data Q2 on the upper side of the inspection target part P1, all the coefficients from the top to the fourth line are set to 0 (zero), and five lines from the top. Each coefficient of the eye is set to the maximum value (5.0), and each coefficient in the 6th to 9th lines from the top is smaller than the coefficient in the 5th line, and gradually becomes a smaller value in the lower line ( 4.4 to 0.7). That is, the center coefficient of the fifth row from the top in the upper-side image smoothing filter F1 corresponds to the pixel of interest of the peripheral image data Q2, and each of the fifth row from the top set to the maximum value. The coefficient is set so that the pixel of interest and each pixel arranged on the left and right of the pixel of interest are enhanced (weighting is increased). The coefficients of the sixth to ninth lines from the top in the smooth filter F1 for the upper side image are pixels that are farther downward from each pixel (each pixel in the fifth line) arranged side by side including the target pixel. The weak coefficient is set to weaken (weighting is weakened), and the coefficient of 0 from the first row to the fourth row from the top excludes the pixel above the target pixel, that is, the pixel on the non-inspection portion P2 side. It is set to be.

The lower-side image smoothing filter F2 corresponding to the peripheral image data Q2 on the lower side portion of the inspection target portion P1 is configured to have a filter arrangement vertically symmetrical with the upper-side image smoothing filter F1. That is, the lower-side image smoothing filter F2 is set so as to emphasize the pixel of interest in the peripheral image data Q2 and the pixels arranged on the left and right of the pixel of interest. The farther away pixels are set to be subjected to the weak tone process, and the lower pixels than the target pixel (the pixels on the inspection non-target part P2 side) are set to be excluded.
Further, the left vertical side image smoothing filter F3 corresponding to the peripheral image data Q2 of the left vertical side of the inspection target part P1 is a filter array obtained by rotating the upper side image smoothing filter F1 by 90 ° counterclockwise. It is comprised. That is, the left vertical side image smoothing filter F3 is set so as to emphasize the target pixel of the peripheral image data Q2 and the pixels arranged above and below the target pixel, and from the pixels arranged above and below including the target pixel. The pixel that is farther to the right is set to be subjected to the weak tone process, and the pixel that is on the left side of the target pixel (the pixel on the non-inspection part P2 side) is set to be excluded.
Further, the right vertical side image smoothing filter F4 corresponding to the peripheral image data Q2 of the right vertical side of the inspection target portion P1 has a filter arrangement symmetrical to the left vertical side image smoothing filter F3. It is configured. In other words, the right vertical side image smoothing filter F4 is set so as to emphasize the pixel of interest in the peripheral image data Q2 and the pixels arranged above and below the pixel of interest, and from the pixels arranged above and below including the pixel of interest. The pixel that is farther to the left is set to be subjected to the weak tone process, and the pixel that is on the right side of the pixel of interest (the pixel on the inspection non-target portion P2 side) is set to be excluded.

  On the other hand, the upper left corner image smoothing filter F5 corresponding to the peripheral image data Q2 at the upper left corner of the inspection target portion P1 has all the coefficients from the top to the fourth row and the coefficients from the left to the fourth column. It is set to 0 (zero), the center coefficient (line 5-column 5) is set to the maximum value (5.0), and each coefficient below and to the right of this center coefficient is The filter arrangement is such that the value gradually decreases as it goes downward (4.4 to 0.1). That is, the center coefficient in the upper left corner image smoothing filter F5 corresponds to the target pixel of the peripheral image data Q2, and is set to emphasize the target pixel (increase the weighting). Then, each of the coefficients on the lower side and the right side of the center coefficient in the upper left corner image smoothing filter F5 is subjected to weakening processing (weakening the weighting) as the pixel is separated from the target pixel in the lower right radial direction. Each coefficient of 0 in the fourth row from the top and the fourth column from the left is set so as to exclude the pixels on the upper side and the left side of the pixel of interest (pixels on the non-inspection part P2 side). Yes.

The upper right corner image smoothing filter F6 corresponding to the peripheral image data Q2 at the upper right corner of the inspection target portion P1 is configured to have a filter arrangement symmetrical to the upper left corner image smoothing filter F5. Yes. That is, the upper right corner image smoothing filter F6 is set so as to emphasize the target pixel of the peripheral image data Q2, and is set so that the pixel farther away from the target pixel in the lower left radial direction is subjected to the weak tone processing. The pixel is set to exclude the pixels on the upper side and the right side of the pixel of interest (pixels on the inspection non-target portion P2 side).
Further, the lower left corner image smoothing filter F7 corresponding to the peripheral image data Q2 in the lower left corner of the inspection target portion P1 is configured to have a filter arrangement vertically symmetrical with the upper left corner image smoothing filter F5. Yes. That is, the lower left corner image smoothing filter F7 is set so as to emphasize the target pixel of the peripheral image data Q2, and is set so as to perform a weak tone process on a pixel farther in the upper right radial direction than the target pixel. The pixels on the lower side and the left side of the target pixel (pixels on the inspection non-target portion P2 side) are set to be excluded.
Further, the lower right corner image smoothing filter F8 corresponding to the peripheral image data Q2 at the lower right corner of the inspection target portion P1 has a filter arrangement that is symmetrical to the lower left corner image smoothing filter F7. Has been. That is, the lower right corner image smoothing filter F8 is set so as to emphasize the target pixel of the peripheral image data Q2, and is set so that the pixel farther from the target pixel in the upper left radial direction is subjected to the weak tone processing. The pixel is set to exclude the pixels on the lower side and the right side of the pixel of interest (pixels on the non-inspection part P2 side).

(Defect detection method)
Next, a defect detection method using the above-described defect detection apparatus 1 will be described with reference to the drawings.
FIG. 8 is a flowchart for explaining the defect detection method.
In the following description, it is assumed that the liquid crystal panel 10 is installed in a panel installation unit (not shown) in the defect detection device 1 and the liquid crystal panel 10 and the panel control device 5 are electrically connected.
The operator operates an operation unit (not shown) of the control device 6 and performs setting input for executing defect detection of the liquid crystal panel 10. And the control part 61 of the control apparatus 6 reads the control program memorize | stored in the memory 62 according to the operation signal output from an operation part, According to a control program, as shown below, the defect of the liquid crystal panel 10 is shown. Perform detection processing.

First, the control unit 61 turns on the light source device 21, introduces the light beam emitted from the light source device 21 into the liquid crystal panel 10, and the image display control unit 611 of the control unit 61 causes the liquid crystal panel 10 to display an inspection image. A control signal is output to the panel control device 5. By this process, the inspection image transmitted through the liquid crystal panel 10 is projected and displayed on the screen 3. In this state, the image data acquisition unit 612 of the control unit 61 outputs a predetermined control signal to the CCD camera 4 to cause the CCD camera 4 to image the projection surface of the screen 3 and acquire inspection image data P (processing ST1: Inspection image data acquisition process). Then, the image data acquisition unit 612 stores the acquired inspection image data P in the memory 62.
After the processing ST1, the background image generation unit 613 of the control unit 61 reads the inspection image data P stored in the memory 62, and generates background image data P ′ based on the inspection image data P (processing ST2: background). Image generation step).

FIG. 9 is a flowchart for explaining the background image generation step ST2.
In the process ST2, first, the area information determination unit 613A of the background image generation unit 613 reads the pattern filter data 621 stored in the memory 62, and uses the pattern filters PT1 to PT8 stored in the pattern filter data 621 as partial image data. Applies to Q. Specifically, the area information determination unit 613A displays the luminance value (numerical value in the figure) of each pixel in the 9 × 9 partial image data Q centered on an arbitrary target pixel as illustrated in FIG. Multiply the corresponding coefficients of the pattern filters PT1 to PT8 shown in FIG. This total value is calculated for each of the pattern filters PT1 to PT8 for one partial image data Q as shown in the following table.
Subsequently, the area information determination unit 613A determines whether or not the total value obtained by multiplying the luminance value and the coefficient of each pixel exceeds a predetermined threshold (for example, 0.5), and the total value is If it is smaller than the threshold value, the partial image Q is determined as the intermediate image data Q1. On the other hand, when the total value exceeds the threshold value, the partial image Q is determined as the peripheral image data Q2, and the pattern filters PT1 to PT8 having the maximum total value are determined, and the pattern filter having the maximum value is determined. The corresponding position and the position of the peripheral image data Q2 are associated with each other and stored as area information in the memory 62 (process ST21: area information determination step).

  As described above, in the process ST21, it is determined whether the partial image data Q is the intermediate image data Q1 or the peripheral image data Q2, and if the partial image data Q is the peripheral image data Q2, the inspection target part P1 is determined. It is determined whether the image is at the position of the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, the upper right corner, the lower left corner, or the lower right corner. Specifically, in the case of the partial image data Q exemplified as the partial image 1 in FIG. 10, the total value (0.932785) when the upper side determination pattern filter PT1 that is the pattern A is applied exceeds the threshold and is the maximum. Therefore, it is determined that the image data is the peripheral image data Q2 of the upper side (pattern A) of the inspection target portion P1. Further, in the case of the partial image data Q illustrated as the partial image 2 in FIG. 10, the total value (0.877976) when the upper left corner determination pattern filter PT5 that is the pattern E is applied exceeds the threshold and is maximum. Therefore, it is determined that the image data Q2 is the peripheral image data of the upper left corner (pattern E) of the inspection target portion P1. On the other hand, in the case of the partial image data Q exemplified as the partial image 3 in FIG. 10, the total value when the pattern filters PT1 to PT8 (pattern A to pattern H) are applied does not exceed the threshold value. It is determined that the image data is Q1 (other than the outer periphery).

After the process STST21, the filter application pixel determination unit 613B of the background image generation unit 613 determines the filter application pixel based on the luminance value of each pixel of the partial image data Q, and stores the position information of the determined filter application pixel in the memory 62. (Process ST22: filter application pixel determination step).
In this process ST22, the average brightness value is calculated from the brightness value of each pixel of the partial image data Q, and the brightness value is lower than the calculated lower limit threshold value and the brightness value larger than the average brightness value. An upper limit threshold is set, and it is determined whether there is a luminance value of each pixel between these lower limit threshold and upper limit threshold (within a predetermined luminance range). Then, a pixel having a luminance value that falls between the lower limit threshold and the upper limit threshold is determined as a filter application pixel. Specifically, for the partial image 1 in FIG. 10, for the calculated average luminance value (79.0), the lower limit threshold is set to average luminance value −20 (59.0), and the upper limit threshold is set to average luminance value +40 ( As a result of setting to 119.0), the pixels indicated by hatching in the partial image 1 of FIG. 11 are determined as the filter application pixels. For the partial image 2 in FIG. 10, for the calculated average luminance value (61.9), the lower limit threshold is set to average luminance value −20 (41.9), and the upper limit threshold is set to average luminance value +40 (101.9). As a result of setting, in the partial image 2 of FIG. 11, pixels indicated by hatching are determined as filter application pixels. For the partial image 3 in FIG. 10, for the calculated average luminance value (127.9), the lower limit threshold is set to average luminance value −20 (107.9), and the upper limit threshold is set to average luminance value +40 (167.9). As a result of setting, in the partial image 3 of FIG. 11, pixels indicated by hatching are determined as filter application pixels.

After the process ST22, the background luminance value calculation unit 613C of the background image generation unit 613 reads the area information of the partial image data Q, the position information of the filter application pixel, and the smoothing filter data 622 stored in the memory 62, and performs smoothing. The smoothing filters F1 to F9 stored in the filter data 622 are selected according to the area information, and the selected smoothing filter is applied to the filter application pixels of the partial image data Q (processing ST23: smoothing filter application step).
Specifically, the background luminance value calculation unit 613C selects the smoothing filter F9 when the partial image data Q is the intermediate image data Q1, and when the partial image data Q is the peripheral image data Q2. Selects a smoothing filter corresponding to the position of the peripheral image data Q2 among the smoothing filters F1 to F8. Then, the filter application pixels in the partial image data Q are multiplied by the coefficients of the smoothing filters F1 to F9 corresponding to these pixel positions, and the multiplied values are summed. Further, by dividing the sum of the luminance value of the multiplied filter application pixel and the coefficients of the smoothing filters F1 to F9 by the sum of the coefficients of the multiplied smoothing filters F1 to F9, the pixel of interest in the partial image data Q is divided. A background luminance value is calculated.
Then, the background image generation unit 613 sets all the pixels of the inspection target portion P1 in the inspection image data P as the target pixel, and repeats the above steps ST21 to ST23 for the partial image data Q centered on the target pixel. Then, the background luminance value of all the pixels is calculated, background image data P ′ based on the background luminance value is generated, and the generated background image data P ′ is stored in the memory 62.

After the process ST2 as described above, the shading correction unit 614 of the control unit 61 performs shading correction as described below (process ST3: shading correction process).
FIG. 12 is a diagram for explaining the shading correction step ST3.
Specifically, FIG. 12A shows the inspection image data P acquired by the image data acquisition unit 612 in the process ST1. FIG. 12B shows the background image data P ′ generated by the background image generation unit 613 in the process ST2. FIG. 12C shows the difference image data R generated by the difference image data generation unit 614A in process ST3. In addition, the upper diagram in FIG. 12 shows images based on the inspection image data P, the background image data P ′, and the difference image data R. The lower diagram in FIG. 12 shows the luminance value (pixel value) for each pixel on the predetermined horizontal line HL in the upper diagram. That is, the horizontal axis indicates the pixel position, and the vertical axis indicates the luminance value.

In the process ST3, the difference image data generation unit 614A reads the inspection image data P and the background image data P ′ stored in the memory 62, and the background image data P from each pixel value of the inspection image data P for each corresponding pixel. Difference of each pixel value of 'is taken and difference image data R is generated. Then, the difference image data generation unit 614A stores the generated difference image data R in the memory 62.
Then, by the shading correction step ST3, as shown in FIG. 12C, difference image data R from which shading has been removed is generated.

After the process ST3, the defect detection unit 615 of the control unit 61 reads the difference image data R stored in the memory 62 and performs a general defect enhancement process for enhancing the display defect component based on the difference image data R. Then, the pixel position that is a display defect is specified (process ST4: defect detection step).
For example, the defect detection unit 615 adds a pixel value of each pixel located in the horizontal direction and the vertical direction with respect to the pixel of interest for each pixel in the difference image data R using a predetermined filter. And the defect detection part 615 specifies the pixel position which is a display defect component by comparing each integrated value integrated | accumulated for every pixel with the preset threshold value.

The first embodiment described above has the following effects.
In the present embodiment, based on the inspection image data P acquired in the inspection image data acquisition step ST1, background image data P ′ is generated from the inspection image data P in the background image generation step ST2, and inspection is performed in the shading correction step ST3. By taking the difference between the image data P and the background image data P ′ to generate the difference image data R and removing the shading, the pixel position may be shifted between the inspection image data P and the background image data P ′. In addition, it is possible to obtain a difference between pixels corresponding to each other, to eliminate the influence of variations in illumination conditions, and to match the shading curve, and to appropriately remove the shading from the inspection image data P. it can. Therefore, it is possible to accurately detect a display defect having a minute contrast.

In the background image generation step ST2, it is determined in the area information determination step ST21 whether the partial image data Q is the intermediate portion image data Q1 or the peripheral portion image data Q2, and in the smoothing filter application step ST23 By applying the smoothing filters F1 to F8 for the peripheral image to the partial image data Q2, it is possible to improve the accuracy of shading correction at the outer peripheral portion of the inspection target portion P1.
Further, in the filter application pixel determination step ST22 of the background image generation step ST2, by excluding pixels having luminance values that do not fall between the lower limit threshold value and the upper limit threshold value, By excluding the pixels of the inspection non-target portion P2 that is a transmission portion or a non-light emitting portion from the filter application pixels, the background image data P ′ and the inspection image data P are not included in the background image data P ′. It is possible to prevent the defect component from being removed from the difference image data R obtained by taking the difference between.

  In the smoothing filter application step ST23, the peripheral image data Q2 is converted into the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, the upper right corner, the lower left corner, and the lower side. A boundary portion between the inspection target part P1 and the inspection non-target part P2 by applying a smoothing filter appropriately selected from the eight types of smoothing filters F1 to F8 depending on which position of the right corner is the image. Can be generated, and the shading removal accuracy in the outer peripheral portion of the inspection target portion P1 can be further enhanced.

[Second Embodiment]
A second embodiment of the present invention will be described with reference to the drawings.
In the second embodiment, the configuration of the smoothing filter is different from that of the first embodiment, and the configuration of the defect detection apparatus 1, the defect detection method, and the like are substantially the same as those of the first embodiment. Hereinafter, the differences will be described in detail.
FIG. 13 is a diagram illustrating the smoothing filters F11 to F14 of the second embodiment.
The four types of smoothing filters F11 to F14 are stored in the smoothing filter data 622 stored in the memory 62. Among these, the three types of smoothing filters F11 to F13 (first to third smoothing filters) are: Peripheral images at the positions of the upper side, lower side, left vertical side, right vertical side, upper left corner, upper right corner, lower left corner, and lower right corner determined by the pattern filters PT1 to PT8 The smoothing filter F14 is the same as the smoothing filter F9, and smoothes the intermediate image data Q1.

  The first smoothing filter F11 corresponds to the peripheral image data Q2 of the upper side and the lower side of the inspection target part P1, and each coefficient in the fifth row from the top is set to the maximum value (5.0). , Each coefficient from the top to the fourth line is set to be smaller than the coefficient of the fifth line and gradually becomes a smaller value (4.4 to 0.7) in the upper line, from the sixth line to the ninth line Each filter is configured to have a filter array that is set to be smaller than the coefficient in the fifth row and gradually become smaller in the lower row (4.4 to 0.7). That is, the center coefficient in the fifth row from the top in the first smoothing filter F11 corresponds to the target pixel of the peripheral image data Q2, and each coefficient in the fifth row from the top set to the maximum value is The pixel of interest and each pixel on the left and right of the pixel of interest are set to be emphasized (weighting is increased). The coefficients from the top to the fourth line and the coefficients from the top to the sixth line in the first smoothing filter F11 are the pixels arranged in the left and right including the pixel of interest (the pixels in the fifth line). ) From the upper side and the lower side, respectively, so that the gradation process is performed (weighting is weakened).

  The second smoothing filter F12 corresponds to the peripheral image data Q2 of the left vertical side and right vertical side of the inspection target part P1, and each coefficient in the fifth column from the left is set to the maximum value (5.0). As well as being set, each coefficient from the left to the fourth column is set to be smaller than the coefficient in the fifth column and gradually decreases to the left column (4.4 to 0.7), and from the right to the fourth column The filter array is configured so that each coefficient is smaller than the coefficient in the fifth column and gradually decreases to the right column (4.4 to 0.7). That is, the center coefficient in the fifth column from the left in the second smoothing filter F12 corresponds to the target pixel of the peripheral image data Q2, and each coefficient in the fifth column from the left set to the maximum value is The pixel of interest and each pixel arranged above and below the pixel of interest are set to be emphasized (weighting is increased). The coefficients from the left to the fourth column and the coefficients from the right to the fourth column in the second smoothing filter F12 are on the left side from the pixels that line up and down including the pixel of interest (the pixels in the fifth column). In addition, it is set so that the pixels that are farther to the right side are subjected to weak gradation processing (weighting is weakened).

  The third smoothing filter F13 corresponds to the peripheral image data Q2 of the upper left corner, the upper right corner, the lower left corner, and the upper right corner of the inspection target portion P1. The fifth row (center) coefficient is set to the maximum value (5.0), and there is a filter array that gradually decreases from the center coefficient toward the radial direction (4.4 to 0.1). Configured. That is, the center coefficient in the third smoothing filter F13 corresponds to the target pixel of the peripheral image data Q2, and is set to emphasize the target pixel (increase the weighting), and is more radial than the target pixel. It is set so that the pixels farther outward are subjected to weak gradation processing (weighting is weakened).

According to 2nd Embodiment mentioned above, in addition to the effect similar to 1st Embodiment, there exist the following effects.
Further, the peripheral image data Q2 is any position of the upper side, the lower side, the left vertical side, the right vertical side, the upper left corner, the upper right corner, the lower left corner, and the lower right corner of the inspection target portion P1. By applying a smoothing filter appropriately selected from the three types of smoothing filters F11 to F13 depending on whether the image is a background image, the background image data P that emphasizes the boundary portion between the inspection target part P1 and the inspection non-target part P2 'Can be generated, and even with a small number of smoothing filters, the shading removal accuracy in the outer peripheral portion of the inspection target portion P1 can be further increased.

It should be noted that the present invention is not limited to the above-described embodiments, and modifications, improvements, and the like within the scope that can achieve the object of the present invention are included in the present invention.
In each of the above embodiments, eight types of smoothing filters F1 to F8 or three types of smoothing filters F11 to F13 are prepared in advance as smoothing filters for the peripheral image data Q2, but the number of smoothing filters and their filter arrangement are particularly The present invention is not limited, and two types of smoothing filters including a smoothing filter for upper and lower sides including corners and a smoothing filter for vertical sides may be applied. Furthermore, four to seven types of smoothing filters may be applied to the peripheral image data Q2, and nine or more types of smoothing filters may be applied.

In the embodiment, the pattern filters PT1 to PT8 are applied and the threshold value is used as a method for determining the area information as to which position of the inspection target portion P1 the partial image data Q is. The determination of information is not limited to the pattern matching, and other appropriate methods can be adopted.
In the above embodiment, the partial image data Q employs an image size composed of 81 pixels of 9 × 9. However, the present invention is not limited to this, and the partial image size may be 3 × 3 or 5 ×. Any image size such as 5, 7 × 7 can be employed.
Moreover, in the said embodiment, the coefficient used for the pattern filter and the smoothing filter is one illustration, and the coefficient of these filters can be set arbitrarily.

Although the best configuration for carrying out the present invention has been disclosed in the above description, the present invention is not limited to this. That is, the invention has been illustrated and described primarily with respect to particular embodiments, but it is not intended to depart from the technical concept and scope of the invention. Various modifications can be made by those skilled in the art in terms of materials, quantity, and other detailed configurations.
Therefore, the description limited to the shape, material, etc. disclosed above is an example for easy understanding of the present invention, and does not limit the present invention. The description by the name of the member which remove | excluded the limitation of one part or all of such restrictions is included in this invention.

The figure which shows the structure of the defect detection apparatus of this invention. The block diagram which shows schematic structure of the control apparatus in the said defect detection apparatus. The figure which shows the picked-up image imaged with the said defect detection apparatus. The figure which shows the pattern filter memorize | stored in the said control apparatus. The figure which shows the smoothing filter memorize | stored in the said control apparatus. The figure which shows the smoothing filter memorize | stored in the said control apparatus. The figure which shows the smoothing filter memorize | stored in the said control apparatus. The flowchart explaining the defect detection method in the said defect detection apparatus. 7 is a flowchart for explaining a background image generation procedure in the defect detection method. The figure for demonstrating the partial image in the said background image generation procedure. The figure for demonstrating the filter application pixel in the said partial image. The figure for demonstrating the shading correction process in the said defect detection apparatus. The figure which shows the smoothing filter of 2nd Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Defect detection apparatus, 10 ... Liquid crystal panel (display device), 21 ... Light source device, 4 ... CCD camera (imaging device), 6 ... Control apparatus, 61 ... Control part, 62 ... Memory (data storage part), 611 ... Image display control unit, 612 ... Image data acquisition unit, 613 ... Background image generation unit, 614 ... Shading correction unit, 615 ... Defect detection unit, 613A ... Area information determination unit, 613B ... Filter application pixel determination unit, 613C ... Background luminance Value calculation unit, 621 ... pattern filter data, 622 ... smooth filter data, F1 to F9, F11 to F14 ... smoothing filter, PT1 to PT8 ... pattern filter, P ... inspection image data (captured image), P1 ... inspection object part, P2 ... non-inspection part for inspection, P '... background image data (background image), Q ... partial image data (partial image), Q1 ... intermediate part image data ( (Intermediate image), Q2 ... peripheral image data (peripheral image), R ... difference image data (difference image), ST1 ... inspection image data acquisition step, ST2 ... background image generation step, ST21 ... area information determination step, ST22 ... Filter application pixel determination step, ST23 ... Smoothing filter application step, ST3 ... Shading correction step, ST4 ... Defect detection step.

Claims (9)

A shading correction method for removing shading from the captured image by taking a difference between a captured image acquired by capturing a display image of a display device to be inspected and a background image generated from the captured image,
In the captured image, a partial image of a predetermined size including an arbitrary pixel of interest is an intermediate image composed only of an image of an inspection target portion of the display device, or an inspection non-target portion outside the inspection target portion An area information determination step for determining area information as to whether the peripheral image is configured to include the image of
An applied pixel determination step of determining whether or not the luminance value for each pixel in the partial image is within a predetermined luminance range and recording the applied pixel determined to be within the predetermined luminance range;
A smoothing filter based on the area information is obtained from at least two types of smoothing filters prepared in advance, including an intermediate image smoothing filter corresponding to the intermediate image and a peripheral image smoothing filter corresponding to the peripheral image. And a background luminance value calculating step of calculating a background luminance value of the pixel of interest by applying the selected smoothing filter to the application pixel of the partial image.
A shading correction method, wherein the background image is generated by executing the area information determination step, the applied pixel determination step, and the background luminance value calculation step with each of the plurality of pixels of the inspection target portion as a target pixel. The shading correction method according to claim 1,
In the area information determination step,
The partial image is a peripheral image at any position of the upper side, lower side, left vertical side, and right vertical side of the inspection target part, or the partial image is the upper left corner, upper right of the inspection target part Determine the area information of the peripheral image at any of the four corners of the corner, lower left corner and lower right corner,
The peripheral image smoothing filter is:
A first peripheral image smoothing filter corresponding to the peripheral image of the upper side and the lower side;
A second peripheral image smoothing filter corresponding to the peripheral image of the left vertical side and the right vertical side;
A third peripheral image smoothing filter corresponding to the peripheral images of the four corners;
At least three types including
In the background luminance value calculation step,
Applying the first peripheral image smoothing filter to the application pixels of the peripheral image of the upper side and the lower side,
Applying a second peripheral image smoothing filter to the application pixels of the peripheral image of the left vertical side and the right vertical side;
A shading correction method, wherein a background luminance value of the pixel of interest in each peripheral image is calculated by applying a third peripheral image smoothing filter to the application pixel of the peripheral image of the four corners. The shading correction method according to claim 2,
The first peripheral image smoothing filter is a filter array that emphasizes the pixel of interest and the pixels arranged on the left and right of the pixel of interest, and performs a weak adjustment process on pixels that are separated upward and downward from the pixels arranged on the left and right Comprising
The second peripheral image smoothing filter has a filter array in which the pixel of interest and the pixels arranged above and below the pixel of interest are emphasized, and the pixels farther left and right from the pixels arranged above and below are subjected to weakening processing. Configured with
The third peripheral image smoothing filter is configured to have a filter array for emphasizing the target pixel and performing a weak adjustment process on a pixel farther away from the target pixel in the radial direction. Correction method. The shading correction method according to claim 1,
In the area information determination step,
The partial image is a peripheral image at any position of the upper side, lower side, left vertical side, and right vertical side of the inspection target part, or the partial image is the upper left corner, upper right of the inspection target part Determine the area information of the peripheral image at any of the four corners of the corner, lower left corner and lower right corner,
The peripheral image smoothing filter is:
A smooth filter for the upper side image corresponding to the peripheral image of the upper side,
A smooth filter for the lower side image corresponding to the peripheral image of the lower side,
A smooth filter for a left vertical side image corresponding to a peripheral image of the left vertical side,
A smooth filter for a right vertical side image corresponding to a peripheral image of the right vertical side;
A smooth filter for the upper left corner image corresponding to the peripheral image of the upper left corner;
A smooth filter for the upper right corner image corresponding to the peripheral image of the upper right corner;
A smooth filter for the lower left corner image corresponding to the peripheral image of the lower left corner,
A smooth filter for the lower right corner image corresponding to the peripheral image of the lower right corner;
At least 8 types including
In the background luminance value calculation step,
Applying the upper side image smoothing filter to the applied pixels of the upper side peripheral image,
Applying the lower side image smoothing filter to the application pixels of the lower side peripheral image,
Applying the left vertical side image smoothing filter to the application pixel of the peripheral image of the left vertical side,
Applying the right vertical side image smoothing filter to the application pixel of the peripheral image of the right vertical side,
Applying the upper left corner image smoothing filter to the application pixels of the peripheral image of the upper left corner portion;
Applying the upper right corner image smoothing filter to the application pixels of the peripheral image of the upper right corner portion;
Applying the lower left corner image smoothing filter to the application pixels of the peripheral image of the lower left corner,
A shading correction characterized by calculating a background luminance value of the pixel of interest in each peripheral image by applying the smooth filter for the lower right corner image to the application pixel of the peripheral image of the lower right corner Method. The shading correction method according to claim 4,
The smoothing filter for the upper side image emphasizes the pixel of interest and the pixels arranged on the left and right of the pixel of interest, performs a weak adjustment process on a pixel farther downward from the pixels arranged on the left and right, and It is configured to have a filter array that excludes pixels above the pixels,
The lower-side image smoothing filter emphasizes the pixel of interest and the pixels arranged on the left and right of the pixel of interest, and performs a weak adjustment process on a pixel farther upward from the pixels arranged on the left and right, and the pixels arranged on the left and right Comprising a filter array for excluding lower pixels,
The left vertical side image smoothing filter emphasizes the pixel of interest and the pixels arranged above and below the pixel of interest, performs a weak adjustment process on the pixels farther to the right side from the pixels arranged above and below, and arranges the pixels vertically It is configured to have a filter array that excludes pixels on the left side of each pixel,
The right vertical side image smoothing filter performs enhancement processing on the pixel of interest and each pixel arranged above and below the pixel of interest, and performs weak adjustment processing on a pixel farther to the left from the pixels arranged above and below, and arranges the pixels vertically It is configured to have a filter array that excludes pixels on the right side of each pixel,
The upper left corner image smoothing filter performs enhancement processing on the target pixel, and performs weak tone processing on pixels lower than the target pixel and in the radial direction on the right side. A filter array configured to exclude pixels;
The upper right corner image smoothing filter performs enhancement processing on the target pixel, and performs weak tone processing on a pixel that is lower than the target pixel in the radial direction on the left side, and the upper pixel and the right side of the target pixel. A filter array that excludes the pixels of
The lower left corner image smoothing filter emphasizes the pixel of interest, and performs a weak tone process on a pixel that is located on the upper side and on the right in the radial direction from the pixel of interest. A filter array configured to exclude pixels;
The lower right corner image smoothing filter performs enhancement processing on the target pixel, and performs weak tone processing on a pixel that is located on the upper side and on the left in the radial direction from the target pixel. The lower pixel and the right side below the target pixel A shading correction method characterized by comprising a filter array for excluding the pixels. In the shading correction method according to any one of claims 1 to 5,
In the area information determination step,
Pre-prepared upper side determination pattern filter, lower side determination pattern filter, left vertical side determination pattern filter, right vertical side determination pattern filter, upper left corner determination pattern filter, upper right corner determination pattern Eight types of pattern filters including a filter, a lower left corner determination pattern filter, and a lower right corner determination pattern filter are used, and the area information is determined by applying each pattern filter to the partial image. Shading correction method. A defect detection method for detecting a display defect of the display device based on a captured image acquired by capturing a display image of a display device to be inspected,
A captured image acquisition step of capturing the captured image by capturing the display device;
A background image generation step of generating a background image from the captured image;
A shading correction step of generating a difference image in which shading is removed from the captured image by taking a difference between the captured image and the background image;
A defect detection step of detecting a display defect of the display device based on the difference image,
The background image generation step includes
In the captured image, a partial image of a predetermined size including an arbitrary pixel of interest is an intermediate image composed only of an image of an inspection target portion of the display device, or an inspection non-target portion outside the inspection target portion An area information determination step for determining area information as to whether the peripheral image is configured to include the image of
An applied pixel determination step of determining whether or not the luminance value for each pixel in the partial image is within a predetermined luminance range and recording the applied pixel determined to be within the predetermined luminance range;
A smoothing filter based on the area information is obtained from at least two types of smoothing filters prepared in advance, including an intermediate image smoothing filter corresponding to the intermediate image and a peripheral image smoothing filter corresponding to the peripheral image. And a background luminance value calculating step of calculating a background luminance value of the pixel of interest by applying the selected smoothing filter to the application pixel of the partial image,
A defect detection method, wherein a background image is generated by executing the area information determination step, the applied pixel determination step, and the background luminance value calculation step with each of a plurality of pixels of the inspection target portion as a target pixel. A defect detection apparatus for detecting a display defect of the display device based on a captured image acquired by capturing a display image of a display device to be inspected,
A light source device for irradiating the display device with a light flux;
An imaging device for imaging a light flux via the display device;
A control device for driving and controlling the display device and the imaging device;
The controller is
An image display control unit for driving the display device to display an inspection image;
An image data acquisition unit that causes the imaging device to capture the inspection image displayed on the display device to acquire captured image data;
A background image data generation unit that generates background image data from the captured image data;
A shading correction unit that generates difference image data obtained by removing shading from the captured image data by taking a difference between the captured image data and the background image data;
A defect detection unit for detecting display defects of the display device based on the difference image data;
A data storage unit for storing various data,
In the background image data generation unit,
In the captured image data, a partial image of a predetermined size including an arbitrary pixel of interest is an intermediate image composed only of an image of the inspection target portion of the display device, or an inspection non-target outside the inspection target portion An area information determination unit that determines area information as to whether or not the image is a peripheral image that includes an image of a part;
An application pixel determination unit that determines whether a luminance value for each pixel in the partial image data is within a predetermined luminance range and records an application pixel determined to be within the predetermined luminance range; and
A background luminance value calculating unit that calculates the background luminance value of the pixel of interest by applying the smoothed filter data read from the data storage unit based on the area information to the application pixel of the partial image data;
In the data storage unit,
Storing at least two types of smoothing filter data including smoothing data for intermediate portion image corresponding to the intermediate portion image data and smoothing filter data for peripheral portion image corresponding to the peripheral portion image data. Feature defect detection device. A defect detection apparatus control method program for driving and controlling the defect detection apparatus according to claim 8,
In the control device constituting the defect detection device,
In the captured image data, the partial image having a predetermined size including any pixel of interest is an intermediate image composed only of the image of the inspection target portion of the display device, or the inspection non-target portion outside the inspection target portion Determination of area information as to whether the image is a peripheral image including the image of
Determination of whether or not the luminance value for each pixel in the partial image data is within a predetermined luminance range, and recording of the applied pixel determined to be within the predetermined luminance range;
Selection of a smoothing filter based on the area information from at least two types of smoothing filters including an intermediate image smoothing filter corresponding to the intermediate image data and a peripheral image smoothing filter corresponding to the peripheral image data When,
Calculating a background luminance value of the pixel of interest by applying the selected smoothing filter to the application pixel of the partial image data;
Generation of the background image by calculating a background luminance value with each of the plurality of pixels of the inspection target portion as a target pixel;
A control method program for a defect detection apparatus, characterized in that