JP2008292309A - Pattern defect inspection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent wrong detection for a pattern defect on a board. <P>SOLUTION: An image input part 11 of an image processing part 7 divides image data obtained by photographing a fluorescent material stripe of a glass substrate 1 with a line sensor camera 5 into a plurality of blocks. A boundary discriminating part 12 performs projection addition processing for the respective blocks to determines whether each divided block is an image including the end of the fluorescent material stripe. A region discriminating part 13 separates each image data into an image belonging to an oblong region inside the boundary of the fluorescent material stripe 21 and an image belonging to the outside of the inside region. A difference image detecting part 14 individually detects a difference image of an adjacent image in a first region between the divided blocks, and detects a difference image of an adjacent image in a second region. A defect detecting part 15 determines whether a pattern defect is found based on the respective difference image detection results. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a pattern defect inspection apparatus for automatically inspecting a coating defect of a phosphor coated on a glass substrate such as a plasma display.

  Conventionally, a pattern defect inspection apparatus for inspecting phosphor coating or printing defects coated or printed on a glass substrate such as a plasma display is formed on a glass substrate by irradiating the glass substrate with ultraviolet rays by an ultraviolet illumination light source, for example. The made phosphor is caused to emit light. This luminescent image is taken by an imaging unit such as a line sensor (one-dimensional sensor).

Since the formed phosphors are red (R), green (G), and blue (B) phosphors, when imaging is performed by the imaging unit, colors corresponding to the R, G, and B phosphors, respectively. A filter is attached to the imaging unit, and an image corresponding to each phosphor is captured.
The image captured by the imaging unit is output to the image processing unit, and the image processing unit detects a phosphor coating defect from the luminance signal level of the image.

  For example, as disclosed in Patent Document 1, there is a method for inspecting phosphor coating defects using a difference image.

In this method, when phosphor stripes of R, G, and B colors are periodically coated on a glass substrate such as a plasma display, image data captured by the imaging unit is sent to the image processing unit to perform image processing. To do. The image processing unit divides the image data into a plurality of blocks such as a block of 4 pixels × 4 pixels, for example, cuts out two adjacent blocks, and outputs them to the difference image detection unit. The difference image detection unit compares the luminance signal levels of the pixels of adjacent blocks. When a defect such as a pinhole is present on the phosphor stripe, the defect is detected as a difference in luminance signal level in the difference image. The output of the difference image detection unit is detected by the defect detection unit as a defect when the difference image is compared with a preset determination level (threshold value) and exceeds the determination level.
JP 2004-245829 A

  However, in the above-described technique, there is a case where a false report is generated due to a difference result using an image including the boundary of the coating surface, and there is a case where a defect cannot be correctly detected. In this case, it is necessary to perform processing such as ignoring the difference image detection result near the boundary based on the coordinate value of the image, and it is impossible to detect the defect near the boundary.

  Accordingly, an object of the present invention is to provide a pattern defect inspection apparatus that can prevent erroneous detection of a pattern defect on a substrate.

  That is, the pattern defect inspection apparatus according to the present invention includes an imaging unit that images a phosphor stripe pattern formed on a substrate, and each block in a predetermined range in the stripe pattern based on the imaging result. Each of the image processing unit for generating the image, the edge determination unit for determining whether or not the generated image includes the edge of the stripe pattern, and the determination result by the edge determination unit, A dividing means for dividing the block into a block belonging to a first area inside the edge of the stripe pattern on the substrate and a block belonging to a second area outside the first area in the stripe pattern; A first comparing means for comparing an image of a block belonging to the first area with an image of another block in the first area; and a block belonging to the second area. And a second comparison unit that compares the image of the second block with an image of another block in the second region, and detects a defect in the pattern on the substrate based on the comparison result by the first comparison unit, A pattern defect on the substrate is detected based on a comparison result by the second comparison means.

  According to the present invention, erroneous detection of a pattern defect on a substrate can be prevented.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic diagram showing a configuration example of a pattern defect detection apparatus according to an embodiment of the present invention.

  This pattern defect detection apparatus includes a transport mechanism 2 for transporting a glass substrate 1 such as a plasma display, an illumination 3 that is an ultraviolet illumination, an interference filter 4 for R (red), G (green), and B (blue), a line sensor The camera 5 includes a drive control unit 6 that controls a drive system for conveyance, an image processing unit 7 that inputs data from the line sensor camera 5, and a display unit 8 that displays inspection results.

As an inspection operation, the illumination 3 is turned on, and the fluorescent material applied to the glass substrate 1 is caused to emit light by reflecting the ultraviolet rays on the glass substrate 1. This emitted light is captured by the line sensor camera 5 through the interference filter 4. At this time, a transport pulse signal is sent from the transport mechanism 2 to the drive control unit 6, and the drive control unit 6 controls the exposure time of the line sensor camera 5 in synchronization with the transport speed. The fetched data is sent to the image processing unit 7, and the image processing unit 7 detects unevenness and scratches based on this data.

  In order to inspect the unevenness and scratches, the image processing unit 7 cuts out the captured image in units of blocks, takes the difference between the adjacent vertical and horizontal blocks, and determines whether there is a defect based on the difference amount.

  The image processing unit 7 includes an image input unit 11 that inputs a captured image from the line sensor camera 5, a boundary determination unit 12, a region determination unit 13, a difference image detection unit 14, and a defect detection unit 15.

FIG. 2 is a diagram for explaining the principle for detecting a defect in the phosphor stripe based on the difference image.
FIG. 3 is a flowchart showing an example of the processing operation of the conventional pattern defect detection apparatus.
Compared with the configuration of the conventional pattern defect detection apparatus shown in FIG. 1, the image processing unit 7 does not include the boundary determination unit 12 and the region determination unit 13.

  Here, as shown in FIG. 2, it is assumed that phosphor stripes 21 of R, G, and B colors are periodically coated on the glass substrate 1. In the following description, a glass substrate on which all of the phosphor stripes 21 of R, G, and B are applied will be described. However, in an actual production line, the phosphor stripes of each color are sequentially applied as described above. Of course, each time it is done, it will be inspected. In that case, when the defect is detected, the next phosphor coating or printing process is stopped, the defective glass substrate is washed, and a new phosphor is coated or printed again. Excellent in eliminating waste. For this purpose, since it is necessary to inspect in accordance with the tact time of the production line, a pattern defect inspection apparatus having a high inspection speed is required.

  Image data captured by the line sensor camera 5 of the conventional pattern defect detection apparatus is sent to the image processing unit 7 and input to the image input unit 11 (step S1). The image input unit 11 divides the image data into a plurality of blocks (step S2). For example, the image input unit 11 includes a well-known 4 pixel × 4 pixel block (hereinafter referred to as a 4 × 4 block), 8 × 8 block, or 32 × 32 blocks 22 and 23. Cut out and output to the difference image detection unit 14. Blocks 22 and 23 are minimum unit blocks for detecting defects, and their sizes are appropriately set based on inspection speed, processing speed, defect detection accuracy, and the like.

  The difference image detection unit 14 compares the block 22 with the block 23. Specifically, the difference image detection unit 14 performs difference image detection between the block 22 and the block 23, for example, by comparing the luminance signal levels of the pixels of the block 22 and the block 23 (step S3).

  When the defect 24 such as a pinhole is on the phosphor stripe 21, the difference image detection unit 14 detects the defect target area from the difference image between the block 22 and the block 23 as a difference in luminance signal level.

  The detection result of the difference image detection unit 14 is sent to the defect detection unit 15. The defect detection unit 15 compares the brightness and size of the defect target area of the difference image with a predetermined determination level (threshold), and detects a defect when the determination level is exceeded (steps S4 and S5).

  The image processing unit 7 performs mask processing of defects detected based on an image related to a predetermined range on the glass substrate 1, for example, a region including the boundary of the phosphor stripe 21 (step S6). ), A pass / fail determination is made based on the number and size of defects on the glass substrate 1 (step S7).

  Accordingly, the block 22 and the block 23 are sequentially moved, and the entire glass substrate is inspected for defects in the entire stripe-like phosphor by detecting the difference image. The inspection data is stored in a storage device (not shown), and analysis of the inspection data can be used for manufacturing quality control. Further, although the comparison based on the luminance signal level has been described as a comparison method, the present invention is not limited to this, and it is needless to say that a difference image can be detected by comparison using a histogram of image signals.

FIG. 4 is a diagram illustrating an example of adjacent images to be compared by a conventional pattern defect detection apparatus. FIG. 5 is a diagram illustrating an example of a difference image between adjacent images by a conventional pattern defect detection apparatus.
In the example shown in FIG. 4, an image of N blocks including a dotted diagonal portion 26 corresponding to a defect on the phosphor stripe 25 which is a blank portion corresponding to the phosphor stripe 21 shown in FIG. 2, and the image N + 1 blocks of images adjacent to each other along the longitudinal direction of the phosphor stripe 25 and having no portion corresponding to a defect are used as comparison targets by the difference image detection unit 14. As shown in FIG. 5, a difference 27 appears in these difference images.

FIG. 6 is a diagram showing an example of a phosphor stripe image photographed by a conventional pattern defect detection device.
FIG. 7 is a diagram illustrating an example of an adjacent image including a boundary portion of phosphor stripes to be compared by a conventional pattern defect detection apparatus.
FIG. 8 is a diagram illustrating an example of a difference image between adjacent images including a boundary portion of a phosphor stripe by a conventional pattern defect detection device.

  The image of the N block and the image of the N + 1 block shown in FIG. 7 are obtained by extracting the boundary portion of the phosphor stripe 25 that is a portion surrounded by a circle in the wide range image shown in FIG. It is an image adjacent along the longitudinal direction of the stripe 25. As shown in FIG. 8, in these difference images, a difference 28 along the longitudinal direction of the phosphor stripe 25 appears. This is an image in which the image of N blocks includes the boundary portion of the phosphor stripe 25, and N + 1. This is false information that appears because the image of the block is not an image including the boundary portion of the phosphor stripe 25.

FIG. 9 is a diagram illustrating an example of an adjacent image including a stripe defect of a phosphor stripe to be compared by a conventional pattern defect detection apparatus.
FIG. 10 is a diagram illustrating an example of a difference image between adjacent images including a stripe defect of a phosphor stripe by a conventional pattern defect detection device.
The image of the N block and the image of the N + 1 block shown in FIG. 9 include the line 31 corresponding to the streak-like defect along the longitudinal direction of the phosphor stripe 25 and the adjacent images along the longitudinal direction of the phosphor stripe 25. It is. As shown in FIG. 10, no difference appears in these difference images.

FIG. 11 is a diagram showing a first example of an adjacent image in a direction orthogonal to the stripe direction, including a phosphor stripe stripe defect to be compared by a conventional pattern defect detection apparatus.
FIG. 12 is a diagram illustrating a first example of a difference image between adjacent images in a direction perpendicular to the stripe direction, including a stripe defect of a phosphor stripe, by a conventional pattern defect detection apparatus.
The image of the N block shown in FIG. 11 is an image including the line 31 corresponding to the stripe defect along the longitudinal direction of the phosphor stripe 25, and the image of the N + 1 block is orthogonal to the longitudinal direction of the phosphor stripe 25. It is an image that does not include the line 31 that is adjacent to the image of the N block along the measured direction and that corresponds to the stripe defect. As shown in FIG. 12, a streak-like difference 32 along the longitudinal direction of the phosphor stripe 25 appears in these difference images.

FIG. 13 is a diagram showing a second example of an adjacent image in a direction orthogonal to the stripe direction, including a phosphor stripe stripe defect to be compared by a conventional pattern defect detection apparatus.
FIG. 14 is a diagram illustrating a second example of a difference image between adjacent images in a direction orthogonal to the stripe direction, including a phosphor stripe stripe defect, by a conventional pattern defect detection apparatus.

The image of the N block shown in FIG. 13 is an image including the line 31 corresponding to the stripe defect along the longitudinal direction of the phosphor stripe 25, and the image of the N + 1 block is orthogonal to the longitudinal direction of the phosphor stripe 25. It is an image that does not include the line 31 that is adjacent to the image of the N block along the measured direction and that corresponds to the stripe defect. However, due to uneven movement of the glass substrate 1 by the transport mechanism 2, the position of the phosphor stripe 25 of the N + 1 block image shown in FIG. 13 is the same as the position of the phosphor stripe 25 of the N block image shown in FIG. Is slightly shifted in the direction perpendicular to the longitudinal direction of the phosphor stripe 25.
As a result, as shown in FIG. 14, in addition to the streak-like difference 32, two streak-like false information 33 along the longitudinal direction of the phosphor stripe 25 appear in these difference images.

  Next, the processing operation of the defect detection apparatus according to the embodiment of the present invention will be described. 15 and 16 are flowcharts showing an example of the processing operation of the defect detection apparatus according to the embodiment of the present invention.

  Image data captured by the line sensor camera 5 of the pattern defect detection apparatus according to the embodiment of the present invention is sent to the image processing unit 7 and input to the image input unit 11 (step S11). The image input unit 11 divides the image data into a plurality of blocks such as 4 × 4 (step S12).

  Then, the boundary determination unit 12 of the image processing unit 7 determines, for each of the divided blocks, a boundary determination process for determining whether the block is an image including the boundary of the phosphor stripe 25, that is, an end. Is performed (step S13).

  Details of the boundary determination processing will be described. FIGS. 17, 18 and 19 are flowcharts showing an example of boundary determination processing by the boundary determination unit of the defect detection apparatus according to the embodiment of the present invention.

FIG. 20 is a diagram illustrating an example of a boundary determination target region in the glass substrate 1 used in the defect detection apparatus according to the embodiment of the present invention.
FIG. 21 is a diagram showing an example of a pattern of adjacent blocks in the vicinity of the boundary of the short side of the stripe in the glass substrate 1 used in the defect detection apparatus according to the embodiment of the present invention. In FIG. 21, N blocks and N + 1 blocks connected to the lower end of the N blocks are adjacent blocks.

  The region 41 shown in FIG. 20, that is, the end of the stripe in the adjacent block near the boundary of the short side of the stripe in the glass substrate 1 is near the upper end of the N block shown in FIG. ) Near the center of the N block shown in FIG. 21, near the lower end of the N block shown in FIG. 21C, or the boundary between the N block and the N + 1 block shown in FIG. 21D.

  FIG. 22 is a diagram illustrating an example of a pattern of adjacent blocks in the vicinity of the boundary of the long side of the stripe in the glass substrate 1 used in the defect detection apparatus according to the embodiment of the present invention. The N + 1 block shown in FIG. 22 is a block connected to the right end of the N block.

  The region 42 shown in FIG. 20, that is, the stripe in the adjacent block near the boundary of the long side of the stripe in the glass substrate 1, is near the center of the N block shown in FIG. 22 (a) or shown in FIG. 22 (b). It is near the left end of the N block, the boundary between the N block and N + 1 block shown in FIG. 22C, or the left end of the N + 1 block shown in FIG.

  FIG. 23 is a diagram illustrating the definition of the first threshold value for boundary determination by the boundary determination unit in the defect detection device according to the embodiment of the present invention. FIG. 24 is a diagram illustrating the definition of the second threshold value for boundary determination by the boundary determination unit in the defect detection device according to the embodiment of the present invention.

  The first threshold value is smaller than the second threshold value, for example, approximately one fifth of the length of one side of the block. The second threshold value is larger than the first threshold value, for example, approximately 4/5 of the length of one side of the block.

  As shown in FIG. 23, the first threshold value for boundary determination by the boundary determination unit 12 depends on the width of the stripe on the block. Also, as shown in FIG. 24, the second threshold value for boundary determination by the boundary determination unit 12 depends on the size of the entire block.

  FIG. 25 is a diagram showing the vertical projection addition results of the first to fourth blocks by the boundary determination unit in the defect detection device according to the embodiment of the present invention. Here, the vertical direction is a direction from the upper end to the lower end of the block.

Here, the first block is the block shown in FIG. 25 (a), the second block is the block shown in FIG. 25 (b), and the third block is shown in FIG. 25 (c). The fourth block is the block shown in FIG. 25 (d).
The longitudinal direction of the band corresponding to the stripe 25 on these blocks is the vertical direction of the block.

  Of these, the length in the short direction of the band on the first block is about one-fourth of one side of the block, and this band extends linearly from the vicinity of the center of the lower end of the block toward the upper end. It is cut near the upper end. The band on the second block extends linearly from the same position as the lower end of the first block to the upper end of the second block. The length of the band in the short direction is the same as the length of the band of the first block in the short direction.

  In addition, there are two bands on the third block, the first band extends linearly from the center and the left end of the lower end of the block to the upper end, and the second band extends from the right end of the lower end of the block. It extends linearly to the upper end of the block. The length in the short direction of the first band is the same as the length in the short direction of the band of the first block, and the length in the short direction of the second band is the short direction of the first band. Is half the length of

The band pattern on the fourth block is obtained by removing the first band from the pattern on the third block.
As shown in FIG. 25, the length of the band of the projection addition result in the vertical direction of the first to fourth blocks exceeds both the first and second thresholds.

  First, the boundary determination unit 12 first selects one of the blocks as the boundary determination process that is the process of step S13, and performs a projection addition process in the vertical direction of the pixels of the block (step S31).

The boundary determination unit 12 determines whether or not the length of the band appearing in the vertical projection addition result exceeds both the first and second thresholds (step S32).
If the selected block is the first to fourth blocks shown in FIG. 25, the boundary determination unit 12 determines “YES” in the process of step S32.
In this case, the boundary determination unit 12 performs a projection addition process in the horizontal direction orthogonal to the above-described vertical direction for the block (step S33).

  The boundary discriminating unit 12 is such that the length of the band appearing in the horizontal projection addition result exceeds the first threshold and is equal to or less than the second threshold, and the band appearing in the projection addition result is cut off in the middle of the vertical direction. It is determined whether or not there is a change point (step S34).

  FIG. 26 is a diagram illustrating a horizontal projection addition result of the first block by the boundary determination unit in the defect detection device according to the embodiment of the present invention. FIG. 27 is a diagram illustrating a horizontal projection addition result of the second block by the boundary determination unit in the defect detection device according to the embodiment of the present invention. FIG. 28 is a diagram showing a horizontal projection addition result of the fourth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.

  When the block selected by the boundary discriminating unit 12 is the first block shown in FIG. 25A, the length of the band of the horizontal projection addition result is the first threshold value as shown in FIG. There is a transition point while exceeding and below the second threshold.

  Further, when the block selected by the boundary determination unit 12 is the fourth block shown in FIG. 25D, the length of the band of the horizontal projection addition result is the first as shown in FIG. Or less.

  In such a case, the boundary determination unit 12 determines “NO” in the process of step S34, and regards the block subjected to the projection addition process as a boundary block including the boundary of the phosphor stripe 25 (step S35). Then, the boundary determination process in step S13 is finished.

  When the block selected by the boundary determination unit 12 is the second block shown in FIG. 25B, the length of the band of the horizontal projection addition result is the first length as shown in FIG. The threshold is exceeded and there is no change point.

  Although not shown, when the block selected by the boundary determination unit 12 is the third block shown in FIG. 25C, the band of the horizontal projection addition result is the same as the second block. The length exceeds the first threshold and there is no change point.

  In this case, the boundary determination unit 12 determines “YES” in the process of step S34, and selects the adjacent block on the assumption that the block adjacent to the upper or lower end of the block may be a boundary block (step S36).

  The boundary determination unit 12 performs the vertical projection addition process on the adjacent block, the projection addition result value exceeds the first threshold value and the second threshold value, and performs the horizontal projection addition process. If there is no change point at step S37 (YES in step S37), it is determined that the band corresponding to the stripe 25 on the adjacent block extends without cutting from the upper end to the lower end of the block. It is assumed that none of the blocks is a boundary block (step S38), and the boundary determination process in step S13 is completed.

  If the boundary determination unit 12 determines “NO” in the process of step S37, there is no band corresponding to the stripe 25 on the adjacent block, or the band is between the upper end and the lower end of the block. It is determined that the block is cut, and the adjacent block is regarded as a boundary block (step S37 → S35).

FIG. 29 is a diagram illustrating vertical projection addition results of the fifth to seventh blocks by the boundary determination unit in the defect detection device according to the embodiment of the present invention.
The fifth block is the block shown in FIG. 29 (a), the sixth block is the block shown in FIG. 29 (b), and the seventh block is the block shown in FIG. 29 (c). .

  The longitudinal direction of the band corresponding to the stripe 25 on the block of the fifth block is the vertical direction of the block, and the longitudinal direction of the band corresponding to the stripe 25 on the block of the sixth and seventh blocks is the vertical direction of the block. Left and right direction.

  Among these, the band pattern on the fifth block is a pattern excluding a part from the upper end of the band on the first block shown in FIG. Specifically, the band on the fifth block extends linearly from the same position as the lower end of the band of the first block toward the upper end of the fifth block, and is closer to the upper end from the center of the lower end and the upper end. Cut at position. The length of the band in the short direction is the same as the length of the band of the first block in the short direction.

  The lower end of the strip on the sixth block is in contact with the lower end of the block and extends linearly from the left end to the right end of the block. The length of the band in the short direction is the same as the length of the band of the first block in the short direction.

  In addition, there are two bands on the seventh block, and the first band is linear from the left end to the right end of the block with a width between the position near the center of the upper end and the lower end of the block and the position closer to the upper end. It extends. The lower end of the second band is in contact with the lower end of the block and extends linearly from the left end to the right end of the block. The length of the first band on the seventh block in the short direction is the same as the length of the first block in the short direction, and the length of the second band in the short direction is the first length. It is half the length of the band in the short direction.

FIG. 30 is a diagram illustrating vertical projection addition results of the eighth and ninth blocks by the boundary determination unit in the defect detection device according to the embodiment of the present invention.
The eighth block is the block shown in FIG. 30A, and the ninth block is the block shown in FIG.

  Of these, the lower end of the band on the eighth block is in contact with the lower end of the eighth block, similar to the band of the sixth block shown in FIG. 29B, and this band is the band of the sixth block. Extends in a straight line from the left end to the right end of the eighth block and is cut near the right end.

Further, the lower end of the band on the ninth block is in contact with the lower end of the block like the eighth block, and this band has the same width as the band of the sixth block shown in FIG. The ninth block extends slightly in a straight line from the left end to the right end, and is cut near the left end.
As shown in FIG. 29 and FIG. 30, the length of the band of the vertical projection addition result of the fifth to ninth blocks exceeds the first threshold and is lower than the second threshold.

  If the boundary determination unit 12 determines “NO” in the process of step S32 described above, whether the length of the band appearing in the vertical projection addition result regarding the selected block exceeds the first threshold value or not. Is discriminated (step S41).

If the selected block is the fifth to ninth blocks shown in FIGS. 29 and 30, the boundary determination unit 12 determines “YES” in the process of step S41. In this case, the boundary determination unit 12 performs a horizontal projection addition process for the block (step S42).
Then, the boundary determination unit 12 determines whether or not the length of the band appearing in the horizontal projection addition result exceeds both the first threshold value and the second threshold value (step S43).

  FIG. 31 is a diagram showing a horizontal projection addition result of the fifth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention. FIG. 32 is a diagram showing a horizontal projection addition result of the sixth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.

  FIG. 33 is a diagram illustrating a horizontal projection addition result of the eighth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention. FIG. 34 is a diagram showing a horizontal projection addition result of the ninth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.

  When the block selected by the boundary discriminating unit 12 is the fifth block shown in FIG. 29A, the length of the band of the horizontal projection addition result is the first threshold value as shown in FIG. And below the second threshold.

  In addition, when the block selected by the boundary determination unit 12 is the sixth block shown in FIG. 29B, the length of the band of the horizontal projection addition result is the first as shown in FIG. Both exceed the second threshold.

  Although not shown, when the block selected by the boundary determination unit 12 is the seventh block shown in FIG. 29C, the band of the horizontal projection addition result is the same as the sixth block. The length exceeds both the first and second thresholds.

  Further, when the block selected by the boundary determination unit 12 is the eighth block shown in FIG. 30A, the length of the band of the horizontal projection addition result is as shown in FIG. Both exceed the second threshold.

  When the block selected by the boundary determination unit 12 is the ninth block shown in FIG. 30B, the length of the band of the horizontal projection addition result is as shown in FIG. Or less.

  If the selected block is the fifth block shown in FIG. 29 (a) or the ninth block shown in FIG. 30 (b), the boundary determination unit 12 sets “NO” in the process of step S43. Determine. As described above, when the band of the vertical projection addition result exceeds the first threshold and falls below the second threshold, a change point exists in the horizontal projection addition result. Therefore, when the boundary determination unit 12 determines “NO” in the process of step S43, the boundary determination unit 12 determines that the selected block is a boundary block (step S44), and ends the boundary determination process of step S13.

  On the other hand, if the selected block is the sixth or seventh block shown in FIGS. 29B and 29C and the eighth block shown in FIG. "YES" is determined in the process of S43.

  In this case, the boundary determination unit 12 performs the vertical projection addition process for the block again (step S45). If there is a change point in the result (YES in step S46), the block is a boundary block. If there is no change point (NO in step S46), it is determined that there is a possibility that the block adjacent to the upper end or the lower end of the block is a boundary block. Transition to processing.

FIG. 35 is a diagram showing vertical projection addition results of the tenth to twelfth blocks by the boundary determination unit in the defect detection device according to the embodiment of the present invention.
The tenth block is the block shown in FIG. 35 (a), the eleventh block is the block shown in FIG. 35 (b), and the twelfth block is the block shown in FIG. 35 (c). .

  Among these, the band pattern on the tenth block is a pattern excluding a part from the upper end of the band on the fifth block shown in FIG. Specifically, the band on the tenth block is linear from the same position as the lower end of the fifth block to the upper end from the center of the tenth block with the same width as the band of the fifth block. It extends slightly and cuts near the lower end.

  Further, the lower end of the band on the eleventh block is in contact with the lower end of the block, and this band has the same width as the second band below the seventh block shown in FIG. 29 (c). The block extends from the left end to the right end.

Further, the lower end of the band on the twelfth block is in contact with the lower end of the block, and this band has the same width as the second band below the seventh block shown in FIG. This block extends from the left end of the block toward the right end and is cut near the right end.
As shown in FIG. 35, the length of the band of the vertical projection addition result of the tenth to twelfth blocks is equal to or less than the first threshold.

If the selected block is the tenth to twelfth block shown in FIG. 35, the boundary determination unit 12 determines “NO” in the process of step S41. In this case, the boundary determination unit 12 performs a horizontal projection addition process for the block (step S51).
Then, the boundary determination unit 12 determines whether or not the length of the band appearing in the horizontal projection addition result exceeds both the first threshold value and the second threshold value (step S52).

FIG. 36 is a diagram illustrating a horizontal projection addition result of the tenth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.
FIG. 37 is a diagram illustrating a horizontal projection addition result of the eleventh block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.
FIG. 38 is a diagram illustrating a horizontal projection addition result of the twelfth block by the boundary determination unit in the defect detection device according to the embodiment of the present invention.

  When the block selected by the boundary discriminating unit 12 is the tenth block shown in FIG. 35A, the length of the band of the projection addition result in the horizontal direction is the first threshold value as shown in FIG. , Below the second threshold, and there is a transition point.

  In addition, when the block selected by the boundary determination unit 12 is the eleventh block shown in FIG. 35B, the length of the band of the horizontal projection addition result is as shown in FIG. Both the first and second thresholds are exceeded, and there is a change point.

  When the block selected by the boundary determination unit 12 is the twelfth block shown in FIG. 35 (c), the length of the band of the horizontal projection addition result is the first as shown in FIG. Both the second threshold and the change point exist.

  If the selected block is the eleventh block shown in FIG. 35 (b) or the twelfth block shown in FIG. 35 (c), the boundary determining unit 12 sets “YES” in the process of step S52. Determine. In this case, the boundary determination unit 12 performs the processing after step S45.

  On the other hand, when the selected block is the tenth block shown in FIG. 35A, the boundary determination unit 12 determines “NO” in the process of step S52. In this case, the boundary determination unit 12 determines whether or not the length of the band appearing in the horizontal projection addition result exceeds the first threshold (step S53).

  If the selected block is the tenth block shown in FIG. 35A, the boundary determination unit 12 determines “YES” in the process of step S53. As described above, when the band of the projection addition result in the vertical direction is equal to or smaller than the first threshold value and the band of the projection addition result in the horizontal direction exists, there is a change point in the projection addition result in the horizontal direction. Therefore, when the boundary determination unit 12 determines “YES” in the process of step S53, the boundary determination unit 12 determines that the selected block is a boundary block (step S54), and ends the boundary determination process of step S13.

  If the boundary determination unit 12 determines “NO” in the process of step S53, it determines that the selected block is not a boundary block (step S55), and ends the boundary determination process of step S13.

  The boundary determination unit 12 determines whether or not the determination as to whether or not it is a boundary block for all the images divided into blocks has been completed by the process of step S13 (step S14). (NO in step S14), another image is selected (step S15), and the process returns to step S13.

  On the other hand, when the boundary determination unit 12 determines “YES” in the process of step S14, the region determination unit 13 first detects the center coordinate value of the image determined as the boundary block, and based on this coordinate value. The image data divided into blocks is divided into an image belonging to the first area which is a rectangular area inside the boundary of the phosphor stripe 21 and an image belonging to the second area which is an area outside the first area described above. Therefore, each image data belonging to the first area is associated with the coordinate value information of the center point of the image and the label information indicating the inner area described above, and the coordinates of the center point of the image are associated with the image data belonging to the second area. The value information is associated with the label information indicating the outer area described above, and this information is stored in the internal memory of the image processing unit 7 (step S16).

  Here, the label information is indicated by numbers. In addition, the label information indicating the outer region includes a case where the longitudinal direction of the boundary block group adjacent in the region is the longitudinal direction of the phosphor stripe 21 and a direction orthogonal to the longitudinal direction of the phosphor stripe 21. It is distinguished by.

  However, in this embodiment, label information is not associated with image data that is an overlapping portion of a block group along the longitudinal direction of the phosphor stripe 21 and a block group along a direction orthogonal to the longitudinal direction of the phosphor stripe 21. The difference image is not detected.

  For example, the label information associated with the coordinate value information of the image belonging to the above inner region is “0”. The label information associated with the coordinate value information excluding the above-described overlapping portion of each image belonging to the outer region and having the longitudinal direction of the boundary block group adjacent in the region being the longitudinal direction of the phosphor stripe 21 is as follows. “1”.

  In addition, it is associated with the coordinate value information excluding the above-described overlapping portion in each image belonging to the outer region and in which the longitudinal direction of the adjacent boundary block group is the direction orthogonal to the longitudinal direction of the phosphor stripe 21 The label information to be recorded is “2”.

FIG. 39 is a diagram showing an example of each region divided by the defect detection device according to the embodiment of the present invention.
A region 51 surrounded by a dotted line shown in FIG. 39 is an image region belonging to the inner region. Further, the region 52 indicated by hatching along the longitudinal direction of the glass substrate 1 shown in FIG. 39 belongs to the outer region, and the longitudinal direction of the boundary block group adjacent in the region is the phosphor stripe 21. It is the area | region of the image which is a longitudinal direction. Further, the region 53 indicated by hatching along the direction orthogonal to the longitudinal direction of the glass substrate 1 shown in FIG. 39 belongs to the outer region, and the longitudinal direction of the boundary block group adjacent in the region is the same. This is a region of an image that is a direction orthogonal to the longitudinal direction of the phosphor stripe 21.

  Next, the difference image detection unit 14 selects either the coordinate value information or the image data with label information stored in the internal memory by the region determination unit 13 (step S17), and the label associated with the image data. By referring to the information, it is determined whether or not the image data is a boundary block image (step S18).

  When the label information of the selected image data is “1” or “2”, the difference image detection unit 14 regards the image as a boundary block image (YES in step S18), and selects the selected image data. If the label information is “1”, the difference image between the selected image data and an image that is adjacent to the image data and is associated with the label information “1” in accordance with the coordinate value information. Perform detection.

  Further, when the label information of the selected image data is “2”, the difference image detecting unit 14 follows the coordinate value information and is adjacent to the selected image data and the label information “ Difference image detection is performed with respect to an image associated with “2” (step S19).

  Further, the difference image detection unit 14 is adjacent to the selected image data whose label information is “1” along the direction orthogonal to the longitudinal direction of the phosphor stripe 21 and is directed to the inside of the glass substrate 1. The image data having the selected label information “2” and the image data of the glass substrate 1 adjacent to the image data along the longitudinal direction of the phosphor stripe 21 may be detected. Difference image detection with an image directed inward may be performed.

  On the other hand, if the label information of the selected image data is “0”, the difference image detection unit 14 regards that the image belongs to an inner area that is not an image of the boundary block (NO in step S18), and performs this selection. Difference image detection is performed between the obtained image data and an image that is adjacent to the image data along the longitudinal direction of the phosphor stripe 21 and that is similarly associated with the label information “0” (step S21).

  Then, the difference image detection unit 14 fluoresces the band that appears in the result of the projection addition process in the direction along the longitudinal direction of the phosphor stripe 21 for the image data that has been selected and the label information is “0”, except at both ends. If there is a change point at a location other than the change portion corresponding to the width from the left end portion to the right end portion of the body stripe 21, it is determined that a streak defect is included in the image.

  After the process of step S19 or step S21, the difference image detection unit 14 determines whether or not all of the difference image detection processes have been performed for each block-divided image (step S22), and “NO” is determined in this process. In that case, the process returns to step S17.

If the difference image detection unit 14 determines “YES” in the process of step S <b> 22, a defect target area from the difference image is detected as a difference in luminance signal level, and the result is sent to the defect detection unit 15. The defect detection unit 15 compares the brightness and size of the defect target area of the difference image with a preset determination level (threshold), and detects a defect when the determination level is exceeded (step S23).
The image processing unit 7 performs a pass / fail determination based on the number and size of defects on the glass substrate 1 (step S24).

  As described above, in the pattern defect detection device according to the embodiment of the present invention, each image obtained by dividing the captured image of the phosphor stripe 21 in the shape of the glass substrate 1 is divided into regions inside the boundary portion of the phosphor stripe 21. And an image belonging to the outer area, and it is determined whether or not there is a pattern defect based on a difference image of a plurality of images in each divided area. Therefore, it is possible not to detect the difference image between the boundary portion of the phosphor stripe and the portion that is not, and as a result, erroneous detection of the pattern defect can be prevented.

  In addition, since it is possible to determine whether or not there is a streak defect based on the presence or absence of a change point by performing projection addition processing of the image divided into blocks, it is possible to prevent erroneous detection of a defect due to movement unevenness or luminance unevenness.

  The present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be omitted from all the components shown in the embodiment. Furthermore, you may combine suitably the component covering different embodiment.

The schematic diagram which shows the structural example of the pattern defect detection apparatus according to embodiment of this invention. Explanatory drawing for detecting the defect of the phosphor stripe by the difference image. The flowchart which shows an example of the processing operation of the conventional pattern defect detection apparatus. The figure which shows an example of the adjacent image compared with the conventional pattern defect detection apparatus. The figure which shows an example of the difference image of an adjacent image by the conventional pattern defect detection apparatus. The figure which shows an example of the image of the fluorescent substance stripe image | photographed with the conventional pattern defect detection apparatus. The figure which shows an example of the adjacent image containing the boundary part of the fluorescent substance stripe compared with the conventional pattern defect detection apparatus. The figure which shows an example of the difference image of the adjacent image containing the boundary part of a fluorescent substance stripe by the conventional pattern defect detection apparatus. The figure which shows an example of the adjacent image containing the stripe-shaped defect of the fluorescent substance stripe compared with the conventional pattern defect detection apparatus. The figure which shows an example of the difference image of the adjacent image containing the stripe-shaped defect of a fluorescent substance stripe by the conventional pattern defect detection apparatus. The figure which shows the 1st example of the adjacent image of the direction orthogonal to a stripe direction containing the stripe-shaped defect of the fluorescent substance stripe compared with the conventional pattern defect detection apparatus. The figure which shows the 1st example of the difference image of the adjacent image of the direction orthogonal to a stripe direction containing the stripe-shaped defect of a fluorescent substance stripe by the conventional pattern defect detection apparatus. The figure which shows the 2nd example of the adjacent image of the direction orthogonal to a stripe direction containing the stripe-shaped defect of the fluorescent substance stripe compared with the conventional pattern defect detection apparatus. The figure which shows the 2nd example of the difference image of the adjacent image of the direction orthogonal to a stripe direction including the stripe-shaped defect of a fluorescent substance stripe by the conventional pattern defect detection apparatus. The flowchart which shows an example of the processing operation of the defect detection apparatus according to embodiment of this invention. The flowchart which shows an example of the processing operation of the defect detection apparatus according to embodiment of this invention. The flowchart which shows an example of the boundary discrimination | determination process by the boundary discrimination | determination part of the defect detection apparatus according to embodiment of this invention (the 1). The flowchart which shows an example of the boundary discrimination | determination process by the boundary discrimination | determination part of the defect detection apparatus according to embodiment of this invention (the 2). The flowchart (the 3) which shows an example of the boundary discrimination | determination process by the boundary discrimination | determination part of the defect detection apparatus according to embodiment of this invention. The figure which shows an example of the area | region of the boundary discrimination | determination object in the glass substrate 1 used for the defect detection apparatus according to embodiment of this invention. The figure which shows an example of the pattern of the adjacent block near the boundary of the short side of a stripe in the glass substrate 1 used for the defect detection apparatus according to embodiment of this invention. The figure which shows an example of the pattern of the adjacent block near the boundary of the long side of the stripe in the glass substrate 1 used for the defect detection apparatus according to embodiment of this invention. The figure explaining the definition of the 1st threshold value for the boundary discrimination | determination by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure explaining the definition of the 2nd threshold value for the boundary discrimination | determination by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the vertical direction of the 1st thru | or 4th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 1st block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 2nd block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. It is a figure which shows the projection addition result of the horizontal direction of the 4th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the vertical direction of the 5th thru | or 7th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the vertical direction of the 8th and 9th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 5th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 6th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 8th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 9th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the vertical direction of the 10th thru | or 12th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 10th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 11th block by the boundary discrimination | determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows the projection addition result of the horizontal direction of the 12th block by the boundary determination part in the defect detection apparatus according to embodiment of this invention. The figure which shows an example of each area | region divided by the defect detection apparatus according to embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Glass substrate, 2 ... Conveyance mechanism, 3 ... Illumination, 4 ... Interference filter, 5 ... Line sensor camera, 6 ... Drive control part, 7 ... Image processing part, 8 ... Display part, 11 ... Image input part, 12 ... Boundary determination unit, 13 ... area determination unit, 14 ... difference image detection unit, 15 ... defect detection unit.

Claims (1)

An imaging means for imaging a stripe pattern of the phosphor formed on the substrate;
An image processing unit that generates an image of each block in a predetermined range in the stripe pattern based on an imaging result by the imaging unit;
An edge determination means for determining whether or not the edge of the stripe pattern is included in the image generated by the image processing unit;
Based on the discrimination result by the edge discrimination means, each block is divided into a block belonging to a first area inside the edge of the stripe pattern on the substrate and the first area in the stripe pattern. A dividing means for dividing into blocks belonging to the second region which is outside;
First comparing means for comparing an image of a block belonging to the first area generated by the image processing unit and divided by the dividing means with an image of another block in the first area;
Second comparing means for comparing an image of a block belonging to the second area generated by the image processing unit and divided by the dividing means with an image of another block in the second area;
A first defect detection unit for detecting a defect of a pattern on the substrate based on a comparison result by the first comparison unit;
A pattern defect inspection apparatus comprising: a second defect detection unit configured to detect a defect of a pattern on the substrate based on a comparison result by the second comparison unit.

Images