Defect inspection method, defect inspection apparatus, and method for manufacturing transparent plate-shaped body

Abstract

To provide a defect inspection device capable of reducing an operation amount required for defect inspection, a defect inspection method, and a manufacturing method of a plate-like body having transparency.SOLUTION: A search unit configured to extract a first defect candidate for a plate-like object from entire first image data obtained by imaging the plate-like object having transparency by an imaging device, calculate a first feature amount of the first defect candidate, and compare the first feature amount with a predetermined threshold value to determine whether or not the defect candidate is a predetermined defect type; SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a defect inspection method and apparatus for detecting defects present in a plate-like body such as a transparent glass substrate, and a method for manufacturing a transparent plate-like body. [Background technology]

[0002] Today, glass substrates (hereinafter referred to as glass plates), which are transparent plate-like bodies, are used in electronic devices such as flat panel displays. It is highly required that the glass plates used in these devices be thin and have very few or no defects such as bubbles. Therefore, during the inspection process, measures such as extracting glass plate parts containing bubbles are necessary. Therefore, various devices have been proposed to inspect manufactured glass plates for defects such as bubbles.

[0003] FIG. 14 is a diagram illustrating an example of a conventional inspection device (hereinafter referred to as the inspection device) for defects such as bubbles present on glass plates. As shown in FIG. 14, the conventional inspection device is provided with a light source 901 on one side of a glass plate 911 being transported. The light source 901 projects linear light, longer than the width of the glass plate 911, toward the glass plate 911 at a predetermined light intensity. The conventional inspection device captures a bright-field image of the transmitted light that has passed through the glass plate 911 using a camera 902 provided on the other side of the glass plate 911 and outputs the captured image to a processing unit 903. Then, in the conventional inspection device, the processing unit 903 extracts dark areas in the bright-field image as defect areas. Note that the conventional inspection device converts the light projected by the light source 901 into pseudo-parallel light using a thin slit extending in the width direction of the glass plate 911. In the conventional inspection device, the camera 902 captures the pseudo-parallel light and the processing unit 903 processes the captured image to detect defects on the glass plate 911 being transported.

[0004] Furthermore, Patent Document 1 proposes providing an illuminator that illuminates the surface of a plate-like body at an angle close to perpendicular to the surface and an illuminator that illuminates at an angle close to parallel to the surface, and then performing image processing on the images obtained using these illuminators to detect defects present on the surface and inside of the plate-like body.

[0005] FIG. 15 is a diagram illustrating the manufacturing process of a glass plate by the float method. In the float process, which is a method for manufacturing flat glass, as shown in Figure 15, molten glass is sent out from a furnace 921 and floated over a float bath 922 containing molten metal with a higher specific gravity than the glass. The molten glass is sent from the float bath 922 to an annealing furnace 923, where it is cooled. The cooled glass solidifies into a sheet shape, becoming a glass ribbon 924, which is then cut to produce glass sheets. General inspections of glass sheets for defects, etc. are performed after cleaning. Inspecting for defects, etc. before cleaning has the advantage of reducing the number of sheets judged to be defective in subsequent processes, but it is affected by inspection disturbances (dust, dirt, etc.) before cleaning. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2002-214158 Summary of the Invention [Problem to be solved by the invention]

[0007] However, with conventional inspection devices and the conventional technology described in Patent Document 1, when defect inspection is performed before the cleaning process, a large amount of dust and transport dirt adheres to the glass, resulting in the detection of a large number of defect candidates, and the inspection processing speed cannot keep up, making it difficult to perform inspection in accordance with the transport speed of the glass ribbon before the cutting process.

[0008] The present invention has been made in consideration of the above problems, and provides a defect inspection device, a defect inspection method, and a method for manufacturing a transparent plate-like body that can reduce the amount of calculation required for defect inspection. [Means for solving the problem]

[0009] The present invention includes the following [1] to [8]. [1] A defect inspection method in a defect inspection device, comprising: a first defect candidate extraction step of extracting a first defect candidate for a transparent plate-like body from first image data of the plate-like body captured by an imaging device; a first feature calculation step of calculating a first feature of the first defect candidate; and a defect type determination step of determining whether the first defect candidate is a predetermined defect type based on the first feature. [2] The defect inspection method of [1] further includes a step of extracting second image data of a predetermined size including pixels of the first defect candidate from the first image data when the defect type determination step determines that the first defect candidate does not belong to the predetermined defect type, a second defect candidate extraction step of extracting the second defect candidate from the second image data, a second feature value calculation step of calculating a second feature value of the second defect candidate, and a step of determining the defect type of the second defect candidate based on the second feature value. [3] A defect inspection method according to [1] or [2], wherein the first feature is a feature that can be used to determine at least one of dust and dirt adhering to the surface of the plate-like body. [4] A defect inspection method according to any one of [1] to [3], wherein the first defect candidate extraction step extracts the first defect candidate by searching the first image data pixel by pixel. [5] A defect inspection method according to any one of [1] to [4], wherein the plate-like body is a glass ribbon formed by continuously flowing molten glass onto the surface of a molten metal bath, a float glass substrate formed by advancing the glass ribbon along the surface of a molten metal bath, a glass substrate for flat panel displays, a glass plate for building materials, or a glass plate for automobiles. [6] A defect inspection device comprising a search unit that extracts a first defect candidate for a transparent plate-like body from the entire first image data obtained by an imaging device capturing the plate-like body, calculates a first feature value of the first defect candidate, and compares the first feature value with a predetermined threshold value to determine whether the first defect candidate is a predetermined defect type. [7] A defect inspection device according to [6], further comprising an identification unit that calculates a second feature amount and determines the type of defect, wherein when the search unit determines that the first defect candidate does not belong to the predetermined defect type, the search unit extracts second image data of a predetermined size that includes pixels of the first defect candidate from the first image data, and generates the second defect candidate from the second image data, and the identification unit calculates the second feature amount of the second defect candidate and determines the type of defect of the second defect candidate based on the second feature amount. [8] A method for manufacturing a transparent plate-like body, comprising: a first defect candidate extraction process for extracting a first defect candidate for the plate-like body from first image data of the transparent plate-like body captured by an imaging device; a first feature calculation process for calculating a first feature of the first defect candidate; and a defect type determination process for determining whether the first defect candidate is a predetermined defect type based on the first feature. [Effects of the Invention]

[0010] According to the present invention, an imaging device extracts a first defect candidate for a plate-like body from first image data obtained by imaging the transparent plate-like body, calculates a first feature amount for the extracted first defect candidate, compares the calculated first feature amount with a predetermined threshold value to determine whether the defect is of a predetermined defect type, and if it is determined that the defect is of the predetermined defect type, determines that the extracted first defect candidate is of the predetermined type. As a result, the image feature amount extracted from the first image data can be used to determine inspection disturbances such as dirt and dust, and non-defects can be excluded from the defect candidates, thereby reducing the amount of calculation required for defect inspection. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram illustrating the schematic configuration of a defect inspection device 10 according to the first embodiment. [Figure 2] FIG. 2 is a diagram illustrating a schematic configuration of the processing device 16 according to the first embodiment. [Figure 3] FIG. 3 is a diagram illustrating the procedure of defect candidate search processing for the entire image according to the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating image data to be searched according to the first embodiment. [Figure 5] FIG. 5 is a diagram illustrating the procedure of the defect image generation process according to the first embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of defect image data according to the first embodiment. [Figure 7] FIG. 7 is a diagram illustrating the procedure of the first defect identification process according to the first embodiment. [Figure 8] FIG. 8 is a diagram illustrating the procedure of the second defect identification process according to the first embodiment. [Figure 9] FIG. 9 is a diagram illustrating the procedure of the quality determination process according to the first embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of image data of a determination area determined to be pass / fail according to the first embodiment. [Figure 11] FIG. 11 is a diagram for explaining the determination result when the defect inspection device 10 according to the first embodiment is used. [Figure 12] FIG. 12 is a diagram illustrating a schematic configuration of a processing device 16a according to the second embodiment. [Figure 13] FIG. 13 is a diagram illustrating an example of cutting a glass plate while avoiding defective product areas according to the second embodiment. [Figure 14] FIG. 14 is a diagram illustrating an example of a conventional inspection device for defects such as bubbles present on a glass plate. [Figure 15] FIG. 15 is a diagram illustrating the manufacturing process of a glass plate by the float method. DETAILED DESCRIPTION OF THE INVENTION

[0012] [First embodiment] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will now be described with reference to the accompanying drawings. Fig. 1 is a diagram illustrating the schematic configuration of a defect inspection device 10 according to this embodiment. As shown in FIG. 1, the defect inspection device 10 comprises a first defect inspection device 12, a second defect inspection device 14, a processing device 16, and a display 32. In FIG. 1, the flow direction of a glass plate G, which is a transparent plate-like body, is the x direction, the width direction of the glass plate G is the z direction, and the direction perpendicular to the x direction and z direction is the y direction. The first defect inspection device 12 and the second defect inspection device 14 are provided in this order from the upstream side along the transport path of the glass sheet G. The glass sheet G is a long plate material that has been removed from a melting furnace and has a predetermined thickness, and is transported in the x-axis direction on a plurality of drive rollers 18 provided on the transport path. Note that in this embodiment, the glass sheet G is a glass ribbon formed by continuously flowing molten glass on the surface of a molten metal bath, or a float glass formed into a sheet by advancing the glass ribbon along the surface of the molten metal bath.

[0013] The first defect inspection device 12 is provided with a first linear light source 20 on the lower surface of the glass plate G, and is provided with a first camera 22 and a light path blocking member 24 on the upper surface of the glass plate G. The first defect inspection device 12 is located at the most upstream side of the conveying side of the defect inspection device 10, and is a device that inspects defects (also called defects) in the glass sheet G. In addition, the first defect inspection device 12 is a device that is suitable for detecting defects such as tiny bubbles present in the glass sheet G.

[0014] The first linear light source 20 is a light source that emits quasi-parallel light, and is, for example, an LED (light-emitting diode) light source. The emission port of the first linear light source 20 extends linearly along the z direction of the glass sheet G. The emission port of the first linear light source 20 is provided at a position, for example, 100 to 900 mm away from the surface of the glass sheet G, and the width of the light source (emission port) along the conveyance direction is set to, for example, 1 to 20 mm. Note that it is preferable that the first linear light source 20 is provided away from the surface of the glass sheet G, since high positional accuracy is not required. Note that the type of light emitted by the LED light source is not particularly limited, and white is preferably used, but red, blue, green, etc. may also be used. The LED light source includes, for example, a light source that emits light, a lens that converts the emitted light into approximately parallel light, a diffusion plate that makes the light intensity approximately uniform, and a slit plate that narrows the light output (not shown). As a result, the first linear light source 20 emits approximately parallel light with approximately uniform light intensity.

[0015] The first camera 22 (imaging device) is a line sensor type camera that is provided at a position facing the first linear light source 20 across the glass plate G, and directly reads the transmitted light that has passed through the glass plate G on its light-receiving surface. A plurality of first cameras 22 are provided in the z direction, and capture images of the same position in the conveyance direction (x direction). Note that FIG. 1 schematically illustrates one first camera 22. The multiple first cameras 22 are set so that the field of view of the glass plate G in the width direction (z direction) partially overlap each other, and are arranged so that there are no non-inspection areas in the inspection portion of the glass plate G. The first camera 22 is installed so that its light receiving surface is located at a position where the imaging lens (not shown) of the first camera 22 is focused, for example, 200 to 400 mm away from the surface of the glass plate G. The first camera 22 is equipped with an optical system including an imaging lens and a diaphragm (not shown) for adjusting the aperture. The image data obtained by the first camera 22 is sent to the processing device 16 sequentially each time a line is read.

[0016] The light path shielding member 24 is a knife-edge-shaped member positioned in front of the first camera 22, blocking part of the optical path of the transmitted light from the glass plate G. The tip portion in the optical path is sharpened to form a blade. The light path shielding member 24 is located in front of the optical system (imaging lens) of the first camera 22, for example, at a distance of 1 to 5 mm. A mechanism that allows the light path shielding member 24 to move in the X direction so that the light path shielding member 24 crosses the optical path is provided in the portion that holds the light path shielding member 24. In this case, when a circle of confusion representing the field of view as seen from the light receiving surface is defined, which is formed on a plane perpendicular to the optical axis of the imaging lens of the first camera 22 at the position where the transmitted light passes through the light path shielding member 24, the area of ​​the portion of the circle of confusion blocked by the light path shielding member 24 preferably corresponds to 30 to 70% of the area of ​​the circle of confusion. If the area is less than 30%, bright areas (described later) are unlikely to appear in the bright-field image, while if the area exceeds 70%, the image is likely to become a dark-field image. Such a range in which the light path is blocked can be achieved, for example, by adjusting the distance between the light path blocking member 24 and the glass plate G and the aperture value of the first camera 22. By setting the range in which the light path is blocked to this range, bright areas can be efficiently formed in the vicinity of dark areas created by defects such as bubbles present in the bright-field image.

[0017] As described above, the image obtained by the first defect inspection device 12 is a bright-field image, and defects in the glass plate G appear as dark areas in the image due to diffuse reflection in the defect area. Furthermore, as described above, a knife-edge-shaped light path blocking member 24 that blocks part of the light path is provided in front of the optical system of the first camera 22. This light path blocking member 24 forms a bright area that is closely opposite or in contact with the dark area in the image. This bright area is generated by refractive error in the defect area, and is brighter than the background area in the bright-field image.

[0018] The second defect inspection device 14 includes a second light source 28 and a second camera 30. The second defect inspection device 14 is a device that illuminates the glass plate G inspected as the inspection target of the first defect inspection device from one side, and receives the illumination light reflected from the front and back surfaces of the glass plate G with the second camera 30, thereby inspecting for defects.

[0019] The second light source 28 is a light source (e.g., an LED light source) that emits substantially parallel light onto the surface of the glass plate G, and the light is incident from a direction oblique to the surface of the glass plate G. As shown in FIG. 1, the second light source 28 is installed at an angle in the y direction. The type of light emitted by the LED light source used for the second light source 28 is not particularly limited; white light is preferably used, but red, blue, green, etc. may also be used. Specifically, the LED light source includes a light-emitting source that emits light, a lens that converts the emitted light into substantially parallel light, a diffusion plate that makes the light intensity substantially uniform, and a slit plate that narrows the light output (not shown). As a result, the second light source 28 emits substantially parallel light with substantially uniform light intensity.

[0020] The second camera 30 (imaging device) is a line sensor type camera that collects reflected light emitted from the surface of the glass plate G and captures a bright field reflected image. The second camera 30 is provided on the same side of the glass plate G as the second light source 28.

[0021] The image captured by the second camera 30 includes an image illuminated by the second light source 28 and reflected on the back surface of the glass plate G (the surface opposite to the surface on which the second light source 28 and the second camera 30 are installed), and the areas of defects present on the glass plate G appear as dark areas. This defect image includes the following two images. The first image is a real image of the defect formed when light is incident on the front surface of the glass plate G from a direction oblique to the surface of the glass plate G, reflected by the back surface of the glass plate G, and passes through the defect area in the optical path of the reflected light. The second image is a mirror image of the defect formed when incident light is incident on the front surface of the glass plate G from a direction oblique to the surface of the glass plate G, passes through the defect area in the optical path within the glass plate G, and is reflected by the back surface of the glass plate G. The image data obtained by the second camera 30 is sent to the processing device 16 each time it is read line by line.

[0022] The processing device 16 extracts defect candidates (first defect candidates) from the captured image data (first image data) as described below. Then, it determines the defect type of the pixels of the extracted first defect candidate, and if it is determined to be a non-defect (pseudo-defect), it excludes the first defect candidate from the defect candidates. The processing device 16 further generates image data of the defect candidate (second image data) including the pixels of the first defect candidate that have not been excluded from the defect candidates. The processing device 16 identifies the defect type in detail based on the generated second image data. The processing device 16 determines whether the glass plate is in a good area or a bad area based on the identification result, and outputs the determination result to the display 32.

[0023] A display 32 is connected to the processing device 16. Images obtained by the first defect inspection device 12 and the second defect inspection device 14, defect detection results, identification results, and / or defect location results are displayed on the display 32.

[0024] Next, an outline of the operation of the processing device 16 according to this embodiment will be described. When there are bubbles or other defects on the glass plate, the image data obtained is white and shiny. On the other hand, when there is dust or other particles on the glass surface, the image data obtained is black. However, even when black image data is obtained, there may be defects such as small bubbles. In particular, in the production of high-quality glass plates, it is necessary to detect even these small bubble defects. Furthermore, when inspection is performed outside the clean room before cleaning, a lot of dust and dirt (pseudo defects) adheres to the glass plate. For this reason, the processing device 16 of this embodiment excludes defect candidates due to dirt or dust from the image data in the early stage of processing, thereby reducing the processing load in the later stage of processing. Furthermore, in the later stage of processing, the processing device 16 of this embodiment identifies whether small black defects that could not be determined in the early stage of processing, defects that appear as glowing white, and dust or dirt that was overlooked in the early stage of processing are defects or not. Furthermore, if the processing device 16 of this embodiment identifies a defect, it identifies the type of defect. Then, based on the identification result, the processing device 16 of this embodiment detects the part identified as a defect as a defective area.

[0025] Next, the processing device 16 will be described with reference to Fig. 2. Fig. 2 is a diagram illustrating the schematic configuration of the processing device 16 according to this embodiment. As shown in FIG. 2, the processing device 16 is connected to an imaging device (first camera 22 and second camera 30) and a display 32. The processing device 16 includes an image acquisition unit 100, a search unit 101, a storage unit 102, a discrimination unit 103, a storage unit 104, a determination unit 105, and a database unit 106.

[0026] The image acquisition unit 100 acquires first image data captured by the imaging device (first camera 22 and second camera 30), and if the acquired image data is an analog signal, converts it into a digital signal and outputs the converted first image data a to the search unit 101. If the image data captured by the imaging device is a digital signal, the acquired image data is output to the search unit 101 as is as the first image data a.

[0027] The search unit 101 scans the entire first image data a output from the image acquisition unit 100, pixel by pixel. The search unit 101 compares the brightness of each scanned pixel with a predetermined first threshold value. The search unit 101 extracts pixels with a brightness higher than the predetermined first threshold value or pixels with a brightness lower than the predetermined first threshold value as first defect candidate pixels, and stores the coordinate values ​​of the extracted first defect candidate pixels in the storage unit 102. The search unit 101 calculates a first feature amount of a portion (first defect candidate) including pixels of the extracted first defect candidate. The search unit 101 compares the calculated first feature amount with a first judgment criterion that is predetermined to classify the type of defect. If the first defect candidate is judged to be a non-defect (a pseudo-defect), the search unit 101 excludes the first defect candidate from the defect candidates. Based on the classification result, the search unit 101 outputs information indicating the defect type, the coordinate values ​​of the pixels of the first defect candidate, and the first feature amount of the first defect candidate to the judgment unit 105 (FIG. 2c). Note that multiple first feature amounts may be calculated. In this case, the search unit 101 compares the calculated multiple first feature amounts with a first judgment criterion that is predetermined for each feature amount to classify the type of defect. The search unit 101 generates second image data including pixels of the first defect candidate that have not been excluded from the defect candidates. Furthermore, the search unit 101 scans the generated second image data pixel by pixel. The search unit 101 compares the brightness of the scanned pixel with a predetermined second threshold value. The search unit 101 extracts pixels with brightness higher than the predetermined second threshold value or pixels with brightness lower than the predetermined second threshold value as second defect candidates, and stores the coordinate values ​​of the extracted pixels of the second defect candidates in the memory unit 102. The search unit 101 outputs the second image data and the scan results to the recognition unit 103 (b in Figure 2).

[0028] The storage unit 102 stores the coordinate values ​​of the first and second defect candidates, the first feature value of the first defect candidate, information indicating the defect type of the first defect candidate, and information indicating whether the coordinates of the first defect candidate are to be excluded from defects, all in association with each other. The storage unit 102 also stores first image data.

[0029] The identification unit 103 calculates a second feature amount using the second image data and the scan result output by the search unit 101. The second feature amount calculated here is the shape (pattern), brightness, area, etc. of the defect. The identification unit 103 identifies and classifies the defect type in detail based on the calculated second feature amount. The identification unit 103 writes the second image data including the second defect candidate classified as a defect into the database unit 106 to update it (e in FIG. 2). Furthermore, the identification unit 103 outputs information indicating the detailed classified defect type related to the second defect candidate based on the identification result to the determination unit 105 (d in FIG. 2).

[0030] The storage unit 104 stores the second image data in association with the second feature amount.

[0031] The determination unit 105 determines whether a determination area of ​​the glass plate is a non-defective area or a defective area, using the information indicating the defect type of the first defect candidate output by the search unit 101, the coordinate values ​​of the pixels of the first and second defect candidates, and the information indicating the detailed classified defect types for the second defect candidate output by the identification unit 103. Based on the determination result, the determination unit 105 writes the information indicating the defect type, the coordinate values ​​of the defect, and the feature amounts of the defect (first and second feature amounts) in association with the second image data of the defect area stored in the database unit 106, and updates the data (f in FIG. 2).

[0032] The database unit 106 stores information indicating the type of defect, coordinate values ​​of the defect, feature amounts of the defect (second feature amounts), and second image data of the defect area in association with each other.

[0033] Next, the defect candidate search process for the entire image performed by the search unit 101 will be described with reference to Fig. 3 and Fig. 4. Fig. 3 is a diagram illustrating the procedure for the defect candidate search process for the entire image according to this embodiment. Fig. 4 is a diagram illustrating the image data to be searched according to this embodiment. In FIG. 4, the direction of glass flow (also referred to as the traveling direction) is the x-direction, and the width direction of the glass is the z-direction. Therefore, the glass flows in the positive x-direction, as indicated by arrow g21. Symbol g1 represents first image data a (if the acquired image data is an analog signal, the first image data a is converted into a digital signal) captured by the imaging device (first camera 22, second camera 30). For example, the x-direction represents A2 pixels, and the z-direction represents A1 pixels. Symbol g2 represents a single pixel. Symbols g11 to g18 represent candidate defect pixels (hereinafter also referred to as "defect pixels"). Symbol g10-1 represents second image data including defect pixels g11 and g12. Symbol g10-2 represents second image data including defect pixel g13. Symbol g10-3 represents second image data including defect pixels g14 to g18.

[0034] (Step S1) The search unit 101 scans all image data g1 for each pixel (g2) using first image data a (if the acquired image data is an analog signal, the first image data a is converted into a digital signal) captured by the imaging device (first camera 22, second camera 30). After step S1 is completed, the process proceeds to step S2. (Step S2) After scanning one pixel, the search unit 101 compares the brightness of the scanned pixel with a predetermined first threshold value. After step S2 is completed, the process proceeds to step S3. (Step S3) If the comparison result shows that the brightness of the scanned pixel is greater than or less than a predetermined threshold value, the search unit 101 designates the pixel as a first defect candidate pixel (note that the former threshold value and the latter threshold value may be different). The search unit 101 stores the coordinate values ​​of the first defect candidate pixel in the storage unit 102, for example, in table format. After step S3 is completed, the process proceeds to step S4.

[0035] (Step S4) The search unit 101 determines a group of adjacent pixels of the first defect candidate as one first defect candidate based on the coordinate values ​​stored in the storage unit 102. An example of the group of adjacent pixels of the first defect candidate is pixels g11 and g12 in FIG. 4. Next, the search unit 101 calculates image features (first features) of the region (first defect candidate) that has been determined as the first defect candidate. The image features include, for example, brightness, area of ​​a part, brightness integral, signal level, brightness information, width length, width length, aspect ratio, density, etc. The luminance integral is the integral value of the luminance of each pixel included in the first defect candidate. After step S4 is completed, the process proceeds to step S5. (Step S5) The search unit 101 requests defect type discrimination condition information (first discrimination conditions) from the judgment unit 105 and acquires the type discrimination condition information. The type discrimination condition information is information for discriminating the type of defect, including the feature values ​​for each defect type and the defect classification processing procedure. The search unit 101 compares the type discrimination condition information with the first feature values ​​of the portion determined as the first defect candidate in step S4 to discriminate the defect type of the portion determined as the first defect candidate. Examples of defect types include air bubbles, foreign matter, dust, and dirt. The type discrimination condition information includes information indicating the size, brightness, size of the circumscribing circle, and density of the defect. The type discrimination condition information also includes information indicating the size, brightness, size of the circumscribing circle, and density of the defect for each of the bright areas, dark areas, and overall image data. Furthermore, the type discrimination condition information is set for each of the imaging devices (first camera 22, second camera 30) shown in FIG. 1. After step S5 is completed, the process proceeds to step S6.

[0036] (Step S6) If the search unit 101 determines that the first defect candidate is a predetermined defect type, it excludes the first defect candidate determined to be a predetermined defect type from the defect candidates. The predetermined defect type is, for example, a non-defect such as dust or dirt adhering to the surface of the glass plate. Dust and dirt adhering to the surface of the glass plate can be removed by cleaning in a subsequent process, so they are excluded from the defect candidates. Next, the search unit 101 adds information indicating exclusion to the coordinate values ​​of pixels included in the first defect candidates determined to be dust or dirt and stored in the storage unit 102. The search unit 101 performs the processes of steps S5 and S6 on all the first defect candidates. Next, the search unit 101 outputs various information c (information indicating the type of defect, coordinate values ​​of the defect, and calculated feature amounts) to the determination unit 105 based on the discrimination result.

[0037] Next, the defect image generation process performed by the search unit 101 will be described with reference to Figs. 5 and 6. Fig. 5 is a diagram illustrating the procedure of the defect image generation process according to this embodiment. Fig. 6 is a diagram illustrating an example of defect image data (second image data) according to this embodiment. In Fig. 6, the direction in which the glass flows is the x direction, and the width direction of the glass is the z direction. In Fig. 6, image data g10-1 is second image data g10-1 including defect pixels g11 and g12 of the image data g1 in Fig. 4, and for example, when the x direction is A 11 Pixel, z direction is A 12 It is a pixel.

[0038] (Step S11) The search unit 101 sequentially refers to the list of coordinate values ​​of the first defect candidate stored in the storage unit 102. In this case, the search unit 101 does not refer to the coordinate values ​​for which information indicating exclusion has been disabled in step S6. After step S11 is completed, the process proceeds to step S12. (Step S12) The search unit 101 searches for the area (A 11 Pixels x A 12 The second image data g10-1 of the first defect candidate pixel is generated from the entire first image data g1. In this case, as shown in FIG. 4, the second image data g10-1 is generated so that the pixel of the first defect candidate is positioned at the center of the second image data to be generated. However, in cases such as image data g10-2 where there is only image data for two pixels in the negative z direction relative to the defect pixel g13, that is, when it is difficult to generate second image data with the defect pixel positioned at the center, the second image data is generated so that the defect pixel is as close to the center as possible. After step S12 is completed, proceed to step S13. (Step S13) The search unit 101 repeats steps S11 and S12 for the list of coordinate values ​​of all first defect candidates stored in the memory unit 102 that do not have information indicating exclusion added, thereby generating second image data. Next, the search unit 101 stores the generated second image data in the storage unit 102. This completes the defect image generation process.

[0039] Next, the first defect identification process for image data of defect candidates performed by the search unit 101 and the identification unit 103 will be described with reference to Fig. 7. Fig. 7 is a diagram illustrating the procedure of the first defect identification process according to this embodiment.

[0040] (Step S21) The search unit 101 sequentially selects one image data from the second image data stored in the storage unit 102 and scans it pixel by pixel. After step S21 is completed, the process proceeds to step S22. The difference between steps S1 and S21 is that in step S1 the entire first image data is scanned, but in step S21 the second image data is scanned. (Step S22) After scanning each pixel, the search unit 101 compares the brightness of the scanned pixel with a predetermined second threshold value. The second threshold value used in step S22 may be the same as or different from the first threshold value used in step S2. After step S22 is completed, the process proceeds to step S23.

[0041] (Step S23) If the comparison result shows that the brightness of the scanned pixel is greater than or less than a predetermined threshold value, the search unit 101 determines that the pixel is a second defect candidate (note that the former threshold value and the latter threshold value may be different). The search unit 101 repeatedly compares the pixels included in the second image data with the second threshold value. Next, the search unit 101 repeats steps S21 to S23 for all image data output by the search unit 101. The search unit 101 outputs second image data b including the second defect candidate to the classification unit 103. After step S23 is completed, the process proceeds to step S24.

[0042] (Step S24) The identification unit 103 calculates various image features (second features) for the second image data b output by the search unit 101. The identification unit 103 calculates, as image features, for example, the shape of the defect (round, linear, etc.), which area of ​​the image data has a high brightness (upward or downward from the center in the y-axis direction), the area of ​​the high brightness range, whether it is continuous, etc., and stores the calculated second features in the storage unit 104 in association with the second image data. This completes the first defect identification process.

[0043] As described above, the search unit 101 extracts a first defect candidate (a portion including pixels of the first defect candidate) from the entire captured first image data and calculates a first feature amount of the extracted first defect candidate. The search unit 101 classifies the type of defect using the calculated first feature amount and a first judgment condition that is predetermined. If the defect candidate is not a defect (a false defect) caused by dust or dirt and has been set as an exclusion target, the search unit 101 excludes the first defect candidate from the defect candidates. Based on the classification result, the search unit 101 outputs information indicating the defect type of the first defect candidate, the coordinate values ​​of the pixels of the defect candidate, and the first feature amount of the defect portion to the judgment unit 105. Note that multiple first feature amounts may be calculated. In this case, the search unit 101 classifies the type of defect using the calculated multiple first feature amounts and a first judgment condition that is predetermined for each feature amount. The search unit 101 generates second image data including pixels of the defect candidates that are first defect candidates that have not been excluded from the defect candidates. Furthermore, the search unit 101 extracts the second defect candidates from the generated second image data, and outputs second image data b including the second defect candidates to the classification unit 103. The classification unit 103 calculates various image feature amounts (second feature amounts) for the second image data b output by the search unit 101.

[0044] Next, the second defect identification process performed by the identification unit 103 on the second image data of the second defect candidate will be described with reference to Fig. 8. Fig. 8 is a diagram illustrating the procedure of the second defect identification process according to this embodiment.

[0045] (Step S31) The classification unit 103 sequentially reads out the second feature amounts of the second image data including the second defect candidates stored in the storage unit 104. After step S31 is completed, the process proceeds to step S32. (Step S32) The identification unit 103 identifies and classifies the defect types of the second defect candidates in detail through a defect classification process described later. After classification, the identification unit 103 outputs information d indicating the defect type of the second defect candidate based on the classification result to the determination unit 105. Next, the identification unit 103 writes second image data e including the classified second defect candidate into the database unit 106 to update the database. The identification unit 103 repeats steps S31 and S32 to classify defects for the second feature amounts of the second image data including all second defect candidates stored in the storage unit 104. Through this classification process, the identification unit 103 classifies defect candidates that could not be classified by the search unit 101 into defects due to dust, foreign matter, etc., and determines whether the classified defect is a non-defect or a defect according to the type of defect, and determines whether it is to be excluded. In addition, the same process is also performed on defect candidates that could be classified by the search unit 101, and a reclassification process is performed. This completes the second defect identification process.

[0046] As described above, the classification unit 103 calculates the second feature amount for the second image data including the second defect candidate generated by the search unit 101. Furthermore, the classification unit 103 determines the defect type of the second defect candidate in detail based on the calculated second feature amount.

[0047] Here, the defect classification process performed by the classification unit 103 will be described.

[0048] The defect classification process in the identification unit 103 determines whether or not the second image data of the second defect candidate contains any areas that shine brighter than the background (areas with higher brightness). If the identification unit 103 determines that there are areas that shine brighter than the background, it determines that the second image data of the second defect candidate contains a defect. The defect type may be, for example, an air bubble or a foreign object. On the other hand, if the identification unit 103 determines that there are no areas that shine brighter than the background, it determines that the second image data of the second defect candidate does not contain a defect, i.e., that the defect candidate is not a defect.

[0049] Next, the pass / fail judgment process performed by the judgment unit 105 will be described with reference to FIGS. Fig. 9 is a diagram illustrating the procedure of the pass / fail determination process according to this embodiment, and Fig. 10 is a diagram illustrating an example of image data of a determination area that has been determined to be pass / fail according to this embodiment. 10, the width direction of the glass sheet G is the z direction, and the direction in which the glass sheet G flows is the x direction. In Fig. 10, symbol g201 represents the determined area, symbol g202 represents a portion determined to be defective, and symbol g203 represents a region determined to be a non-defective area.

[0050] (Step S201) The judgment unit 105 judges the pass / fail of each first defect candidate for the judgment area (usually the entire area of ​​the first image data a) using each piece of information c (information indicating the defect type of the first defect candidate, coordinate values, first feature amount) output by the search unit 101 and information d indicating the defect type of the second defect candidate output by the identification unit 103. In addition, the pass / fail judgment comparison threshold for each first defect candidate can be changed arbitrarily.

[0051] (Step S202) The judgment unit 105 judges areas where the first defect candidate was not detected and areas where the first defect candidate was detected but was judged to be dirt or dust attached to the surface of the glass plate G (areas that do not contain defective defects) as good areas. (Step S203) If the judgment unit 105 judges that the first defect candidate is a bubble or foreign matter occurring inside or on the surface of the glass plate G and that it is of a size that makes it considered to be a defective product, the judgment unit 105 judges the area including the first defect candidate (the area including the defective defect) to be a defective product area. As a result, as shown in FIG. 10, the determining unit 105 determines the area g202 containing the defect as a defective area, and the area g203 as a non-defective area. Next, the determination unit 105 writes defect information (defect type, coordinate values, first and second feature amounts, etc.) for the area determined to be a defective area into the database unit 106 to update it. This completes the pass / fail judgment.

[0052] FIG. 11 is a diagram for explaining the determination results when the defect inspection device 10 according to this embodiment is used. In Figure 11, row r1 represents the processing results of the defect candidate search process (steps S1 to S6), row r2 represents the processing results of the defect image generation process (steps S11 to S13). Row r3 represents the processing results of the first defect identification process (steps S21 to S24), and row r4 represents the processing results of the second defect identification process (steps S31 to S32). Furthermore, column c1 represents the processing results when the defect is dirt, column c2 represents the processing results when the defect is dust, column c3 represents the processing results when the defect is a bubble (air bubble), and column c4 represents the processing results when the defect is a foreign object.

[0053] In this embodiment, the search unit 101 scans (searches) the entire image data and determines the defect type for the defect candidate portion (first defect candidate) extracted. As in case 2 of columns c1 to c4, if the defect type of the first defect candidate can be determined in the defect candidate search process of row r1 and it is determined to be a predetermined defect type, no other determination process needs to be performed on the first defect candidate, thereby reducing the amount of calculation. Furthermore, in the case of Case 1 in columns c1 to c4, when the defect candidate search process in row r1 is unable to determine the type of defect, the defect inspection device 10 of this embodiment is able to determine the type of defect by performing the defect image generation process in row r2, the first defect identification process in row r3, and then the second defect identification process in row r4. In this way, the search unit 101 scans the entire first image data and excludes dirt and dust that are not defects (pseudo-defects) from the pixels that are identified as first defect candidates, thereby significantly reducing the processing load on the identification unit 103 and the judgment unit 105.

[0054] As described above, in the defect inspection device 10 according to this embodiment, the search unit 101, which performs processing in the initial stage, extracts first defect candidates from the first image data. The search unit 101 then calculates feature quantities for the extracted first defect candidates and identifies the defect type based on the calculation results. For a first defect candidate identified at this stage, if it is set as a candidate for exclusion, second image data is not generated, i.e., the second image data is not sent to the identification unit 103. In particular, when inspection is performed before cutting, before cleaning, or outside a clean room, conventional defect inspection devices extract a large number of first defect candidates due to the large amount of dirt adhering to the glass plate. In contrast, in the defect inspection device 10 according to this embodiment, the search unit 101 can extract and exclude these non-defect (pseudo-defect) defect candidates, such as dirt and dust, adhering to the surface of the glass plate during the initial stage of processing. As a result, the processing device 16 according to this embodiment can reduce the amount of data output to the identification unit 103 and further reduce the amount of calculations performed by the identification unit 103. Furthermore, in the defect inspection device 10 according to this embodiment, the search unit 101 extracts second image data including pixels of the first defect candidate that could not be identified in the initial stage processing. Then, the identification unit 103 calculates second feature amounts from the second image data and identifies and classifies the type of defect in detail. That is, the identification unit 103 also classifies the second defect candidate that could not be identified by the search unit 101 and determines whether it is a defect or not. Then, the determination unit 105 determines that a position including pixels that the search unit 101 and the identification unit 103 have identified as a defect is a defective area.

[0055] [Second embodiment] FIG. 12 is a diagram illustrating a schematic configuration of a processing device 16a included in a defect inspection device 10a according to the second embodiment. As shown in FIG. 12, an imaging device (a first camera 22, a second camera 30) and a glass plate cutting control device 200 are connected to the processing device 16a. The processing device 16a includes an image acquisition unit 100, a search unit 101, a storage unit 102, a recognition unit 103, a storage unit 104, a determination unit 105a, a database unit 106, and a size determination unit 107. The differences from the processing device 16 in the first embodiment are the determination unit 105a and the size determination unit 107.

[0056] The determining unit 105a determines whether the determination area is a defective area or a non-defective area, as in the first embodiment, and outputs the coordinate values ​​of the defective area to the size determining unit 107 based on the determination result.

[0057] Based on the coordinate values ​​of the defective area output by the determining unit 105a, the size determining unit 107 outputs an instruction to the glass plate cutting control device 200 to cut the glass plate G while avoiding the defective area, as shown in FIG.

[0058] FIG. 13 is a diagram illustrating an example of cutting a glass plate while avoiding defective product areas according to this embodiment. In Fig. 13, the flow direction of the glass sheet G is the x direction, and the width direction of the glass sheet G is the z direction. In Fig. 13, symbols g301 (g301-1, g301-31, g301-32, and g301-51) represent defective areas including defects, and symbol g311 represents non-defective areas (g311-12, g311-21, g311-22, g311-41, g311-42, and g311-51) and defective areas (g311-11, g311-3, and g311-52). 13, symbols L1 to L5 represent cutting positions in the x direction of the glass plate G. The widths in the x direction at positions L1, L2, L4, and L5 are equal, and the width in the x direction at position L3 is shorter than the width in the x direction at position L1, for example.

[0059] 13, at position L1, the size determination unit 107 determines that the glass plate g311-11 includes the defective area g301-1. Next, at position L1, the size determination unit 107 determines that the glass plate g311-12 can be cut out while avoiding this defective area g301-1, and generates an instruction to cut out the glass plate g311-12. At position L2, the size determining unit 107 determines that no defective area is included, and generates an instruction to cut out the glass plates g311-21 and g311-22. At position L3, the size determination unit 107 determines that the glass plate cannot be cut out while avoiding the defective areas g301-31 and g301-32, and therefore generates an instruction not to cut out the area including the defective areas g301-31 and g301-32. At position L4, the size determining unit 107 determines that no defective area is included, and generates an instruction to cut out the glass plates g311-41 and g311-42. At position L5, the size determination unit 107 determines that the glass plate g311-52 includes the defective area g301-51. Next, at position L5, the size determination unit 107 determines that the glass plate g311-51 can be cut out while avoiding this defective area g301-51, and generates an instruction to cut out the glass plate g311-51.

[0060] As described above, similarly to the first embodiment, the determination unit 105a detects defective areas before cutting. Then, the size determination unit 107 generates instructions for efficiently cutting out the glass sheet G while avoiding the defective areas based on the determination result of the determination unit 105a. As a result, the glass sheet cutting control device 200 can efficiently cut out the glass sheet while avoiding the defective areas.

[0061] The glass plate cutting control device 200 may cut out a glass plate including a defective area described in FIG. 13, for example, g311-11, into a glass plate smaller than the glass plate g311-12 while avoiding the defective area g301-1.

[0062] The defect inspection devices 10 and 10a of the first and second embodiments can be used in the manufacturing process of glass sheets for FPDs (flat panel displays) such as liquid crystal displays, plasma displays, and organic EL (electroluminescence) displays, as well as glass sheets for building materials and automotive glass. In this case, as described in Fig. 13, defective areas can be detected before cutting, and the detected defective areas can be avoided to efficiently manufacture glass sheets.

[0063] 1 and 2 or the defect inspection device 10a of the embodiment in Fig. 12 may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed to perform processing of each part. Note that the "computer system" here includes hardware such as an OS and peripheral devices. Furthermore, if a WWW system is used, the "computer system" also includes the homepage provision environment (or display environment). Additionally, "computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs (Read Only Memory), and CD-ROMs, as well as USB memory connected via a USB (Universal Serial Bus) interface, and storage devices such as hard disks built into computer systems. Furthermore, "computer-readable recording media" also includes devices that retain programs for a certain period of time, such as volatile memory within a server or client computer system. The program may be one that implements part of the aforementioned functions, or it may be one that can implement the aforementioned functions in combination with a program already stored in the computer system.

[0064] While the present invention has been described in detail and with reference to specific embodiments, it will be apparent to those skilled in the art that various changes and modifications can be made without departing from the spirit and scope of the invention. [Explanation of symbols]

[0065] REFERENCE SIGNS LIST 10 defect inspection device, 12 first defect inspection device, 14 second defect inspection device, 16 processing device, 20 first linear light source, 22 first camera, 24 light path blocking member, 28 second light source, 30 second camera, 32 display, 100 image acquisition unit, 101 search unit, 102, 104 storage unit, 103 identification unit, 105 judgment unit, 106 database unit, 107 size judgment unit

Claims

1. A defect inspection method in a defect inspection device, comprising: a first defect candidate extraction step of extracting a first defect candidate for the plate-like object from first image data obtained by capturing an image of the plate-like object having transparency using an imaging device; a first feature amount calculation step of calculating a first feature amount of the first defect candidate; a defect type determination step of determining whether the first defect candidate is a predetermined defect type based on the first feature amount; A defect inspection method comprising:

2. When the first defect candidate is determined not to be of the predetermined defect type in the defect type determination step, extracting second image data of a predetermined size including pixels of the first defect candidate from the first image data to generate the second image data; a second defect candidate extraction step of extracting second defect candidates from the second image data; a second feature amount calculation step of calculating a second feature amount of the second defect candidate; and determining a defect type of the second defect candidate based on the second feature amount; The defect inspection method according to claim 1, further comprising:

3. The first feature amount is a feature that can be used to determine at least one of dust and dirt adhering to the surface of the plate-like body; The defect inspection method according to claim 1 .

4. 2. The defect inspection method according to claim 1, wherein said first defect candidate extraction step extracts said first defect candidate by searching said first image data for each pixel.

5. 5. The defect inspection method according to claim 1, wherein the plate-like body is any one of a glass ribbon formed by continuously flowing molten glass on a surface of a molten metal bath, a float glass substrate formed by advancing the glass ribbon along the surface of the molten metal bath, a glass substrate for a flat panel display, a glass plate for a building material, and a glass plate for an automobile.

6. A defect inspection device comprising a search unit that extracts a first defect candidate for a transparent plate-like body from the entire first image data obtained by an imaging device capturing an image of the plate-like body, calculates a first characteristic value of the first defect candidate, and compares the first characteristic value with a predetermined threshold value to determine whether the first defect candidate is a predetermined defect type.

7. further comprising a discrimination unit that calculates the second feature amount and determines the type of defect; The search unit If it is determined that the first defect candidate is not of the predetermined defect type, second image data of a predetermined size including pixels of the first defect candidate is extracted from the first image data to generate; extracting second defect candidates from the second image data; The identification unit Calculating the second feature amount of the second defect candidate; determining a defect type of the second defect candidate based on the second feature amount; The defect inspection device according to claim 6.

8. A method for producing a transparent plate-like body, comprising: a first defect candidate extraction step of extracting a first defect candidate for the plate-like object from first image data obtained by capturing an image of the plate-like object having transparency using an imaging device; a first feature amount calculation step of calculating a first feature amount of the first defect candidate; a defect type determination step of determining whether the first defect candidate is a predetermined defect type based on the first feature amount; A method for producing a transparent plate-like body, comprising:

Images