Glass plate inspection method

Abstract

To provide a glass plate inspection method capable of non-contact inspection to determine whether a glass plate has a deformation exceeding a predetermined size.SOLUTION: A photographed image of a glass plate is generated by photographing the glass plate using the Schlieren method. Then, on the basis of a relationship between a gradation value (a gradation value near a maximum gradation value) of pixels in a white region close to a region 21 corresponding to bubbles, and a gradation value of pixels in a background region 24, it is determined whether the amount of surface deformation generated on the glass plate is equal to or less than a predetermined value.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a glass plate inspection method for inspecting whether or not deformation of a predetermined magnitude or more has occurred on the surface of a glass plate. [Background technology]

[0002] If bubbles are generated or foreign matter gets into the glass plate during the manufacturing process, the surface of the glass plate will bulge due to the presence of the bubbles or foreign matter. FIG. 11 is a schematic diagram showing deformation of a glass plate. If a bubble 82 is present inside a glass plate 81, a deformed portion 83 may be generated on the surface of the glass plate 81 at the location of the bubble 82. The deformed portion 83 is raised from the surface of the glass plate 81. The height H from the flat portion of the surface of the glass plate 81 to the highest point of the deformed portion 83 is referred to as the amount of surface deformation. Note that the deformed portion 83 will also be generated if a foreign matter is present inside the glass plate 81.

[0003] The surface deformation amount H can be directly measured using a laser microscope.

[0004] Patent Document 1 also describes a glass substrate in which at least one surface is not bulged. It also describes a method for determining whether or not such a glass substrate exists. In this inspection method, light is irradiated onto the glass substrate during transportation, and the reflected light is incident on a line sensor camera, which then captures an image of the glass substrate. Based on the resulting image, the distance D from the surface of the glass substrate to the bubble within the glass substrate and the equivalent sphere diameter e of the bubble are calculated. Then, depending on whether D and e satisfy a predetermined relationship, it is determined whether or not the glass substrate exists in a state in which at least one surface is not bulged.

[0005] The Schlieren method is also known as a method for visualizing refractive index variations in an object. The Schlieren method captures an image in which the refractive index variations are represented by light and dark, utilizing the fact that the path of light passing through an object varies depending on the location of the object. FIG. 12 is an explanatory diagram showing an example of the Schlieren method. Note that in FIG. 12, the glass plate to be photographed is not present. In the Schlieren method, as shown in FIG. 12, a light source 93 is provided so that light enters a line sensor camera 91. Light that passes through a focal position 95 of the lens 92 enters a lens 92 of the line sensor camera 91. In the example shown in FIG. 12, a light shielding plate 94 is provided between the focal position 95 and the lens 92. The light shielding plate 94 is positioned so as to block a portion of the light beam from the light source 93 that enters the lens 92. The light shielding plate 94 is sometimes referred to as an edge knife. If the light shielding plate 94 were not present, light within the range of angle θ1 shown in FIG. 12 would be incident on the lens 92. However, due to the presence of the light shielding plate 94, light within the range of angle θ2 (see FIG. 12) narrower than θ1 is incident on the lens 92. Therefore, by providing the light shielding plate 94, the captured image is darker by angle θ3 (= angle θ1 - angle θ2) compared to when the light shielding plate 94 is not present. Here, angle θ1 is narrower than the camera angle of view (angle θ1 + angle θ4 × 2), and refers to the angle at which light emitted from the light-emitting unit 99 of the light source 93 passes through the focal position 95 and enters the lens 92.

[0006] 13 and 14 show examples of the path of light when a glass plate to be photographed is placed at focal position 95. A glass plate 96 is transported so as to pass through focal position 95 of a lens 92, and in this state, a line sensor camera 91 photographs the glass plate 96. If a deformed portion 83 (see FIG. 11) caused by a bubble or the like occurs in the glass plate 96, the path of the light that has passed through focal position 95 also changes.

[0007] 13 shows a case where the path of light that has passed through focal position 95 is moved away from light blocking plate 94 due to refraction equivalent to angle θ5 caused by surface deformation of the portion of glass plate 96 that passes through focal position 95. In the example shown in FIG. 13, the light that has passed through focal position 95 is not blocked by light blocking plate 94. Therefore, in the captured image, the portion of glass plate 96 that passes through focal position 95 becomes bright.

[0008] 14 shows a case where the path of light that has passed through focal position 95 is shifted toward the light blocking plate 94 due to refraction at an angle θ6 caused by surface deformation of the portion of glass plate 96 that passes through focal position 95. In the example shown in FIG. 14, the light that has passed through focal position 95 is blocked by light blocking plate 94. Therefore, in the captured image, the portion of glass plate 96 that passes through focal position 95 appears dark. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] International Publication No. 2012 / 077542 Brochure Summary of the Invention [Problem to be solved by the invention]

[0010] The applicant predicted that there is a correlation between the major axis of a bubble and the surface deformation amount H corresponding to that bubble (see FIG. 11). Based on the prediction that the surface deformation amount H is equal to or greater than a specified value when the major axis of a bubble is equal to or greater than a threshold, the applicant attempted to classify a glass sheet containing bubbles equal to or greater than the threshold as a glass sheet whose surface deformation amount H is equal to or greater than a specified value (i.e., a defective product).

[0011] However, when the relationship between the major diameter of a bubble and the corresponding surface deformation amount H was actually confirmed, there was no correlation between the two. FIG. 15 is a schematic diagram showing the results of confirming the relationship between the major diameter of a bubble and the surface deformation amount H. A glass sheet whose surface deformation amount H is less than the specified value A is deemed to be acceptable as a non-defective product. As shown in FIG. 15, since there is no correlation between the major diameter of a bubble and the surface deformation amount H, even if glass sheets whose major diameter of a bubble is less than the threshold value Th0 are deemed to be non-defective, some of these glass sheets have a surface deformation amount H equal to or greater than the specified value A. Therefore, it is not possible to determine whether a glass sheet is non-defective based solely on the major diameter of a bubble.

[0012] It is also possible to inspect the glass plate by directly measuring the surface deformation amount H using a laser microscope. However, since measuring the surface deformation amount H using a laser microscope takes time, this method is not practical.

[0013] Therefore, an object of the present invention is to provide a glass plate inspection method capable of non-contactly inspecting whether or not a glass plate has been deformed by a predetermined amount or more. [Means for solving the problem]

[0014] The inventors of the present invention have found that in an image obtained by photographing a glass plate using the Schlieren method, there is a correlation between the image feature amount derived from the gradation value near the maximum gradation value and the gradation value of the background region corresponding to the portion other than defects (bubbles, foreign matter, etc.), and the surface deformation amount H. Based on this finding, the inventors have made the following invention.

[0015] The glass plate inspection method according to the present invention is characterized in that it generates a photographed image of the glass plate by photographing the glass plate using the Schlieren method, and determines whether the amount of surface deformation that has occurred in the glass plate is equal to or greater than a predetermined value based on the relationship between the gradation value near the maximum gradation value in the photographed image of the glass plate and the gradation value of a background region that corresponds to a portion other than defects in the photographed image.

[0016] The method may involve extracting an area containing a defect from a photographed image of the glass plate, and determining whether the amount of surface deformation occurring in the glass plate is equal to or greater than a predetermined value based on the relationship between the gradation value near the maximum gradation value in that area and the gradation value of the background area.

[0017] Alternatively, a method may be used in which a predetermined number of pixels are selected in descending order from the maximum gradation value, the average of the gradation values ​​of the selected predetermined number of pixels is corrected by the gradation value of the background region, and the result is defined as an image feature of the photographed image of the glass plate, and the amount of surface deformation occurring on the glass plate is determined to be equal to or greater than a predetermined value, provided that the image feature is equal to or greater than a threshold value.

[0018] Alternatively, the method may define the result of subtracting the gradation value of the background region from the maximum gradation value, correcting the value with the gradation value of the background region, as the image feature of the photographed image of the glass plate, and determine that the amount of surface deformation occurring on the glass plate is equal to or greater than a predetermined value, provided that the image feature is equal to or greater than a threshold value.

[0019] Alternatively, the method may define the value obtained by subtracting the gradation value of the background region from the maximum gradation value as the image feature of the photographed image of the glass plate, and determine that the amount of surface deformation occurring on the glass plate is equal to or greater than a predetermined value, provided that the image feature is equal to or greater than a threshold value. [Effects of the Invention]

[0020] According to the present invention, it is possible to inspect, without contact, whether or not a glass plate has been deformed by a predetermined amount or more. [Brief explanation of the drawings]

[0021] [Figure 1] 1 is an explanatory diagram showing an example of a glass plate inspection device that is applied to a glass plate inspection method of the present invention. [Figure 2] Schematic diagrams showing examples of captured images of a glass plate. [Figure 3] Schematic diagrams showing dynamic ranges before and after change. [Figure 4] FIG. 10 is a schematic diagram showing an example of a circumscribing rectangle of an area corresponding to a bubble. [Figure 5]5A and 5B are schematic diagrams showing examples of areas extracted by a surrounding area extraction unit. [Figure 6] 1 is a flowchart showing an example of a glass plate inspection method of the present invention. [Figure 7] FIG. 10 is a schematic diagram showing an example of a histogram of gradation values ​​of a captured image after the processing in step S2. [Figure 8] 3 is a schematic diagram showing an example of a histogram of the gradation values ​​of each pixel in the region 25. FIG. [Figure 9] 3 is a schematic diagram showing an example of a histogram of the gradation values ​​of each pixel in the region 26. FIG. [Figure 10] 10 is a schematic diagram showing the relationship between image feature amounts and surface deformation amounts H. FIG. [Figure 11] Schematic diagram showing deformation of a glass plate. [Figure 12] FIG. 1 is an explanatory diagram showing an example of a Schlieren method. [Figure 13] FIG. 10 is an explanatory diagram showing an example of a light path when a glass plate to be photographed is placed at a focal position. [Figure 14] FIG. 10 is an explanatory diagram showing an example of a light path when a glass plate to be photographed is placed at a focal position. [Figure 15] Schematic diagram showing the results of confirming the relationship between the major axis of the bubble and the amount of surface deformation H. DETAILED DESCRIPTION OF THE INVENTION

[0022] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0023] FIG. 1 is an explanatory diagram showing an example of a glass plate inspection device that can be applied to the glass plate inspection method of the present invention. The glass plate inspection device shown in FIG. 1 includes a light source 3, a line sensor camera 1, and a light shielding plate 4 as elements for photographing a glass plate using the Schlieren method. The glass plate inspection device also includes a calculation device 10. A photographed image of the glass plate obtained by photographing using the Schlieren method is input to the calculation device 10. Then, based on the photographed image, the calculation device 10 determines whether the amount of surface deformation H that has occurred in the glass plate is equal to or greater than a predetermined value. This predetermined value is a reference value that is set in advance, and a glass plate that has experienced a surface deformation equal to or greater than this predetermined value is deemed to be a defective product.

[0024] The light source 3 is positioned so that light is incident on the lens 2 of the line sensor camera 1. In addition, a light shielding plate 4 is provided between the focal position 5 of the lens 2 and the lens 2. The light shielding plate 4 is positioned so as to block a portion of the light component of the light from the light source 3 that is incident on the lens 2. Light that has passed through the focal position 5 of the lens 2 is incident on the lens 2. However, due to the presence of the light shielding plate 4, a portion of the light that has passed through the focal position 5 is blocked by the light shielding plate 4.

[0025] Assume that nothing is placed at the focal position 5. When comparing the image obtained when the light shielding plate 4 is provided with the image obtained when the light shielding plate 4 is not provided, the image obtained when the light shielding plate 4 is provided (as shown in Figure 1) is darker.

[0026] A glass plate 6 to be inspected for surface deformation of a predetermined value or more is conveyed in one direction so as to pass through a focal position 5 of a lens 2. At this time, the glass plate 6 is conveyed with its main surface perpendicular to the optical axis of the lens 2. The line sensor camera 1 then captures an image of the glass plate 6 as it passes through the focal position 5. The line sensor camera 1 then generates a captured image of the glass plate 6. This imaging mode corresponds to an imaging mode using the Schlieren method.

[0027] If a defect (a bubble or foreign matter) exists in the glass plate 6, resulting in a deformed portion 83 (see FIG. 11), a difference in brightness occurs in the image of the glass plate 6 captured by the line sensor camera 1 due to refraction caused by the surface deformation at the point passing through the focal position 5. The following explanation will be given using an example in which the defect is a bubble, but the following explanation also applies to all points with surface deformation, even when the defect is a foreign matter.

[0028] FIG. 2 is a schematic diagram showing an example of a captured image of a glass plate 6. In the schematic diagram shown in FIG. 2, positions that pass through the focal position 5 earlier appear on the right side of the captured image, and positions that pass through the focal position 5 later appear on the left side of the captured image. In the captured image, areas with lower gradation values ​​are darker, and areas with higher gradation values ​​are brighter. In the captured image, area 21 corresponding to a bubble is black. In the captured image, area 22 to the right of area 21 (i.e., the area that passes through the focal position 5 just before the bubble) is white. In the captured image, area 23 to the left of area 21 (i.e., the area that passes through the focal position 5 just after the bubble) is black. However, there is a difference in the degree of darkness between area 23 and area 21, with area 21 being a darker black than area 23. In addition, background area 24 (excluding areas 22 and 23) that does not correspond to a defect is gray (i.e., a dark intermediate tone). Note that FIG. 2 shows an example of the regions 22 and 23 that appear in the image, and the extent and area of ​​the regions 22 and 23 are not limited to the example shown in FIG.

[0029] The magnitude relationship of the gradation values ​​of the pixels in each region of the image is (gradation value of region 21)<(gradation value of region 23)<(gradation value of background region 24)<(gradation value of region 22).

[0030] The calculation device 10 includes a dynamic range change unit 11, a defective area detection unit 12, a surrounding area extraction unit 13, and a determination unit .

[0031] The dynamic range modification unit 11 modifies the gradation values ​​of each pixel in the captured image of the glass plate 6 so as to expand the dynamic range of the gradation values ​​of the captured image. The dynamic range is the difference between the maximum and minimum gradation values ​​obtained from the image data of the captured image. Figure 3 is a schematic diagram showing the dynamic range before and after modification. The upper part of Figure 3 schematically shows a histogram of the gradation values ​​of pixels in the captured image before the dynamic range modification, and the lower part of Figure 3 schematically shows a histogram of the gradation values ​​of pixels in the captured image after the dynamic range modification of the same captured image as the upper part. In both histograms shown in Figure 3, the horizontal axis represents gradation value and the vertical axis represents frequency. The captured image before the dynamic range modification had a narrow dynamic range (see the upper part of Figure 3). The dynamic range modification unit 11 modifies the gradation values ​​of each pixel to expand this dynamic range. As a result, as shown schematically in the lower part of Figure 3, the difference in peak positions on the histogram becomes clearer, resulting in an image in which the difference in gradation values ​​in each region is easily recognized.

[0032] The dynamic range change unit 11 may perform processing to widen the dynamic range of the gradation values ​​using a known method.

[0033] When an image contains an area corresponding to bubbles, three frequency peaks appear, as shown in the histogram in the lower part of Figure 3. Peak A in the lower part of Figure 3 corresponds to the gradation value of area 21. Peak B in the lower part of Figure 3 corresponds to the gradation value of background area 24. Peak C in the lower part of Figure 3 corresponds to the gradation value of white area 22.

[0034] The defect area detection unit 12 identifies a circumscribing rectangle of the region 21 corresponding to a defect (a bubble in this example) from the captured image after the dynamic range has been changed. FIG. 4 is a schematic diagram showing an example of the circumscribing rectangle of the region 21. For example, the defect area detection unit 12 sets threshold information such that a region consisting of a series of pixels with a gradation value of 80 or less corresponds to a defect. In this example, 80 is used as the threshold, but any threshold capable of identifying the region 21 corresponding to a bubble may be determined in advance. The defect area detection unit 12 identifies a region consisting of a series of pixels with a gradation value of less than the threshold (80 in the above example) from the captured image after the dynamic range has been changed. As a result, the region 21 corresponding to a bubble is determined. The defect area detection unit 12 then further identifies a circumscribing rectangle of the region (a region consisting of a series of pixels with a gradation value of less than the threshold; in other words, the region 21 corresponding to a bubble). As a result, the region 25 shown in FIG. 4 is identified.

[0035] The surrounding area extraction unit 13 extracts the area surrounding the area 21 based on the area 25 of the circumscribing rectangle of the area 21 identified by the defect area detection unit 12. FIG. 5 is a schematic diagram showing an example of an area extracted by the surrounding area extraction unit 13. The surrounding area extraction unit 13 identifies a rectangular area 26 in which the lengths of the short and long sides of the circumscribing rectangle (area 25) of the area 21 are multiplied by a predetermined factor. Here, an example will be described in which the lengths of the short and long sides of the circumscribing rectangle area 25 are doubled, but the value of this multiplication factor is not limited to two. In this case, the surrounding area extraction unit 13 determines, as the area 26, a rectangular area whose center is the same as that of the area 25 and whose long sides are parallel to the long sides of the area 25. As a result, the area 26 shown in FIG. 5 is identified. The surrounding area extraction unit 13 extracts the area 26 from the captured image after the dynamic range has been changed. In addition to the pixels that fall within the region 21, a region 26 surrounding the region 21 also includes pixels that fall within the region 22 and pixels that fall within the background region 24. Pixels that fall within the region 23 are also included.

[0036] The determination unit 14 determines whether the amount of surface deformation of the glass plate 6 to be inspected is equal to or greater than a predetermined value, based on the relationship between the gradation values ​​near the maximum gradation value in the region 26 extracted by the surrounding region extraction unit 13 and the gradation values ​​of the background region 24. In this embodiment, the determination unit 14 selects a predetermined number of pixels (10 in this example) from the region 26 extracted by the surrounding region extraction unit 13 in descending order from the maximum gradation value, and calculates the average of the gradation values ​​of the selected 10 pixels. The determination unit 14 also identifies the gradation values ​​of pixels that fall within the background region 24, and divides the result of this average calculation by the gradation values ​​of pixels that fall within the background region 24. In this embodiment, the result of this division is used as the image feature of the captured image.

[0037] Then, the determination unit 14 determines that the surface deformation amount H of the glass plate 6 is equal to or greater than a predetermined value, provided that this image feature amount is equal to or greater than a predetermined threshold (a threshold for determining whether the glass plate is a good product). That is, if the image feature amount is equal to or greater than the threshold, the determination unit 14 determines that a deformed portion 83 having a surface deformation amount H equal to or greater than a predetermined value has occurred in the glass plate 6, and that the glass plate is not a good product. On the other hand, if the image feature amount is less than the threshold, the determination unit 14 determines that a deformed portion 83 having a surface deformation amount H equal to or greater than a predetermined value has not occurred in the glass plate 6, and that the glass plate is a good product.

[0038] FIG. 6 is a flowchart showing an example of the glass plate inspection method of the present invention.

[0039] First, the line sensor camera 1 (see FIG. 1) photographs a glass plate by the Schlieren method to generate a photographed image of the glass plate (step S1). Specifically, with light emitted from the light source 3, the glass plate 6 to be inspected is transported so as to pass through the focal position 5 of the lens 2 of the line sensor camera 1. In this state, the line sensor camera 1 photographs the glass plate 6 as it passes through the focal position 5.

[0040] The line sensor camera 1 inputs the captured image of the glass plate 6 generated in step S1 to the calculation device 10. Then, the dynamic range change unit 11 changes the gradation value of each pixel of the captured image so as to widen the dynamic range of the gradation values ​​of the captured image (step S2). Figure 7 is a schematic diagram showing an example of a histogram of gradation values ​​of the captured image processed in step S2. Figure 7 illustrates an example where frequency peaks appear at gradation values ​​of 50, 100, and 110.

[0041] After step S2, the defective area detection unit 12 identifies a circumscribing rectangle area 25 (see FIG. 4) of the area 21 corresponding to the bubble from the captured image processed in step S2 (step S3). As in the above example, assume that "80" is predetermined as the threshold for identifying the area 21. The defective area detection unit 12 determines from the captured image that an area of ​​consecutive pixels having a gradation value equal to or less than the threshold 80 is the area 21 corresponding to the bubble, and identifies the area 25 that is the circumscribing rectangle of the area 21. FIG. 8 is a schematic diagram showing an example of a histogram of the gradation values ​​of each pixel in the area 25. The area 25 is mostly occupied by the area 21 corresponding to the bubble. Therefore, in the histogram shown in FIG. 8, the frequency of gradation values ​​equal to or less than 80 is high.

[0042] In step S3, if there are no pixels whose gradation values ​​are equal to or less than the threshold, this means that there are no regions 21 corresponding to bubbles. In this case, the defective region detection unit 12 notifies the determination unit 14 that there are no regions 21 corresponding to bubbles. When the determination unit 14 receives this notification, it determines that no deformed portion 83 with a surface deformation amount H equal to or greater than a predetermined value has occurred in the glass plate 6, and that the glass plate is a non-defective product. In this case, the processing from step S4 onwards is not performed.

[0043] The following description will be given on the assumption that a region 21 corresponding to a bubble exists and a region 25 that is a circumscribing rectangle of the region 21 has been identified.

[0044] After step S3, the surrounding area extraction unit 13 extracts an area 26 (see FIG. 5) around the area 21 based on the area 25 identified in step S3 (step S4). The surrounding area extraction unit 13 extracts a rectangular area with the lengths of the short sides and long sides double those of the area 25, with the center position being the same as that of the area 25 and the long sides parallel to the long sides of the area 25, as the area 26 around the area 21.

[0045] Next, the determination unit 14 creates a histogram of the gradation values ​​of the region 26 extracted in step S4 and identifies the gradation values ​​of the pixels in the background region (step S5). Fig. 9 is a schematic diagram showing an example of a histogram of the gradation values ​​of each pixel in the region 26. The region 26 includes pixels that fall within the white region 22 and pixels that fall within the background region 24, and the frequency peaks near the gradation values ​​of the region 22 and the background region 24. The determination unit 14 determines the gradation value corresponding to the frequency peak B that appears second in descending order of gradation values ​​in this histogram as the gradation value of the pixel in the background region 24.

[0046] In this embodiment, an example has been described in which the gradation values ​​of pixels in background region 24 are determined from a histogram of gradation values ​​in region 26, but other methods may be used to derive the gradation values ​​of pixels in background region 24. For example, the gradation values ​​of pixels in the background region may be considered to be approximately constant, and the gradation values ​​of pixels in the background region may be determined in advance as constants.

[0047] Then, the determination unit 14 calculates the image feature amount of the captured image (step S6). In this embodiment, the determination unit 14 selects a predetermined number of pixels (10 in this example) from the pixels in the region 26 in descending order from the maximum gradation value, and calculates the average of the gradation values ​​of the 10 pixels. The determination unit 14 then divides this average by the gradation value of the pixel in the background region 24, and sets the result of this division as the image feature amount. That is, the determination unit 14 calculates the image feature amount by the calculation of the following equation (1).

[0048] Image feature value = (average of the top 10 pixel gradation values) / gradation value of pixels in the background area Formula (1)

[0049] The above 10 pixels can be considered to be pixels in the white region 22 (see FIG. 5). That is, the image feature amount is calculated from the gradation values ​​of the pixels in the white region 22 and the gradation values ​​of the pixels in the background region 24. Note that the number of pixels selected in step S6 may be other than 10.

[0050] Then, the determination unit 14 determines whether or not a deformed portion 83 with a surface deformation amount H of a predetermined value or more has occurred in the glass plate to be inspected by determining whether or not the image feature amount is equal to or greater than a threshold (step S7). If the image feature amount is equal to or greater than the threshold, the determination unit 14 determines that a deformed portion 83 with a surface deformation amount H of a predetermined value or more has occurred in the glass plate 6, and that the glass plate is not a good product. On the other hand, if the image feature amount is less than the threshold, the determination unit 14 determines that a deformed portion 83 with a surface deformation amount H of a predetermined value or more has not occurred in the glass plate 6, and that the glass plate is a good product.

[0051] The inventors have confirmed the relationship between the image feature value calculated by equation (1) and the surface deformation amount H. FIG. 10 is a schematic diagram showing the results of this confirmation. As shown in FIG. 10, there is a relationship in which the larger the value of the image feature value, the larger the surface deformation amount H. Therefore, based on the relationship shown in FIG. 10, a threshold value Th is defined for determining whether the surface deformation amount H is equal to or greater than a predetermined value A. In step S7, by determining whether the image feature value is equal to or greater than the threshold value Th, it is possible to determine whether a deformed portion 83 in which the surface deformation amount H is equal to or greater than the predetermined value has occurred on the glass plate.

[0052] Furthermore, according to this embodiment, the line sensor camera 1 captures an image of the glass plate 6, and the computing device 10 calculates image features from the captured image of the glass plate 6 and determines whether or not a deformed portion 83 with a surface deformation amount H equal to or greater than a predetermined value has occurred on the glass plate. Therefore, it is possible to determine whether or not a deformation equal to or greater than a predetermined value has occurred on the glass plate without coming into contact with the glass plate.

[0053] Alternatively, the arithmetic unit 10 may calculate the image feature amount by using the entire captured image to find the average of the gradation values ​​of the top 10 pixels in gradation value and the gradation values ​​of the pixels in the background area. However, as in the above embodiment, the amount of calculation can be reduced by extracting area 26 from the entire captured image and calculating the image feature amount for area 26. Therefore, it is more preferable to extract area 26 and perform the processing from step S5 onwards for area 26, as in the above embodiment.

[0054] Next, a modified example of this embodiment will be described. In the above embodiment, the determination unit 14 calculates the image feature amount by the calculation of the above-mentioned formula (1). A value obtained by a calculation other than formula (1) may be used as the image feature amount. Below, an example will be shown in which the image feature amount is obtained by a calculation other than formula (1). Note that the processing of steps S1 to S5 is the same as the processing of steps S1 to S5 in the above embodiment.

[0055] In step S6, the determination unit 14 identifies the maximum value of the gradation values ​​of the pixels in the region 26 (i.e., the maximum gradation value). The determination unit 14 may then subtract the gradation values ​​of the pixels in the background region 24 from the maximum gradation value, and calculate the result of this subtraction by the gradation value of the pixels in the background region 24 as the image feature amount. That is, the determination unit 14 may calculate the image feature amount by the calculation of the following equation (2).

[0056] Image feature value = (maximum gradation value - gradation value of pixel in background area) / gradation value of pixel in background area Formula (2)

[0057] The above maximum gradation value can be considered as the gradation value of the pixels in the white area 22. Therefore, when calculating equation (2), the image feature amount is calculated from the gradation values ​​of the pixels in the white area 22 and the gradation values ​​of the pixels in the background area 24.

[0058] There is also a relationship between the image feature quantity calculated by equation (2) and the surface deformation amount H such that the larger the value of the image feature quantity, the larger the surface deformation amount H. Therefore, from this relationship, a threshold value can be set in advance for determining whether the surface deformation amount H is equal to or greater than a predetermined value. Then, in step S7, the determination unit 14 determines whether the image feature quantity calculated by equation (2) is equal to or greater than the threshold value, thereby determining whether a deformed portion 83 in which the surface deformation amount H is equal to or greater than a predetermined value has occurred on the glass plate to be inspected.

[0059] Furthermore, when the gradation value of a pixel in the background region 24 is considered to be a constant that takes an approximately constant value, the determination unit 14 may use the result of subtracting the gradation value of a pixel in the background region 24 from the maximum gradation value in the region 26 as the image feature amount. That is, the determination unit 14 may calculate the image feature amount by the calculation of the following equation (3).

[0060] Image feature value = maximum gradation value - gradation value of pixels in background area Equation (3)

[0061] There is also a relationship between the image feature quantity calculated by equation (3) and the surface deformation amount H such that the larger the value of the image feature quantity, the larger the surface deformation amount H. Therefore, from this relationship, a threshold value can be set in advance for determining whether the surface deformation amount H is equal to or greater than a predetermined value. Then, in step S7, the determination unit 14 determines whether the image feature quantity calculated by equation (3) is equal to or greater than the threshold value, thereby determining whether a deformed portion 83 in which the surface deformation amount H is equal to or greater than a predetermined value has occurred on the glass plate to be inspected.

[0062] The same effect as in the above embodiment can be obtained when calculating the image feature amount using equation (2) or equation (3). It is possible that the maximum gradation value is the gradation value of white spot noise that occurs in the captured image. In this case, if the image feature amount is calculated using equation (2) or equation (3), the image feature amount will be calculated using the gradation value of the white spot noise, not the gradation values ​​of the pixels in the white area 22. On the other hand, when equation (1) is used, even if the top 10 pixels in terms of gradation value include pixels that correspond to white spot noise, the influence of the white spot noise can be minimized. Therefore, it is more preferable to calculate the image feature amount using equation (1). [Industrial Applicability]

[0063] The present invention is suitably applied to a glass plate inspection method for inspecting whether or not deformation of a predetermined magnitude or more has occurred on the surface of a glass plate. [Explanation of symbols]

[0064] 1 Line sensor camera 2 lenses 3 light source 4 Shade 5 Focus position 6 glass plates 10 Arithmetic unit 11 Dynamic range change section 12 Defective area detection unit 13 Surrounding area extraction unit 14 Judgment section

Claims

1. generating a photographed image of the glass plate by photographing the glass plate using a Schlieren method; Whether or not the amount of surface deformation occurring in the glass plate is equal to or greater than a predetermined value is determined based on the relationship between the gradation value near the maximum gradation value in the photographed image of the glass plate and the gradation value of a background region corresponding to a portion other than the defect in the photographed image. A glass plate inspection method comprising:

2. A region including a defect is extracted from a photographed image of the glass plate, and it is determined whether the amount of surface deformation occurring in the glass plate is equal to or greater than a predetermined value based on the relationship between the gradation value near the maximum gradation value in the region and the gradation value in the background region. The glass plate inspection method according to claim 1 .

3. A predetermined number of pixels are selected in descending order from the maximum gradation value, and the average of the gradation values ​​of the selected predetermined number of pixels is corrected by the gradation value of the background region, and the result is defined as an image feature amount of the photographed image of the glass plate. If the image feature amount is equal to or greater than a threshold value, it is determined that the amount of surface deformation occurring in the glass plate is equal to or greater than a predetermined value. The glass plate inspection method according to claim 1 or 2.

4. The result of subtracting the gradation value of the background region from the maximum gradation value and correcting the value by the gradation value of the background region is defined as an image feature amount of the photographed image of the glass plate, and the amount of surface deformation occurring in the glass plate is determined to be equal to or greater than a predetermined value, provided that the image feature amount is equal to or greater than a threshold value. The glass plate inspection method according to claim 1 or 2.

5. The value obtained by subtracting the gradation value of the background region from the maximum gradation value is defined as an image feature amount of the photographed image of the glass plate, and the amount of surface deformation occurring in the glass plate is determined to be equal to or greater than a predetermined value, provided that the image feature amount is equal to or greater than a threshold value. The glass plate inspection method according to claim 1 or 2.

Images