JP2017161435A - Inspection method of transparent body - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an inspection method capable of inspecting more surely and distinctively a bubble and a foreign matter entering a transparent body.SOLUTION: There is provided an inspection method for detecting contaminants 51, 52 contained in a transparent body by performing image processing of an image obtained by imaging the transparent body. In the image processing, a brightness is determined in segment 6 units comprising one pixel or a plurality of pixels from the image obtained by imaging the transparent body, and it is determined that, when a brightness difference between adjacent segments is large, the contaminant is a foreign matter 51, and that when the brightness difference is small, the contaminant is a bubble 52.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a method for inspecting a transparent body such as a glass plate or a resin plate.

  In the production process of a transparent body such as a glass plate or a resin plate, there is a possibility that foreign matters or bubbles may be mixed in the transparent body. And the foreign material and bubble which were mixed will be a factor of a performance fall, when a transparent body is an object for optical components. For this reason, it is required to inspect the transparent body for foreign matter or bubbles.

  On the other hand, since the optical influence is different between the case where the mixed substance is a foreign substance and the case where it is a bubble, it is desirable that the inspection method can distinguish the two.

  And as an inspection method corresponding to such a point, there is one disclosed in Patent Document 1. This means that when binarizing an image obtained by photographing a transparent body, the binarization level is set to a level at which a ring shape can be obtained in consideration of the occurrence of a ring-shaped low-luminance portion for round bubbles. Thus, the bubbles can be identified.

JP-A-8-285789

  However, the bubbles are not always round (spherical), and binarization is indistinguishable from the case of foreign matter even though they are bubbles when binarized at a predetermined binarization level. It often happens that only information can be obtained, and as a result, bubbles and foreign substances cannot be distinguished.

  The present invention has been made in view of the above problems, and it is an object of the present invention to provide a method for inspecting a transparent body that can more reliably perform the inspection by distinguishing between bubbles and foreign matters mixed in the transparent body. To do.

  According to the transparent body inspection method of the present invention, in the transparent body inspection method for detecting contaminants contained in the transparent body by performing image processing on an image of the transparent body, the transparent body is imaged. The luminance is obtained from an image in a unit of one or more pixels, and when the luminance difference between adjacent segments is large, the contaminant is a foreign substance, and when the luminance difference is small, the contaminant Is characterized in that it is determined to be a bubble.

  In the present invention, regardless of the size of the bubble or the shape of the bubble, it can be identified whether it is a foreign object or a bubble. It can be done reliably.

It is a block diagram which shows the outline of the test | inspection apparatus for enforcing the test | inspection method of this invention. A is an explanatory diagram of a captured image, B is an explanatory diagram of size detection by edge detection, and C is an explanatory diagram of segment scanning. It is explanatory drawing of an example of the segment scan of the order of t1, t2, t3, t4. A is a line graph regarding the change in luminance difference, and B is a graph showing the change in the primary differential value of the change in luminance difference. It is explanatory drawing which showed the determination result whether it is a defect or a non-defect. It is a flowchart about one Example of this invention.

  The present invention provides an inspection method for a transparent body in which an image obtained by imaging a transparent body is subjected to image processing to detect a contaminant contained in the transparent body. The luminance is calculated in units of pixels, and when the luminance difference between adjacent segments is large, it is determined that the contaminant is a foreign substance, and when the luminance difference is small, the contaminant is a bubble. .

  Then, in addition to the image processing for obtaining the brightness, the edge of the image obtained by imaging the transparent body is detected to determine the size of the contaminant, and whether the contaminant is a foreign matter based on the size determination of the contaminant and the brightness difference. It is preferable to determine whether the contaminant is defective or non-defective by determining whether it is a bubble.

  Further, the edge of an image obtained by imaging a transparent body may be detected to determine the size of the contaminant, and the luminance may be obtained in segment units having a size corresponding to the obtained size.

  Hereinafter, the present invention will be described in detail with reference to the illustrated embodiment. FIG. 1 shows an inspection apparatus for inspecting a resin-made transparent body 1 formed as a so-called light guide. An illuminator 2 that illuminates the image, an imaging unit 3 that includes a CCD camera or the like that captures light transmitted through the transparent body 1, and a processing circuit 4 that processes an image obtained by the imaging unit 3 and performs a determination process from the processing result Consists of. A computer can be suitably used as the processing circuit 4.

  The contaminant 5 mixed in the transparent body 1 may be a foreign matter 51 or a bubble 52, which appears in the image as shown in FIG. 2A. In the case of the bubbles 52, since a part of the light passes, the decrease in the luminance is less than that in the case of the foreign matter 51. In this case, the foreign matter 51 is painted black and the bubbles 52 are hatched. ing.

  The processing circuit 4 temporarily stores the luminance signal of the image sent from the imaging means 3 in the storage unit 41, and then detects the luminance of each pixel stored in the x and y directions as shown in FIG. 2B. Then, the size S of the contaminant 5 (foreign matter 51, bubble 52) is detected. In this processing, the processing circuit 4 detects the size with the edge sensitivity (binarization level) that can detect the bubbles 52 as well as the foreign matter 51.

  Then, the processing circuit 4 determines from the above-described size detection results that if the size S of the foreign matter 51 and the bubble 52 is less than the extremely small size Smin, these are not regarded as defects. Is determined.

  Next, the processing circuit 4 rescans the luminance of each pixel stored in the storage unit 41 with respect to the foreign matter 51 and the bubble 52 that are the temporary defects. However, in this rescanning, the processing circuit 4 calculates an average luminance value in each segment 6 for each segment 6 with a set of a plurality of pixels as a segment 6 as shown in FIG. 2C.

  The change in average luminance of each segment 6 when the region having the contaminant 5 as a temporary defect is scanned in the order indicated by t1, t2, t3, and t4 in units of segments 6 as shown in FIG. It will be shown in Here, the line 5 a indicates when the contaminant 5 is a foreign substance 51, and the line 5 b indicates when the contaminant 5 is a bubble 52. That is, the average luminance difference between the adjacent segments 6 is large when the contaminant 5 is the foreign matter 51 and is small when the contaminant 5 is the bubble 52.

  This is because the difference between the bubbles 52 and the foreign matter 51 is small when the brightness of the peripheral portion of the contaminant 5 is compared between the bubbles 52 and the foreign matter 51. However, the brightness of the central portion of the contaminant 5 is higher than that of the foreign matter 51 because light easily passes through the bubble 52. Therefore, the difference between the portion at t4 where the central portion of the contaminant 5 is mainly extracted and the portions at t1, t2 and t3 where the peripheral portion of the contaminant 5 is mainly extracted This is because the part becomes larger.

  Therefore, as shown in FIG. 4B, when the primary differential value of the change in average luminance due to scanning is larger than the predetermined value K, the temporary defect is due to the foreign matter 51 and the primary differential value is less than the predetermined value K. When it is, it can be determined that the temporary defect is due to the bubble 52.

  Thus, since the processing circuit 4 can determine whether the contaminant 5 is the foreign matter 51 or the bubble 52, when the contaminant 5 is the foreign matter 51 and when the contaminant 5 is the bubble 52. When it is, it is possible to set different criteria for determining the defect.

  Incidentally, FIG. 5 shows the contaminant 5 as a temporary defect. The contaminant 5 in the upper left and the contaminant 5 in the upper right are both determined to be a foreign substance 51 as a result of the above determination, so that it is designated as a defect F, and the contaminant 5 in the lower right. Is determined as a bubble 52 as a result of the determination, but the size S is equal to or greater than the predetermined size Smax, so that it is defined as a defect F, and the lower left contaminant 5 is determined to be a bubble 52 as a result of the determination. In this example, the non-defective NF is determined because the thickness S is smaller than the predetermined size Smax. FIG. 6 shows a flowchart of the determination operation performed in the processing circuit 4.

  If the size 5 of the contaminant 5 (foreign matter 51 and bubble 52) is greater than or equal to the predetermined size Smax at the stage where the size of the contaminant 5 is detected by edge detection, the processing circuit 4 determines that all of these are defects. Processing may be performed, and the processing based on the luminance difference may be skipped for the contaminant 5 having the size Smax or more.

  Further, the contaminants 5 having an extremely small size Smin may not be detected by adjusting the edge sensitivity at the edge detection stage described above.

  Here, the size of the segment 6 as a set of a plurality of pixels is determined from the relationship between the size Smin and the predetermined value K. In addition, the size of the segment may be changed according to the size of the contaminant 5 obtained by the edge detection.

  By the way, in the scanning of the segment 6 shown in FIG. 3, some pixels are overlapped by the adjacent segments 6 and 6 by setting segment size> segment movement amount. This is to increase the resolution, but segment size = segment movement amount may be used, and calculation speed may be improved by segment size <segment movement amount.

  In any case, in order to detect the bubble 52 or the foreign object 51 based on the average luminance in the segment 6 as a set of a plurality of pixels, the bubble 52 is formed by adjusting the edge sensitivity (binarization level). Compared to the conventional example in which the distinction is made, the foreign object 51 or the air bubble 52 can be identified regardless of the size of the air bubble 52 and the shape of the air bubble 52.

  It should be noted that the segment 5 may be one pixel when it is necessary to determine whether the contaminant 5 is less than the extremely small size Smin as the foreign matter 51 or the bubble 52. In this case, the foreign object 51 and the bubble 52 are identified based on the luminance difference between adjacent pixels.

DESCRIPTION OF SYMBOLS 1 Transparent body 2 Illuminator 3 Imaging means 4 Processing circuit 5 Contaminant 6 Segment 51 Foreign object 52 Bubble

Claims (3)

A method for inspecting a transparent body, wherein an image obtained by imaging the transparent body is subjected to image processing to detect contaminants contained in the transparent body,
In the above image processing, the brightness is obtained from an image obtained by imaging a transparent body in units of one or more pixels, and when the brightness difference between adjacent segments is large, the contaminant is a foreign substance, and the brightness When the difference is small, it is determined that the contaminant is a bubble.   In addition to the image processing for obtaining the brightness, edge detection is performed on an image obtained by imaging a transparent body to determine the size of the contaminant, and the size of the contaminant is determined. 2. The method for inspecting a transparent body according to claim 1, wherein whether or not the contaminant is a defect is determined by determining whether or not the contaminant is present.   The edge detection of the image which imaged the transparent body is performed, the magnitude | size of a contaminant is determined, The said brightness | luminance is calculated | required per segment of the magnitude | size according to the obtained magnitude | size. Inspection method of transparent body.

Images