TW201233993A

TW201233993A - Method and system for measuring defect in glass ribbon

Info

Publication number: TW201233993A
Application number: TW100145663A
Authority: TW
Inventors: Shinji Fujii; Makoto Kurumisawa; Makoto Yoshikawa
Original assignee: Asahi Glass Co Ltd
Priority date: 2010-12-09
Filing date: 2011-12-09
Publication date: 2012-08-16
Also published as: CN103250047A; JPWO2012077683A1; WO2012077683A1; KR20130140058A

Abstract

Provided is a method for measuring a defect in a glass ribbon, which makes it possible to measure the position of a defect in a glass ribbon in the height direction even when the defect is located closed to an interface of the glass ribbon or the defect is large. A circumscribing rectangle is specified for two overlapping images from among pictures obtained by imaging a glass ribbon being transported. A height (h) in real space corresponding to a side (24) of the circumscribing rectangle is then calculated. Furthermore, a height in real space corresponding to the distance from a center region (21a) of an image (21) from among the pictures to the closest short side is calculated, and a length (s) that is twice said height is calculated. Thereafter, h - s is calculated to obtain a movement distance (yd) of a defect from the position where a first image is captured to the position where a second image is captured, and the position of the defect in the height direction is calculated using the movement distance (yd) and an angle of refraction (ss).

Description

201233993 六、發明說明：【發明所屬之技術領域】本發明係關於一種進行與玻璃帶内之缺陷相關之測定的玻璃帶内缺陷測定方法及玻璃帶内缺陷測定系統，尤其關於一種測定玻璃帶中之缺陷之高度方向位置等的破璃帶内缺陷測定方法及玻璃帶内缺陷測定系統。【先前技術】提出有各種測定玻璃帶内之缺陷之高度方向位置等之方法。作為測定玻璃帶内之缺陷之高度方向位置之一般方法，有於拍攝缺陷時調節相機之焦點而測定缺陷之高度方向位置之方法。為方便起見將該方法稱作第！測定方法。圖i〇A 係模式性地表示第1測定方法之說明圖。第丨測定方法中，如圖10A所示，於使光通過玻璃帶82之狀態下搬送玻璃帶 82而且，利用線陣相機Gine camera)8 1拍攝所搬送之玻璃帶82之内部。若於玻璃帶82之内部存在缺陷以，則缺陷 83被拍攝。圖10B表示所拍攝之缺陷之圖像之例。圖i〇a 中，將缺陷83模式性地以長方形表示，Wl〇B中亦將玻璃帶之圖像内出現之缺陷之像86以長方形表示，但缺陷之形 &並不限於長方形。其中，以下所示之圖nA、圖nB、圖UA、圖12β、圖13、及圖14中亦將缺陷模式性地以長方形表不。再者，圖1〇B所示之箭頭為玻璃帶Μ之搬送方向。於利用線陣相機81拍攝玻璃帶82之内部時，調節相機之焦點，使缺陷之存在位置與相機之焦點一致，測定自線 160767.doc 201233993 陣相機8 1至缺陷為止之絕 > 絕對距離’根據該距離計算缺陷之局度方向位置。作為調節相々凋郎相機之焦點而使缺陷之存在位置與相機之焦點-致之方法，有_(以帥—F。㈣對焦深度）法等。又，關於缺陷之尺寸，對所拍攝之圖像進行圖像處理而測定缺陷之尺寸。調節相機之焦點而測定社一州疋缺陷之尚度方向位置之方法及裝置揭示於例如專利文獻1〜3等中。广’作為測定玻璃帶内之缺陷之高度方向位置之其他一般方法，有如下方法：利用入射至玻璃帶之光之反射光，於2個位置拍攝到同—缺陷，根據其結果所得之2個像之位置關係収缺陷之高度方向位置。為方便起見將該方法記乍第/則定方法。圖1 i A係模式性地表示第2測定方法之說明圓。第2測定方法中’例如圖11A所示，於玻璃帶82中，使光於與線陣相機81同-側入射，其反射光到達線陣相機 81。而且，搬送玻璃帶82，利用線陣相機。拍攝玻璃帶μ 之内部。關於玻璃帶内之光之路徑，之後參照圓13之上段所不之側視圖進行敍述。缺陷83隨著玻璃帶82之搬送而移動，與反射前之光之路徑重疊時及與反射後之光之路徑重疊時，分別於線陣相機81中作為像而被捕捉。其結果為，即便缺陷83為1個，於所拍攝之圖像中亦映出2個像。圖 11B係第2測定方法中所拍攝之圖像之例。如圖i ΐβ所示，對於同一缺陷映出2個像84、85。於第2測定方法中，根據圖11B所例示之圖像中之2個像之位置關係，計算缺陷以之高度方向位置。又，關於缺陷之尺寸，對所拍攝之圖像進 160767.doc 201233993 行圖像處理而測定缺陷之尺寸。再者，圖丨丨B所示之箭頭為玻璃帶82之搬送方向。利用入射至透明基板等之光之反射光，於2個位置拍攝到同一缺陷，根據2個像之位置關係測定缺陷之高度方向位置的方法及裝置揭示於例如專利文獻4〜6、8等中。又’有如下方法：於玻璃帶之兩面，與第2測定方法同樣地拍攝圖像，根據於玻璃帶之各個面所拍攝之圖像内之像之位置關係測定缺陷之高度方向位置。為方便起見將該方法記作第3測定方法。圖12A係模式性地表示第3測定方法之說明圖。第3測定方法中，例如圖12A所示，於玻璃帶 82中，使光於與第1線陣相機8丨a同一側入射，其反射光到達第1線陣相機81a 〇同樣地，使光於與第2線陣相機8“同一側入射，其反射光到達第2線陣相機811> ^而且，搬送玻璃帶82，利用第1線陣相機及第2線陣相機8ib分別拍攝玻璃帶82之内部。於是，於第1線陣相機813中，與第2測定方法之情形同樣地捕捉到2個像。又，於第2線陣相機 8U中亦捕捉到2個像。圖12B係第3測定方法中所拍攝之圖像之例。第3測定方法中，如圖12B所示，獲得一線陣相機自玻璃帶之上側所拍攝之圖像、及另—線陣相機自玻璃帶之下側所拍攝之圖像。於各圖像中，分別映出2個像。第3 測定方法中，根據自玻璃帶之上側及下側所拍攝之各圖像中之像之位置關係，計算缺陷83之高度方向位置。再者，圖12B中，例示自上側所拍攝之圖像中像重疊之情形。又’關於缺陷之尺寸，對所拍狀圖像進行圖像處理而測 160767.doc 201233993 定缺陷之尺寸。再者，圖丨2B所示之箭頭為玻璃帶82之搬送方向。自透明基板等之兩側拍攝圖像而求出缺陷之高度方向位置之方法揭示於例如專利文獻7中。第2 J丨疋方法及第3測疋方法中’以同一缺陷之像於圖像内不重疊為條件’計算缺陷之高度方向位置。再者，第3 測定方法中，於如圖12Β所例示般在一圖像中像重疊之情形時’利用另一圖像計算缺陷之高度方向位置即可。以下，示出根據第2測定方法中所拍攝之圖像内之2個像之位置關係，測定缺陷之高度方向位置之具體例。圖13係表示所搬送之玻璃帶内之缺陷被線陣相機拍攝時之位置之說明圖。圖13之上段所示之圖為玻璃帶之側視圖，圖^之下段左側所示之圖為對應於圖13之上段所示之側視圖之俯視圖又圖13之下段右側所示之圖表示拍攝所搬送之玻璃帶82内之1個缺陷83時所獲得之圖像。201233993 VI. Description of the Invention: [Technical Field] The present invention relates to a method for measuring defects in a glass ribbon and a method for measuring a defect in a glass ribbon, which are related to a defect in a glass ribbon, and more particularly to a method for measuring a glass ribbon A method for measuring defects in a broken glass ribbon such as a position in a height direction of a defect, and a method for measuring a defect in a glass ribbon. [Prior Art] Various methods for measuring the position of the height direction of the defect in the glass ribbon and the like have been proposed. As a general method for measuring the position in the height direction of a defect in a glass ribbon, there is a method of adjusting the focus of the camera to measure the height direction of the defect when shooting a defect. This method is called the first for convenience! test methods. Figure i〇A is an explanatory view schematically showing the first measurement method. In the third measurement method, as shown in Fig. 10A, the glass ribbon 82 is conveyed while passing the light through the glass ribbon 82, and the inside of the conveyed glass ribbon 82 is imaged by a line camera Gine camera 81. If there is a defect inside the glass ribbon 82, the defect 83 is photographed. Fig. 10B shows an example of an image of a defect photographed. In Fig. ia, the defect 83 is schematically represented by a rectangle, and the image 86 of the defect appearing in the image of the glass ribbon is also indicated by a rectangle in Wl〇B, but the shape of the defect is not limited to a rectangle. Here, the defects shown in the following figures nA, nB, UA, Fig. 12β, Fig. 13, and Fig. 14 are also schematically represented by a square. Further, the arrow shown in Fig. 1B is the conveyance direction of the glass ribbon. When the inside of the glass ribbon 82 is photographed by the line camera 81, the focus of the camera is adjusted so that the position of the defect coincides with the focus of the camera, and the absolute distance is determined from the line 160767.doc 201233993 array camera 8 1 to the defect. 'According to the distance, the position of the defect is calculated. As a method of adjusting the focus of the camera, the position of the defect and the focus of the camera - there are _ (to the handsome - F. (four) focus depth) method. Further, regarding the size of the defect, the captured image is subjected to image processing to measure the size of the defect. A method and apparatus for measuring the position of the singularity of the defect in the state of the camera are disclosed in, for example, Patent Documents 1 to 3. As another general method for measuring the position of the height direction of the defect in the glass ribbon, there is a method of photographing the same-defect at two positions by using the reflected light of the light incident on the glass ribbon, and obtaining two according to the result. The positional relationship like the position of the height of the defect. For the sake of convenience, the method is described as a method of determining the method. Fig. 1 i A schematically shows the circle of the second measurement method. In the second measurement method, for example, as shown in Fig. 11A, in the glass ribbon 82, light is incident on the same side as the line camera 81, and the reflected light reaches the line camera 81. Further, the glass ribbon 82 is conveyed, and a line camera is used. Take the inside of the glass ribbon μ. The path of the light in the glass ribbon will be described later with reference to the side view of the upper portion of the circle 13. The defect 83 moves as the glass ribbon 82 is moved, overlaps with the path of the light before reflection, and overlaps with the path of the reflected light, and is captured as an image in the line camera 81. As a result, even if the number of defects 83 is one, two images are reflected in the captured image. Fig. 11B is an example of an image taken in the second measurement method. As shown in Fig. ΐβ, two images 84, 85 are reflected for the same defect. In the second measurement method, the position of the defect in the height direction is calculated based on the positional relationship of the two images in the image illustrated in Fig. 11B. Further, regarding the size of the defect, the image taken was subjected to image processing at 160767.doc 201233993, and the size of the defect was measured. Further, the arrow shown in Fig. B is the conveying direction of the glass ribbon 82. A method and apparatus for measuring the position of the defect in the height direction based on the positional relationship between the two images by using the reflected light of the light incident on the transparent substrate or the like, and the like, are disclosed in, for example, Patent Documents 4 to 6, 8 and the like. . Further, there is a method of photographing an image on the both sides of the glass ribbon in the same manner as in the second measurement method, and measuring the position in the height direction of the defect based on the positional relationship of the image in the image captured on each side of the glass ribbon. This method is referred to as the third measurement method for the sake of convenience. Fig. 12A is an explanatory view schematically showing a third measurement method. In the third measurement method, for example, as shown in FIG. 12A, light is incident on the same side as the first line camera 8A in the glass ribbon 82, and the reflected light reaches the first line camera 81a. The second line camera 811 is incident on the same side as the second line camera 8 and the reflected light reaches the second line camera 811. ^ Further, the glass ribbon 82 is conveyed, and the glass ribbon 82 is taken by the first line camera and the second line camera 8ib, respectively. Then, in the first line camera 813, two images are captured in the same manner as in the case of the second measurement method. Further, two images are captured in the second line camera 8U. (3) An example of an image taken in the measurement method. In the third measurement method, as shown in Fig. 12B, an image taken from the upper side of the glass ribbon by a line camera and a line camera from the glass ribbon are obtained. The image captured on the side. Two images are respectively reflected in each image. In the third measurement method, the defect is calculated based on the positional relationship of the images in the images taken from the upper side and the lower side of the glass ribbon. The height direction position of 83. Furthermore, in Fig. 12B, the image overlaps in the image taken from the upper side. In the case of the size of the defect, the image of the captured image is imaged and measured. 160767.doc 201233993 The size of the defect is determined. Furthermore, the arrow shown in Fig. 2B is the direction in which the glass ribbon 82 is transported. A method of obtaining an image in the height direction of a defect on both sides of a substrate or the like is disclosed in, for example, Patent Document 7. In the second J丨疋 method and the third method, the image of the same defect is not in the image. The overlap condition is 'the height direction position of the defect. In the third measurement method, when the image overlaps in one image as illustrated in FIG. 12A, 'the height direction position of the defect is calculated by using another image. The following is a specific example of measuring the position in the height direction of the defect based on the positional relationship between the two images in the image captured by the second measurement method. Fig. 13 is a view showing the defect line in the conveyed glass ribbon. An illustration of the position of the camera when shooting. The figure shown in the upper part of Fig. 13 is a side view of the glass ribbon, and the figure shown on the left side of the lower part of Fig. 13 is a top view corresponding to the side view shown in the upper part of Fig. 13. 13 below the right The figure shown on the side shows an image obtained when one defect 83 in the conveyed glass ribbon 82 is taken.

在2個。In 2 pieces.

而，入射之光到達玻璃帶82中之與入到達線陣相機8 1之光自破搬送之玻璃帶82入射。繼與入射侧為相反側之界面 160767.docOn the other hand, the incident light reaches the glass ribbon 82 of the glass ribbon 82 which is incident on the glass ribbon 82 which has been transported to the line camera 81. Following the interface opposite the incident side 160767.doc

S 201233993 時，於該界面反射，通過入射側之界面而到達線 :卜到達線陣相機81之光之入射角_存於線陣相機81: 設置位置。藉由固定線陣相機81之設置八射角α你為固定值被決定。又’光之折射角Ρ依存於光之人射玻璃帶82之折射率„而定。此處，人㈣α及折射知’從而折射角β亦作為固定值被決^。關於折射率打、入射角α及折射角β，式〇)之關係成立。 n=sina/sinp 式（1) 因此，若入射角a及折射率n為已知，則藉由關於β來解開式（1)而求出折射角β。又，圖13所示之例中，自玻璃帶82中之與線陣相機"為相反側之面至缺陷83為止之高度方向位置d為測定對象。線陣相機81持續拍攝玻璃帶82之内部。缺陷以與玻璃帶 82—併沿搬送方向移動。而且，若缺陷趵移動至與入射至玻璃帶82並於界面反射後到達線陣相機8丨之光之路徑的最初之交又位置91，則線陣相機81拍攝第丨個像（以下記作第 1像）作為缺陷83之像。進而，若缺陷83移動至與光之路徑之第2次之交又位置92，則線陣相機81拍攝第2個像（以下記作第2像）作為缺陷83之像。其結果，如圖13之下段右側所示’所拍攝之圖像中出現第1像98及第2像99。再者，於缺陷83為透光性之情形時，穿透缺陷83之光到達線陣相機8 1而作為像被捕捉。於缺陷8 3為遮光性之缺陷之情形時’缺陷83作為黑色之像映於圖像上。缺陷83無論疋否為遮光性’於移動至位置91、92時均作為像被捕捉。 160767.doc 201233993 又，如圖13所不’將自第1像之拍攝位置91至第2像之拍攝位置92為止之缺陷83之移動距離設為〜。又，將線陣相機81之正面方向之拍攝位置之連線稱作中央線％。更具體而吕，將線陣相機81之正面方向之拍攝位置之連線正投影於玻璃帶82之界面而獲得之真料中央線％。根據於所拍攝之圖像（參照圖13之下段右側）中，將第丨像卯及第2像99 正投影於相當於中央線95之圖像内之線96時的像98、99之距離，可測定yd。若根據圖像’已測定出yd之值，則利用折射角P，藉由計算以下所示之式（2)，可求出缺陷83之高度方向位置d。 d=yd/(2-tanp) 式（2) 又，將自線陣相機81朝向第丨像之拍攝位置91之直線正投影於玻璃帶之界面之直線與t央線95所成之角設為θ。此時，於所拍攝之圖像（參照圖13之下段右側）中，通過第1 像98及第2像99之各中心、之直線與線96所成之角亦為θ。再者，此時，可以如下方式算出。以下，對圖13之下段左側之俯視圖所示之yc進行說明後，對tane之計算進行說明。〇圖13中，表示缺陷83自線陣相機81之正面偏移之情形。於如圖14所示，假定缺陷83存在於線陣相機^之正面之情形時’將第2像被拍攝之位置92正投影於玻璃㈣之界面之位置與線陣相機81之透鏡部分正投影於玻璃帶82之界面之位置的距離稱作拍攝距離ye。#中’拍攝距離^根據缺陷83之高度方向位η而變化。鸱最大時，拍攝距離為最 I60767.doc 201233993 小值y〗，d為最小時，拍攝距離yc為最大值^(參照圖14之上段所示之側視圖）。即，yi$yc^y2。如上所述，嚴格而言yc依存於d，但7。亦可例如於yisycgy2i範圍内預先決定。即便yc為不正確之值，只要為yiSyc$y22範圍之值，則tan0僅包含可忽視之誤差。又’將缺陷83自線陣相機之正面方向之偏移量記作 xcc(參照圖13之下段左側）。^。可根據於所拍攝之圖像（參照圖13之下段右側）中自相當於中央線95之線96至第2像99 為止之距離而特定。即，於圖像内，對相當於自線96至第 2像99為止之距離之像素數進行計數。由於線陣相機8丨之位置為固定，故而每1像素之實際空間中之距離亦作為固定值被決定。藉由相當於自線96至第2像99為止之距離之像素數乘以每1像素之實際空間中之距離，可算出、之長度。此處，tan0可利用yc&Xcc，如以下之式（3)所示，由近似式表示。即，tane可利用yc及Xce且藉由式（3)之計算而求出。 [數1] tan Θ —-—— « yd+yc y〇式(3) 又，專利文獻8中，揭示有一面使玻璃板移動„面使光入射至玻璃板，以該入射光及反射光檢測缺陷，且對缺陷之兩度方向位置進行運算之方法。專利文獻8所揭示之方法中，於檢測缺陷之圖案之情形時，在玻璃板之移動方向 160767.doc 201233993 上無大致同一大小之圖案時，即，在玻璃板之背面附近存在缺陷時或缺陷較大時，將該缺陷之高度方向位置判定為 0。因此’專利文獻8所揭示之方法中，於上述情形時無法正確地求出缺陷之高度方向之位置。先前技術文獻專利文獻專利文獻1:曰本專利特開2001-305072號公報專利文獻2 :日本專利特開2004-361384號公報專利文獻3 :曰本專利特開2008-76071號公報專利文獻4:曰本專利第2920056號公報專利文獻5 :曰本專利特開平9-61139號公報專利文獻6:曰本專利特表2003-508786號公報專利文獻7 :國際公開第2006/057 125號專利文獻8:曰本專利特開2〇1〇_8177號公報【發明内容】發明所欲解決之問題上述第1測定方法中，藉由調節相機之焦點，算出自相機至缺陷為止之絕對距離，而求出玻璃帶内之缺陷之高度方向位置。然而’玻璃帶有時於搬送中上下振動。因此，第1測疋方法中，存在因玻璃帶之上下振動而缺陷之高度方向位置之測定結果易產生誤差之問題。第2測定方法及第3測定方法中，以所拍攝之圖像内2個像不重疊為條件，計算缺陷之高度方向位置。因此，若如圖11B所示，圖像内2個像84、85不重疊，則可計算缺陷之 160767.doc 201233993 面度方向位置。然而’於缺愔幵”姐w 、缺陷存在於破螭帶之界面附近之個:;Γ之情形時’同-缺陷之2個像會重疊。若2 像重邊’則於第2測定方法及第3測定缺陷之高度方向位置。 ‘…凌Τ异再者’第3測Μ法中’於缺陷存在於玻璃帶之界面附近：情形時，如圖12Β所示’由與該界面為相反側之線陣㈣㈣攝之圖像中2個像重疊’而由存在缺陷之側之線陣相機所拍攝之圖像中2個圖像未重叠。於該情形時，根據2個圖像中像未重疊之圓像，可計算缺陷之高度方向位置。然而’即便於第3敎方法中，在缺陷較大之情形時，亦有時由2個線陣相機81a、81b(參照圖12Α)所拍攝之各圖像令均發生2個像重4 4該情料，無法計算缺陷之高度方向位置。因此，本發明之目的在於提供一種即便於玻璃帶之界面附近存在缺陷之情形或缺陷較大之情形時，亦可測定玻璃帶内之缺陷之高度方向位置的玻璃帶内缺陷測定方法及玻璃帶内缺陷測定系統。解決問題之技術手段本發明之玻璃帶内缺陷測定方法之特徵在於包括：拍攝步驟’其係自光源（例如為光源2)對所搬送之玻璃帶（例如為玻璃帶5)照射光，藉由配置於以玻璃帶反射之光所到達之位置上的拍攝機構（例如為線陣相機3)拍攝玻璃帶；及運算步驟’其係根據由拍攝機構所拍攝之圖像内之起因於玻 160767.doc 201233993 璃帶之同一缺陷之2個重疊的像且該缺陷之種類所固有之形狀之2個像的位置關係，算出玻璃帶内之缺陷之高度方向位置。亦可為如下方法：於運算步驟中，計算自起因於同一缺陷之2個重疊之像（例如為像21、22)中之一個像的拍攝位置至另個像之拍攝位置為止之缺陷的移動距離（例如為 y<〇,藉由計算出之移動距離、及玻璃帶内之光之折射角，而算出破璃帶内之缺陷之高度方向位置。亦可為如下方法：於運算步驟中，藉由自起因於同一缺陷之2個重疊之像（例如為像21、22)之外接矩形中的與相當於^帶之搬送方向之方向平行之邊的像素數所對應之實際空間中之長度(例如為h)中，減去與搬送方向平行之缺陷直拴之長度（例如為s)，而算出移動距離。亦可為如下方法：於運算步驟卜根據起因於同一缺陷之2個重疊之像之位置關係，利用包含玻璃帶之寬度方向之像之位置作為變數（例如為變數u)的預先決定之算式 ^如為式（4)或式（5))，算出缺陷之特徵值（例如為s或^，利用該特徵值，算出移動距離。阶、可為如下方法··特徵值為與玻璃帶之搬送方向平行之矩^徑之長度（例如為…藉由自2個重疊之像之外接矩形中的與相營切4 + , Μ ' ' 。之方向平行之邊的像素數所對應之實際空間中之動距離。 X 減去该直徑之長度，而算出移亦可為如下方法特徵值為缺陷之2個直徑之比（例如為 160767.docIn S 201233993, it is reflected at the interface, and reaches the line through the interface on the incident side: the incident angle of the light reaching the line camera 81 is stored in the line camera 81: the set position. By setting the line camera 81, the angle of incidence α is determined for a fixed value. Moreover, the angle of refraction of light depends on the refractive index of the light-emitting glass belt 82 of the light. Here, the human (four) α and the refractive knowledge are known, and thus the refraction angle β is also determined as a fixed value. The relationship between the angle α and the refraction angle β, 〇) holds. n=sina/sinp (1) Therefore, if the incident angle a and the refractive index n are known, the equation (1) is solved by β. Further, in the example shown in Fig. 13, the position d in the height direction from the surface of the glass ribbon 82 opposite to the line camera "to the defect 83" is the measurement target. The inside of the glass ribbon 82 is continuously photographed. The defect moves with the glass ribbon 82 and in the transport direction. Further, if the defect 趵 is moved to the path of the light incident on the glass ribbon 82 and reflected by the interface and reaching the line camera 8 When the initial position is 91, the line camera 81 captures the next image (hereinafter referred to as the first image) as the image of the defect 83. Further, if the defect 83 moves to the second intersection with the path of the light, 92, the line camera 81 captures a second image (hereinafter referred to as a second image) as an image of the defect 83. As a result, the first image 98 and the second image 99 appear in the image taken as shown on the right side of the lower part of Fig. 13. Further, when the defect 83 is translucent, the light reaching the defect 83 reaches the line. The camera 8 is captured as an image. When the defect 83 is a defect of light blocking, the defect 83 is reflected as a black image on the image. The defect 83 is moved to the position 91 regardless of whether it is a light blocking property. In the case of 92, the image is captured as an image. 160767.doc 201233993 Moreover, as shown in FIG. 13, the moving distance of the defect 83 from the imaging position 91 of the first image to the imaging position 92 of the second image is set to ~. The line connecting the photographing positions in the front direction of the line camera 81 is referred to as a center line %. More specifically, the line connecting the photographing positions in the front direction of the line camera 81 is orthographically projected on the interface of the glass ribbon 82. The center line %. According to the image taken (see the right side of the lower part of FIG. 13), when the first image 第 and the second image 99 are projected onto the line 96 corresponding to the image of the center line 95, For distances of 98 and 99, yd can be measured. If the value of yd has been determined based on the image, the fold is used. In the angle of incidence P, the position d in the height direction of the defect 83 can be obtained by calculating the equation (2) shown below. d=yd/(2-tanp) Equation (2) Further, the line camera 81 is directed toward the The angle between the straight line of the image capturing position 91 and the straight line projected at the interface of the glass ribbon and the center line 95 is set to θ. At this time, in the captured image (refer to the right side of the lower part of Fig. 13), The angle between the straight line of each of the first image 98 and the second image 99 and the line 96 is also θ. Further, in this case, it can be calculated as follows. Hereinafter, the top view on the left side of the lower part of Fig. 13 is shown. After yc is explained, the calculation of tane will be described. 〇 In Fig. 13, the case where the defect 83 is shifted from the front side of the line camera 81 is shown. As shown in FIG. 14, assuming that the defect 83 is present on the front side of the line camera ^, the position where the position where the second image is captured is projected to the position of the interface of the glass (4) and the lens portion of the line camera 81 is orthographically projected. The distance from the position of the interface of the glass ribbon 82 is referred to as the shooting distance ye. The #中' shooting distance ^ changes according to the height direction bit η of the defect 83. When 鸱 is the maximum, the shooting distance is the most I60767.doc 201233993 small value y, and when d is the minimum, the shooting distance yc is the maximum value ^ (refer to the side view shown in the upper part of Fig. 14). That is, yi$yc^y2. As mentioned above, strictly speaking, yc depends on d, but 7. It can also be predetermined in advance, for example, within the scope of yisycgy2i. Even if yc is an incorrect value, as long as it is a value in the range of yiSyc$y22, tan0 contains only errors that can be ignored. Further, the offset of the defect 83 from the front direction of the line camera is referred to as xcc (refer to the left side of the lower portion of Fig. 13). ^. It can be specified based on the distance from the line 96 corresponding to the center line 95 to the second image 99 in the captured image (refer to the right side of the lower portion of Fig. 13). That is, the number of pixels corresponding to the distance from the line 96 to the second image 99 is counted in the image. Since the position of the line camera 8 is fixed, the distance in the actual space per pixel is also determined as a fixed value. The length can be calculated by multiplying the number of pixels corresponding to the distance from the line 96 to the second image 99 by the distance in the actual space per pixel. Here, tan0 can be represented by an approximate expression using yc & Xcc as shown in the following formula (3). That is, tane can be obtained by calculation of the equation (3) using yc and Xce. [Number 1] tan Θ ———— « yd+yc y 〇 (3) Further, in Patent Document 8, it is disclosed that one side of the glass plate is moved to face the light to be incident on the glass plate, and the incident light and the reflected light are used. A method of detecting a defect and calculating the position of the defect in two directions. In the method disclosed in Patent Document 8, when detecting the pattern of the defect, there is no substantially the same size in the moving direction of the glass plate 160767.doc 201233993. In the case of a pattern, that is, when there is a defect near the back surface of the glass sheet or when the defect is large, the height direction position of the defect is determined to be 0. Therefore, in the method disclosed in Patent Document 8, the above situation cannot be accurately obtained. The position of the height direction of the defect is disclosed. Patent Document 1: Patent Document No. 2001-305072 Patent Document 2: Japanese Patent Laid-Open No. 2004-361384 Patent Document 3: Japanese Patent Application Publication No. 2008 Japanese Laid-Open Patent Publication No. Hei. No. Hei. No. Hei. No. Hei. No. Hei. No. Hei. No. Hei. No. Hei. [Patent Document No. 2006/057125] Japanese Patent Application Laid-Open No. Hei. No. Hei. No. Hei. The absolute distance from the camera to the defect is obtained, and the height direction position of the defect in the glass ribbon is obtained. However, when the glass is carried, the glass vibrates up and down during the conveyance. Therefore, in the first measurement method, there is vibration due to the upper and lower sides of the glass ribbon. The measurement result in the height direction of the defect is liable to cause an error. In the second measurement method and the third measurement method, the height direction position of the defect is calculated on the condition that the two images in the captured image are not overlapped. As shown in FIG. 11B, if the two images 84 and 85 in the image do not overlap, the position of the defect 160767.doc 201233993 can be calculated. However, the defect is present in the broken zone. The one near the interface:; in the case of Γ, the same image of the same-defect will overlap. If the 2 image is heavy, the position is in the height direction of the second measurement method and the third measurement defect. '... Τ Τ 再 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' Two images in the image taken by the line camera with the side of the defect overlap are not overlapped. In this case, the position of the height direction of the defect can be calculated based on the image of the two images that do not overlap. However, even in the case of the third method, in the case where the defect is large, sometimes two image weights 4 4 are generated by the image sequences of the two line cameras 81a and 81b (refer to FIG. 12A). In this case, the position of the height direction of the defect cannot be calculated. Therefore, an object of the present invention is to provide a method for measuring defects in a glass ribbon and a glass ribbon which can measure the position in the height direction of a defect in the glass ribbon even when there is a defect in the vicinity of the interface of the glass ribbon or a large defect. Internal defect measurement system. MEANS FOR SOLVING THE PROBLEMS The method for measuring defects in a glass ribbon according to the present invention is characterized in that it includes a photographing step of irradiating light to a conveyed glass ribbon (for example, a glass ribbon 5) from a light source (for example, a light source 2). The photographing mechanism (for example, the line camera 3) disposed at a position where the light reflected by the glass ribbon reaches is photographed with a glass ribbon; and the operation step 'is based on the glass 160767 in the image taken by the photographing mechanism. Doc 201233993 The positional relationship between the two images of the same defect of the same defect and the shape of the defect is calculated in the height direction of the defect in the glass ribbon. Alternatively, in the calculation step, the movement of the defect from the shooting position of one of the two overlapping images (for example, 21, 22) of the same defect to the shooting position of the other image may be calculated. The distance (for example, y<〇, by calculating the moving distance and the angle of refraction of the light in the glass ribbon, the height direction position of the defect in the broken glass ribbon is calculated. The method may also be as follows: The length in the real space corresponding to the number of pixels in the rectangle parallel to the direction parallel to the direction of the transport direction of the strip by the two overlapping images (for example, 21, 22) due to the same defect (For example, in h), the length of the defect straight parallel to the transport direction (for example, s) is subtracted, and the moving distance is calculated. It may also be a method in which the operation step is based on two overlaps caused by the same defect. In the positional relationship of the image, the predetermined value of the image including the position of the image in the width direction of the glass ribbon (for example, the variable u) is calculated as the equation (4) or the equation (5)), and the characteristic value of the defect is calculated (for example, For s or ^, profit Using the characteristic value, the moving distance is calculated. The order may be a method in which the characteristic value is a length parallel to the conveying direction of the glass ribbon (for example, by connecting the rectangle from the two overlapping images) The moving distance in the real space corresponding to the number of pixels parallel to the direction of the phase cut 4 + , Μ ' '. X minus the length of the diameter, and the calculated shift can be the following method. 2 diameter ratios (for example, 160767.doc

S 12- 201233993 r)，藉由相當於拍攝機構之正面方向之拍攝位置的圖像内 =線與通過2個像之各中心之線所成之角、及上述比，而异出移動距離。又本發明之玻璃帶内缺陷測定系統之特徵在於包含：搬送機構（例如為搬送輥1)，其搬送缺陷之高度方向位置成為測定對象之玻璃帶；光源（例如為光源2)，其對玻璃帶照射光；拍攝機構（例如為線陣相機3)，其配置於以玻璃帶反射之光所到達之位置上’拍攝玻璃帶；及運算機構(例如為運算裝置4)，其根據由拍攝機構所拍攝之圖像内之起因於玻璃帶之同一缺陷之2個重疊的像且缺陷之種類所固有之形狀之2個像的位置關係，算出玻璃帶内之缺陷之高度方向位置。亦可為如下構成：運算機構計算自起因於同一缺陷之2 個重疊之像中之-個像的拍攝位置至另—個像之拍攝位置為止之缺陷的移動距離（例如為…），藉由計算出之移動距離、及玻璃帶内之光之折射角，巾算出玻璃帶内之缺陷之兩度方向位置。亦可為如下構成：運算機構藉由自起因於同一缺陷之2 個重疊之像之外接矩形中的與相當於玻璃帶之搬送方向之方向平行之邊的像素數所對應之實際空間中之長度（例如為h)中，減去與搬送方向平行之缺陷之直徑之長度（例如為 S)’而算出移動距離。亦可為如下構成：運算機構根據起因於同—缺陷之2個重疊之像之位置關係，利用包含玻璃帶之寬度方向上之像 I60767.doc 13 201233993 之位置作為變數(例如為變數U)的預先決定之算式(例如為 5弋（））算出缺陷之特徵值（例如為s或r)，利用該特徵值，算出移動距離。發明之效果根據本發明，即便於玻璃帶之界面附近存在缺陷之情形或缺陷較大之情形時，亦可測定玻璃帶内之缺陷之高度方· 向位置β - 【實施方式】以下，參照圖式對本發明之實施形態進行說明。本發明中’成為測定玻璃帶内之高度方向位置之對象之缺陷的種類為已知。又’拍攝該缺陷而獲得之圖像内之像設為包含特徵性之點（以下記作特徵點）之一定形狀之像。換言之，本申請案中之成為高度方向位置之測量對象之缺陷為滿足作為包含特徵點之一定形狀之像被拍攝之條件的缺陷。作為如上所述之缺陷之例，可舉出玻璃帶内之氣泡。氣泡於玻璃帶内為橢圓體。而且’氣泡作為中心為白色之橢圓形之像而映於圖像上，可將中心之白色部分作為特徵點而使用以下之實施形態中，以缺陷為氣泡之情形為例進行說明。 .S 12-201233993 r) The movement distance is different by the angle formed by the line in the image corresponding to the shooting position in the front direction of the photographing mechanism and the line passing through the centers of the two images, and the above ratio. Further, the glass ribbon inner defect measuring system according to the present invention includes: a conveying mechanism (for example, a conveying roller 1) that is a glass ribbon to be measured in a height direction of a conveying defect; a light source (for example, a light source 2), which is a pair of glass a illuminating light; a photographing mechanism (for example, a line camera 3) disposed at a position where the light reflected by the glass ribbon reaches; a photographing glass ribbon; and an arithmetic mechanism (for example, an arithmetic unit 4) according to the photographing mechanism In the captured image, the positional relationship between the two images of the shape of the same defect of the glass ribbon and the shape of the defect is determined, and the position in the height direction of the defect in the glass ribbon is calculated. The calculation unit may calculate a moving distance (for example, ...) of a defect from a shooting position of one of the two overlapping images caused by the same defect to a shooting position of another image. Calculate the moving distance and the angle of refraction of the light in the glass ribbon, and calculate the position of the defect in the glass ribbon in two directions. Alternatively, the arithmetic unit may be configured to calculate the length of the real space corresponding to the number of pixels in the rectangle parallel to the direction corresponding to the direction in which the glass ribbon is transported by the two overlapping images of the same defect. (For example, in h), the moving distance is calculated by subtracting the length (for example, S) of the diameter of the defect parallel to the conveying direction. Alternatively, the arithmetic unit may use a position including an image of the image I60767.doc 13 201233993 in the width direction of the glass ribbon as a variable (for example, a variable U) based on a positional relationship between two overlapping images of the same defect. The predetermined value (for example, 5 弋()) is used to calculate a feature value of the defect (for example, s or r), and the moving distance is calculated using the feature value. Advantageous Effects of Invention According to the present invention, even when there is a defect or a defect in the vicinity of the interface of the glass ribbon, the height of the defect in the glass ribbon and the position β can be measured. [Embodiment] Hereinafter, reference is made to the drawings. The embodiment of the present invention will be described. In the present invention, the type of defect which is a target for measuring the position in the height direction in the glass ribbon is known. Further, the image in the image obtained by capturing the defect is an image having a certain shape including a characteristic point (hereinafter referred to as a feature point). In other words, the defect of the measurement target in the height direction position in the present application is a defect that satisfies the condition that the image of a certain shape including the feature point is photographed. As an example of the defects as described above, bubbles in the glass ribbon can be cited. The bubbles are ellipsoidal inside the glass ribbon. Further, the 'bubble is reflected on the image as a white elliptical image at the center, and the white portion of the center can be used as a feature point. In the following embodiment, the case where the defect is a bubble will be described as an example. .

[實施形態1] . 圖1係表示本發明之玻璃帶内缺陷測定系統之構成例之模式圖。本發明之玻璃帶内缺陷測定系統包含搬送輥i、光源2、線陣相機3、及運算裝置4。搬送輥1支持玻璃帶5，且沿固定方向以固定速度搬送玻 160767.doc ·14·[Embodiment 1] Fig. 1 is a schematic view showing a configuration example of a glass ribbon inner defect measuring system of the present invention. The glass ribbon inner defect measuring system of the present invention includes a conveying roller i, a light source 2, a line camera 3, and an arithmetic unit 4. The conveying roller 1 supports the glass ribbon 5 and conveys the glass at a fixed speed in a fixed direction. 160767.doc ·14·

S 201233993 璃帶5。再者，作為玻璃帶5之製造方法，可舉出例如浮式法。玻璃帶5沿製造時之主要之延伸方向由搬送輥1搬送。玻璃帶之主要之延伸方向並非指利用導引構件之向玻璃帶之寬度方向之延伸，而是指沿玻璃帶之行進方向之延伸之方向。以f，將玻璃帶之主要之延伸方向僅記作玻璃帶之 I伸方向又，本發明中，測定自玻璃帶5中之搬送輥丨側之表面至缺陷（氣泡）為止之高度方向位置（距離）。光源2配置於玻璃帶5之兩面中之一面之側，且朝向玻璃帶5照射光。該光自界面8入射至玻璃帶5,通過玻璃帶内’於與人射側為相反側之界面9反射^反射之光通過入射側之界面8而到達線陣相機3。再者，圖i中將光之路徑簡化而表示，但如圖13之上段之側視圖所示，光之路徑於光入射至界面8時及於界面9反射後通過界面叫分別會進行折射。線陣相機3配置於自光源2照射並於玻璃^反射之光所到達之位置上。具體而言’以玻璃帶5為基準，配置於與光源2同一側。又’例如線陣相機3以光源2為基準，配置於玻璃帶5之搬送方向上。而且’線陣相機3拍攝玻璃帶5 之内部，且生成圖像作為拍攝結果。由於光源2及線陣相機3之配置位置已決定，而於光之路徑上入射角參照圖13之上段)亦作為固定值被決定。進而，玻璃帶5之折射率η亦為已知，藉由解出式⑴，而自光源2至線陣相機3為止之光之路徑中之折射角β之值亦作為固定值被決定。 ρ^ f ^ 160767.doc •15- 201233993 玻璃帶5被搬送，且線陣相機3於固定位置上持續進行玻璃帶5之拍攝。因此’隨著時間經過，玻璃帶5中被拍攝之部位發生變化。於是，將線陣相機3之正面方向之拍攝位置之連線在正投影於玻璃帶5之界面8時表示為直線。將該直線稱作中央線。圖2A係表示中央線之說明圖，圖叫系表不圖像内之相當於中央線之線之說明圖。㈣係玻璃帶 5之俯視圖。隨著玻璃帶5之搬送’線陣相機3之正面之拍攝位置發生變化，將其連線向界面之正投影圖示為中央線 …又，圖2B表示藉由線陣相機3所拍攝之圓像。於圖像内，由單點劃線表示相當於中央線95之線％。該線％可稱為與線陣相機3之正面方向之拍攝位置對應之像素的連線。又，中央線95與玻璃帶5之搬送方向平行，從而可稱為相當於中央線95之圖像内之線96表示圖像内之相當於玻璃帶5之搬送方向之方向。將相當於中央線95之圖像内之線96記作搬送方向線。再者’圖㈣，為進行說明而圖示有搬送方向線96，但實際夕杯媒固 1一頁u之拍攝圖像中，搬送方向線96並不會映於圊像内。於玻璃帶5内存在缺陷（本例中為氣泡）之情形時，起因於1個缺陷，在線陣相機3所拍攝之圖像内出現2個該缺陷之像。又，本例中，由於缺陷為氣泡，故而圖像内出現之像為橢圓形，其中心部為白色。運算裝置4參照藉由線陣相機3所拍攝之圖像，測定缺陷之高度方向位置。該缺陷之高度方向位置為圓13之上段之側視圖中表示為「d」之長度。_，玻璃帶5中，自與光源 160767.docS 201233993 Ribbon 5. Further, as a method of producing the glass ribbon 5, for example, a floating method can be mentioned. The glass ribbon 5 is conveyed by the conveyance roller 1 in the main extension direction at the time of manufacture. The main direction of extension of the glass ribbon does not refer to the extension of the direction of the width of the glass ribbon by the guiding member, but rather the direction in which the direction of travel of the glass ribbon extends. In f, the main direction of extension of the glass ribbon is simply referred to as the direction in which the glass ribbon is stretched. In the present invention, the position in the height direction from the surface on the side of the transport roller to the defect (bubble) in the glass ribbon 5 is measured ( distance). The light source 2 is disposed on one side of one of the two faces of the glass ribbon 5, and is irradiated with light toward the glass ribbon 5. This light enters the glass ribbon 5 from the interface 8, and the light reflected by the interface 9 on the opposite side to the human side in the glass ribbon passes through the interface 8 on the incident side to reach the line camera 3. Further, in Fig. i, the path of the light is simplified, but as shown in the side view of the upper part of Fig. 13, the path of the light is refracted by the interface when the light is incident on the interface 8 and after being reflected by the interface 9. The line camera 3 is disposed at a position where the light reflected from the light source 2 reaches the light reflected by the glass. Specifically, the glass ribbon 5 is disposed on the same side as the light source 2. Further, for example, the line camera 3 is disposed in the transport direction of the glass ribbon 5 with reference to the light source 2. Further, the line camera 3 photographs the inside of the glass ribbon 5, and an image is generated as a result of the shooting. Since the arrangement positions of the light source 2 and the line camera 3 have been determined, and the incident angle on the light path is referred to the upper portion of Fig. 13, it is also determined as a fixed value. Further, the refractive index η of the glass ribbon 5 is also known, and by solving the equation (1), the value of the refraction angle β in the path of the light from the light source 2 to the line camera 3 is also determined as a fixed value. ρ^ f ^ 160767.doc •15- 201233993 The glass ribbon 5 is conveyed, and the line camera 3 continues the shooting of the glass ribbon 5 at a fixed position. Therefore, as time passes, the portion of the glass ribbon 5 that is photographed changes. Then, the line connecting the shooting positions in the front direction of the line camera 3 is indicated as a straight line when it is projected onto the interface 8 of the glass ribbon 5. This line is called the center line. Fig. 2A is an explanatory view showing a center line, and is a diagram for explaining a line corresponding to a center line in an image. (4) A plan view of the glass ribbon 5 . As the position of the front side of the line camera 3 is changed as the glass belt 5 is moved, the orthographic projection of the line to the interface is shown as a center line... Again, FIG. 2B shows the circle taken by the line camera 3. image. Within the image, the line % corresponding to the center line 95 is indicated by a one-dot chain line. This line % can be referred to as a line connecting pixels corresponding to the shooting position in the front direction of the line camera 3. Further, the center line 95 is parallel to the conveyance direction of the glass ribbon 5, and the line 96 in the image corresponding to the center line 95 indicates the direction in the image corresponding to the conveyance direction of the glass ribbon 5. A line 96 corresponding to the image of the center line 95 is referred to as a transport direction line. Further, in the figure (4), the conveyance direction line 96 is shown for the sake of explanation, but in the captured image of the actual one-cup medium, the conveyance direction line 96 is not reflected in the image. When there is a defect (bubble in this example) in the glass ribbon 5, two defects are present in the image taken by the line camera 3 due to one defect. Further, in this example, since the defect is a bubble, the image appearing in the image is elliptical, and the center portion thereof is white. The arithmetic unit 4 refers to the image captured by the line camera 3, and measures the position in the height direction of the defect. The height direction of the defect is the length indicated as "d" in the side view of the upper portion of the circle 13. _, glass belt 5, self and light source 160767.doc

S -16 - 201233993 2為相反側之界面9至缺陷為止之距離。運算裝置4於拍攝共同之缺陷而獲得之成對之像重疊之情形時，根據缺陷之種類（本例中為氣泡）所固有之形狀之像（即，橢圓形之像）的位置關係，算出玻璃帶5内中之缺陷之高度方向位置。具體而έ ’運算裝置4計算下述值：自圖像内2個重疊之像之外接矩形中的與相當於玻璃帶之搬送方向之方向平行之邊的像素數所對應之實際空間中之距離中，減去缺陷（氣泡）之直徑中與搬送方向平行之直徑之長度。再者，於圖像内，與相當於玻璃帶之搬送方向之方向平行係指與搬送方向線96(參照圖2Β)平行。運算裝置4藉由以上述減法運算所求出之值、及玻璃帶5中之折射角β，而算出缺陷之高度方向位置。關於該計算，之後參照圖5進行敍述。又，玻璃帶内之氣泡之長軸與搬送輥丨之搬送方向（換言之為玻璃帶5之延伸方向）大致平行。如圖3所示，氣泡之長軸72之方向與利用搬送輥！之玻璃帶5之搬送方向71的偏移量最大為10。。如上所述，由於氣泡之長軸72與搬送輥1 之搬送方向71為大致平行，故而於線陣相機3所拍攝之圖像中，表現為橢圓形之缺陷之像之長軸與搬送方向線 96(參照圖2B)亦大致平行。以下’以所拍攝之圖像中缺陷之像之長軸與搬送方向線96平行之情形為例進行說明。再者’於成對之像不f疊之情形時，運算裝置4藉由公知之方法而算出缺陷之高度方向位置即可。又，線陣相機3之配置位置被固定。因此，與線陣相機3 所拍攝之圖像中之m素對應之實際空間中的距離亦作為 160767.doc 201233993 固定值被決定。設為對應於圖像中之丨像素之實際空間中的距離為已知。其次，對動作進行說明。圖4係表示本實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流程圖。首先，光源2對玻璃帶5開始光之照射（步驟$ 1)。繼而，搬送輥1沿固定方向搬送配置於搬送輥丨上之玻璃帶5，線陣相機3持續拍攝所搬送之玻璃帶$之内部。而且，線陣相機3生成圖像作為拍攝結果（步輝S2)。線陣相機3將藉由拍攝所獲得之圖像發送至運算裝置4。於玻璃帶5之内部存在缺陷之情形時，在步驟“中所獲得之圖像中包含缺陷之像。本例中，由於缺陷為氣泡故而於圖像内映出橢圓形之像。又，如圖13中所說明，於缺陷移動至與反射前之光之路徑重疊之位置（圖13之上段之側視圖所示之位置91)時、及缺陷移動至與反射後之光之路徑重疊之位置（圖1 3之上段之側視圖所示之位置92)時，分別作為像被映於圖像上。因此，於存在1個缺陷之情形時，圖像上映出2個像》又，於缺陷較大之情形或缺陷存在於玻璃帶5之界面9(參照圖1)之附近之情形時，該2個像發生重疊。運算裝置4若接收步驟S2中所生成之圖像，則自圖像中檢測2個重疊之像之外接矩形之區域。而且，對該外接矩形之邊中之於圖像内與相當於玻璃帶之搬送方向之方向平行之邊（即，與圖像内之搬送方向線平行之邊）的像素數進行計數。繼而，運算裝置4藉由將該邊之像素數乘以每1像 I60767.doc .18- 201233993 素之實際空間中之距離，而算出對應於該邊之像素數之實際空間中之長度（步驟S3)。圖5係表示2個重疊之像之外接矩形之區域的說明圖。如圖5所示，作為重疊之2個像21、22之外接矩形，決定為圖 5所示之外接矩形23。像21、22為橢圓，且可看作全等。圖5所示之例中’設為外接矩形23之長邊與搬送方向線（參照圖2B)平行。於該情形時，運算裝置4對像21、22之外接矩形23之長邊24之像素數進行計數，將該像素數乘以每j 像素之實際空間中之距離。將對應於該長邊24之實際空間中之長度以「h」表示。h之單位例如為μπι。又於缺陷為氣泡之情形時，像21之中心部2 1 a於圖像上為白色。該中心部21a為像21之特徵點。運算裝置4對自一個像21之中心部21a至外接矩形23之短邊中距其較近一方之短邊為止之像素數進行計數。即，對圖5中符號A所示之部分之像素數進行計數。運算裝置4將該像素數乘以每】像素之實際空間中之距離。該乘法運算結果為相當於圖5 所示之A之部分所對應之實際空間中之長度，具體而言，為與搬送方向平行之缺陷之直徑（缺陷之直徑中與搬送方向平行之直徑）之1/2長度所示之例中，該直徑為缺陷之長轴•運算裝置4藉由將上述乘法運算結果乘以2,而算出與搬送方向平行之缺陷之直徑之長度（步驟S4)。將該缺陷之直徑之長度設為s。s之單位例如為。與實際空間中之s/2之長度對應之圖像内之部位為圖5中符號a所示之部分。又’由於2個像21、22可看作全等，故而圖^，可看 160767.doc 201233993 作 a=a'。 /再者，此處以利用像21之令心部2“計算s之情形订說明，但亦可利用像22之中心部計算s。又’圖5中’以缺陷之像之絲與搬送方向線平行之形為例進行說明，但亦有缺陷之像之長抽與搬送方向線不月完全平行之情形。然而’玻璃帶内之氣泡之長抽與玻璃帶之搬送方向之偏移量最大僅為1〇。（參照圖3)。因此，即便缺陷之像之長轴與搬送方向線不完全平行’亦可看作兩者平行，且與上述步驟S3、84同樣地計算h' s。即，求^ 時’只要對重叠之2個像之外接矩形之長邊之像素數進行計數，將該像素數乘以每丨像素之實際空間中之距離即可。又，求s時，只要對自一個像之中心部至外接矩形之短邊中距其較近-方之短邊為止之像素數進行特，將該像素數乘以每1像素之實際空間中之距離，且將該乘法運算結果乘以2即可。即便缺陷之像之長軸與搬送方向線不完全平行，如上述般計算h、s，且利用該h、s計算缺陷之冋度方向位置時亦僅包含可忽視之程度之誤差。其-人，運算裝置4自步驟S3中所算出之h中減去步驟§4中所算出之s(步驟S5)。將該減法運算結果設為…。％為自第 1個像被拍攝之位置至第2個像被拍攝之位置為止之缺陷之移動距離。即，步驟S5中所算出之yd為缺陷之像被拍攝之 2點間之距離。再者，對應於實際空間中之之長度的圖像内之部位為圖5中符號B所示之部分。運算裝置4利用步驟S5中所算出之yd、及預先決定之折 160767.doc 201233993 射角β，進行式⑺之計算，而言十算缺陷之高度方向位置“ 即，計算yd/(2.t_)，將其計算結果作為d(步驟％)。缺陷之高度方向位置d為自玻璃帶5之界面9(參照圖至缺陷止之距離。㈢根據本實施形態，即便起因於同一缺陷之2個像重疊，亦可測定該缺陷之高度方向位置。因&，即便於玻璃帶之界面附近存在缺陷之情形或缺陷較大之情形時，亦可測定玻璃帶内之缺陷之高度方向位置。又，步驟S4中所算出之s為缺陷之長軸之長度。又，即便於缺陷之像之長軸與搬送方向線不完全平行之情形時，將步驟S4中所算出之s看作缺陷之長軸之長度時亦僅包含可忽視之程度之誤差。因此，亦可算出缺陷之大小（長轴之長度）。又，根據本實施形態，於玻璃帶5之單側配置光源2及線陣相機3即可。因此，與第3測定方法（參照圖12A)相比’ 可減少光源2及線陣相機3之設置數，從而可降低測定所需之成本》 [實施形態2] 本發明之第2實施形態與第1實施形態同樣地包含搬送輥 1、光源2、線陣相機3、及運算裝置4(參照圖丨）。光源2及線陣相機3相對於玻璃帶5之位置關係與第1實施形態相同，而省略說明。第2實施形態中，利用運算裝置4之缺陷之高度方向位置之測定方法與第1實施形態不同。第2實施形態中，運算裝置4算出玻璃帶$内之缺陷之特 160767.doc 201233993 徵值。繼而’運算裝置4利用該特徵值，計算下述值：自重疊之2個像之外接矩形中的與相當於玻璃帶之搬送方向之方向平行之邊的像素數所對應之實際空間中之長度中，減去與玻璃帶之搬送方向平行之缺陷之直徑（缺陷之直徑中與搬送方向平行之直徑）之長度。又，運算裝置4於上述特徵值時，根據重疊之2個像之位置關係，使用預先決定之算式計算特徵值。又，第2實施形態中，作為特徵值，計算與玻璃帶之搬送方向平行之缺陷之直徑的長度。用以算出上述特徵值之式作為下述函數被預先決定，該函數將以玻璃帶之端部為基準之像之特徵點所對應之位置的座標、第1實施形態中所說明之h、及2個重疊之像之面積設為變數。用以決定該特徵值（缺陷之直徑中與搬送方向平行之直徑）之算式能夠以例如以下之式（4)表示。 ssay+ay+ay+aeh+ashp+awp+ayu+ash+agp+a]。式⑷ 式（4)中，「u」為以玻璃帶之端部為基準之像之特徵點所對應之位置的座標，具體而言，為自與搬送方向平行之玻璃帶之侧面至缺陷之中心為止之距離。此處，u之單位設為mm e「h」為根據拍攝到缺陷之圖像，藉由與第丨實施形態中之步驟S3相同之計算而獲得之值。此處，h之單位設為μπι。p為拍攝到缺陷之圖像中2個像所佔區域（2個像之區域之並集）之面積，具體而言，以圖像内之像素數表示式（4)中之ai〜aio為係數。又，式（4)中之s為與玻璃帶之搬送方向平行之缺陷之直徑。由於成為特徵值之直徑s 160767.docS -16 - 201233993 2 is the distance from the interface 9 on the opposite side to the defect. When the paired images obtained by the arithmetic device 4 are captured by the common defect, the positional relationship between the image of the shape (ie, the elliptical image) inherent to the type of the defect (in this example, the bubble) is calculated. The height direction position of the defect in the glass ribbon 5. Specifically, the arithmetic unit 4 calculates the distance from the actual space corresponding to the number of pixels in the rectangle parallel to the direction corresponding to the direction in which the glass ribbon is transported, from the two overlapping images in the image. In the middle, the length of the diameter of the defect (bubble) parallel to the transport direction is subtracted. Further, in the image, the direction parallel to the direction corresponding to the conveyance direction of the glass ribbon is parallel to the conveyance direction line 96 (see Fig. 2A). The arithmetic unit 4 calculates the height direction position of the defect by the value obtained by the above subtraction calculation and the refraction angle β in the glass ribbon 5. This calculation will be described later with reference to FIG. 5. Further, the long axis of the bubble in the glass ribbon is substantially parallel to the conveying direction of the conveying roller (in other words, the extending direction of the glass ribbon 5). As shown in Fig. 3, the direction of the long axis 72 of the bubble and the use of the transfer roller! The shift amount of the glass ribbon 5 in the transport direction 71 is at most 10. . As described above, since the long axis 72 of the bubble is substantially parallel to the transport direction 71 of the transport roller 1, the long axis and the transport direction line of the image of the defect of the elliptical shape are displayed in the image captured by the line camera 3. 96 (see Fig. 2B) is also substantially parallel. Hereinafter, a case where the long axis of the image of the defect in the captured image is parallel to the transport direction line 96 will be described as an example. Further, when the paired images are not stacked, the arithmetic unit 4 calculates the position in the height direction of the defect by a known method. Further, the arrangement position of the line camera 3 is fixed. Therefore, the distance in the actual space corresponding to the m element in the image taken by the line camera 3 is also determined as a fixed value of 160767.doc 201233993. The distance in the actual space set to correspond to the pixels in the image is known. Next, the action will be described. Fig. 4 is a flow chart showing an example of a processing procedure of the in-glass defect measuring system in the embodiment. First, the light source 2 starts the irradiation of light to the glass ribbon 5 (step $1). Then, the conveyance roller 1 conveys the glass ribbon 5 disposed on the conveyance roller 沿 in the fixed direction, and the line camera 3 continues to capture the inside of the conveyed glass ribbon $. Moreover, the line camera 3 generates an image as a shooting result (step S2). The line camera 3 transmits the image obtained by the shooting to the arithmetic unit 4. In the case where there is a defect inside the glass ribbon 5, the image obtained in the step "includes an image of the defect. In this example, the image of the ellipse is reflected in the image because the defect is a bubble. As illustrated in Fig. 13, when the defect moves to a position overlapping the path of the light before reflection (the position 91 shown in the side view of the upper stage of Fig. 13), and the position where the defect overlaps with the path of the reflected light (Position 92 shown in the side view of the upper part of Fig. 13.), respectively, is imaged on the image. Therefore, when there is one defect, the image is projected with two images. When a large situation or defect exists in the vicinity of the interface 9 (refer to FIG. 1) of the glass ribbon 5, the two images overlap. When the arithmetic device 4 receives the image generated in step S2, the image is self-imaged. Detecting two overlapping images and connecting the area of the rectangle, and the side of the circumscribed rectangle in the image parallel to the direction corresponding to the direction in which the glass ribbon is transported (ie, the direction of transport within the image) Count the number of pixels in the parallel side of the line. Then, The arithmetic unit 4 calculates the length in the real space corresponding to the number of pixels of the side by multiplying the number of pixels of the side by the distance in the real space of each image I60767.doc.18-201233993 (step S3) Fig. 5 is an explanatory view showing a region in which two overlapping images are connected to a rectangular rectangle. As shown in Fig. 5, the two overlapping images 21 and 22 are connected to a rectangle, and are determined as a rectangle 23 as shown in Fig. 5. 21 and 22 are elliptical and can be regarded as congruent. In the example shown in Fig. 5, the long side of the circumscribed rectangle 23 is parallel to the transport direction line (see Fig. 2B). In this case, the arithmetic device 4 is opposed. 21, 22 is counted by the number of pixels of the long side 24 of the rectangle 23, and the number of pixels is multiplied by the distance in the actual space of each j pixel. The length in the actual space corresponding to the long side 24 is "h" Said. The unit of h is, for example, μπι. Further, in the case where the defect is a bubble, the central portion 2 1 a of the image 21 is white on the image. The center portion 21a is a feature point of the image 21. The arithmetic unit 4 counts the number of pixels from the center portion 21a of the image 21 to the short side of the short side of the circumscribed rectangle 23 from the shorter side thereof. That is, the number of pixels of the portion indicated by the symbol A in Fig. 5 is counted. The arithmetic unit 4 multiplies the number of pixels by the distance in the actual space of each pixel. The multiplication result is the length in the real space corresponding to the portion corresponding to A shown in FIG. 5, specifically, the diameter of the defect parallel to the transport direction (the diameter of the defect which is parallel to the transport direction) In the example shown by the length of 1/2, the long axis of the defect is the arithmetic unit 4, and by multiplying the result of the multiplication by 2, the length of the diameter of the defect parallel to the transport direction is calculated (step S4). The length of the diameter of the defect is set to s. The unit of s is for example. The portion in the image corresponding to the length of s/2 in the real space is the portion indicated by the symbol a in Fig. 5. Also, since the two images 21 and 22 can be regarded as congruent, the picture can be viewed as 160767.doc 201233993 as a=a'. / Again, here, the situation is calculated by using the heart 2 of the 21st to calculate the s, but the center of the image 22 can also be used to calculate s. In the 'Fig. 5', the wire and the direction of the line of the defect are The parallel shape is described as an example, but the long-drawn drawing of the defective image is completely parallel with the conveying direction line. However, the distance between the long pumping of the bubble in the glass ribbon and the conveying direction of the glass ribbon is the largest. Therefore, it is 1 (refer to Fig. 3). Therefore, even if the long axis of the image of the defect is not completely parallel to the transport direction line, it can be regarded that the two are parallel, and h's is calculated in the same manner as the above steps S3 and 84. For the case of '^,' you only need to count the number of pixels of the long side of the overlapped rectangle, and multiply the number of pixels by the distance in the actual space of each pixel. From the center of the image to the short side of the circumscribed rectangle, the number of pixels is shorter than the short side of the square, multiplying the number of pixels by the distance in the real space per pixel, and multiplying the multiplication Multiply the result by 2. Even if the long axis of the image of the defect and the direction of the transport are not It is completely parallel, and h and s are calculated as described above, and the error direction position of the defect is calculated by using the h and s, and only the error of the degree that can be ignored is included. The human, the computing device 4 calculates the h from the step S3. The s calculated in step §4 is subtracted (step S5). The result of the subtraction is set to ....% is the movement of the defect from the position where the first image is taken to the position where the second image is taken. The distance yd calculated in step S5 is the distance between the two points at which the image of the defect is captured. Furthermore, the portion in the image corresponding to the length in the real space is represented by the symbol B in FIG. The arithmetic unit 4 calculates the height direction of the defect by using the yd calculated in step S5 and the predetermined folding angle 160767.doc 201233993, and calculates the yd/(2). .t_), the result of its calculation as d (step %). The height direction position d of the defect is the interface 9 from the glass ribbon 5 (refer to the distance from the drawing to the defect. (3) According to the present embodiment, even if two images due to the same defect overlap, the height direction position of the defect can be measured. The height direction position of the defect in the glass ribbon can be measured even if there is a defect near the interface of the glass ribbon or a large defect, and the s calculated in step S4 is the long axis of the defect. Further, even when the long axis of the image of the defect is not completely parallel with the direction of the transport direction, the s calculated in step S4 is regarded as the length of the long axis of the defect, and only the error of the degree of negligibility is included. Therefore, the size of the defect (the length of the long axis) can be calculated. Further, according to the present embodiment, the light source 2 and the line camera 3 can be disposed on one side of the glass ribbon 5. Therefore, the third measurement method (see In the second embodiment, the second embodiment of the present invention includes a conveying roller in the same manner as the first embodiment. 1 The light source 2, the line camera 3, and the arithmetic unit 4 (see FIG. 。). The positional relationship between the light source 2 and the line camera 3 with respect to the glass ribbon 5 is the same as that of the first embodiment, and the description thereof will be omitted. In the second embodiment, The method of measuring the height direction position of the defect by the arithmetic unit 4 is different from that of the first embodiment. In the second embodiment, the arithmetic unit 4 calculates the characteristic value of the defect 160767.doc 201233993 in the glass ribbon $. Then the arithmetic unit 4 Using the characteristic value, the following value is calculated: the length in the real space corresponding to the number of pixels in the outer rectangle of the two images overlapping the direction corresponding to the direction in which the glass ribbon is transported, minus The length of the diameter of the defect in which the glass ribbon is conveyed in parallel (the diameter of the diameter of the defect parallel to the transport direction). Further, when the characteristic value of the arithmetic unit 4 is based on the positional relationship of the two images superimposed, the predetermined use is used. In the second embodiment, the length of the diameter of the defect parallel to the direction in which the glass ribbon is conveyed is calculated as the characteristic value. The equation is determined in advance as a function of the coordinates of the position corresponding to the feature point of the image based on the end of the glass ribbon, the h described in the first embodiment, and the area of the two overlapping images. The equation for determining the characteristic value (the diameter of the diameter of the defect parallel to the transport direction) can be expressed, for example, by the following formula (4): ssay+ay+ay+aeh+ashp+awp+ayu+ash + agp+a] In the formula (4), "u" is a coordinate at a position corresponding to a feature point of an image based on an end portion of the glass ribbon, and specifically, a glass parallel to the transport direction. The distance from the side of the belt to the center of the defect. Here, the unit of u is set to mm e "h" as the image obtained by taking the defect, and is obtained by the same calculation as step S3 in the third embodiment. value. Here, the unit of h is set to μπι. p is the area of the area occupied by the two images (the combination of the two image areas) in the image in which the defect is captured. Specifically, the ai~aio in the formula (4) is expressed by the number of pixels in the image. coefficient. Further, s in the formula (4) is a diameter of a defect parallel to the conveying direction of the glass ribbon. Due to the diameter of the characteristic value s 160767.doc

S •22- 201233993 易於受玻璃帶之寬度方向之缺陷之位置所影響，故而將包含上述變數U之算式（例如為上述式（4))用於s之計算。又’於所拍攝之圖像中，缺陷之像之長軸與搬送方向線平行之情形時，上述s相當於缺陷之長軸。但是，於圖像中’缺陷之像之長軸與搬送方向線不完全平行之情形時，亦由於兩者大致平行，故而上述特徵值s可看作缺陷之長軸。即便如上所述看作缺陷之長軸’亦僅包含可忽視之程度之誤差’不影響缺陷之高度方向位置之算出。式（4)中之係數係藉由最小平方法而預先求出。具體而S，利用成為樣本之缺陷，實際測量s、u。又對於包含成為樣本之缺陷之玻璃帶，進行與第i實施形態中所說明之步驟S1〜S3相同之處理而獲得he又，自當時步驟S2 中所獲得之圖像中’對2個像之並集之區域之像素數p進行計數。準備複數個成為樣本之缺陷，對於料各缺陷，以上述方式獲得s、u、h、卜若獲得複數組s、u、h、p之組，則自該等S、U、h、p之組中，藉由最小平方法而求出式（4)中之係數a,〜a1〇即可。 s與u、h、p之間存在關聯，藉稽由竑小平方法，可求出式 (4)中之各係數。運算裝置4根據藉由拍攝缺 μ A 陷之同度方向位置成為測對象之玻璃帶而獲得之圖像，令 ^ 衣出u、h、P，並代入式( 中，藉此算出s。繼而，運算举裝置4s十算h-s(=yd)，利用計算結果及折射角β，算出缺度方向位置。其次，對第2實施形態之動作進灯說明。圖6係表示| 160767.doc •23· 201233993 實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流程圖。對於與圖4所示之處理相同之處理，標註與圖4相同之符號且省略說明。至步驟S3中算出h為止之動作與第1實施形態相同。圖7係表示圖像内所映出之玻璃帶之例之說明圖。於存在缺陷之情形時，圖像内亦映出缺陷之像21、22。又，圖 7所示之例中’作為像之特徵點之各像21、22之中心部分 2 U ' 22a亦作為白色之區域而出現於圖像内。再者，雖圖示有像21、22之外接矩形23，但外接矩形23並不會映於圖像内。於步驟S3之後，運算裝置4對自圖像内之玻璃帶之端部 3 1至像之特徵點為止之像素數進行計數。即，對圖7中符號c所不之部分之像素數進行計數。繼而，運算裝置4將該像素數乘以每1像素之實際空間中之距離（步驟su)。該乘法運算結果相當於自實際空間中之玻璃帶之端部（側面）至缺陷為止之距離u。即，於步驟S11中算出u。其中’於上述步驟S11之說明中，.為易於說明’以破璃帶之端部3 1映於圖像内之情形為例進行說明。於玻璃帶之端部3 1未映於圖像内之情形時，以下述方式計算距離u即可。由於線陣相機3之設置位置為固定，故而可預先求出自玻璃帶之端部至藉由線陣相機3所拍攝之圖像内之破璃帶端部側之端為止的實際空間中之距離（設為。繼而，運算裝置4計算自所拍攝之圖像中之該端之部分至像之特徵點為止之距離。於該計算中，例如只要對自圖像中之該 160767.docS • 22- 201233993 is susceptible to the position of the defect in the width direction of the glass ribbon, so the equation containing the above variable U (for example, the above formula (4)) is used for the calculation of s. Further, in the captured image, when the long axis of the image of the defect is parallel to the direction of the transport direction, the above s corresponds to the long axis of the defect. However, in the case where the long axis of the image of the defect is not completely parallel with the transport direction line in the image, since the two are substantially parallel, the feature value s can be regarded as the long axis of the defect. Even if the long axis 'of the defect as described above contains only the error of the degree of negligence', the calculation of the height direction position of the defect is not affected. The coefficients in the formula (4) are obtained in advance by the least squares method. Specifically, S, using the defects of the sample, actually measures s, u. Further, the glass ribbon including the defect as the sample is subjected to the same processing as the steps S1 to S3 described in the i-th embodiment to obtain he, and the image obtained from the step S2 at that time is 'two images. The number of pixels p in the area of the union is counted. Prepare a plurality of defects that become samples, and for each defect, obtain s, u, h, and if the group of complex arrays s, u, h, and p are obtained in the above manner, then from the S, U, h, and p In the group, the coefficient a, ~a1 中 in the equation (4) can be obtained by the least square method. There is an association between s and u, h, and p. The coefficients in equation (4) can be found by the method of 竑 Xiaoping. The arithmetic unit 4 obtains an image obtained by photographing the glass ribbon to be measured in the direction of the same direction of the lack of μ A, and then extracts u, h, and P, and substitutes the equation (in this, thereby calculating s. The calculation device 4s calculates hs (=yd), and uses the calculation result and the refraction angle β to calculate the position in the direction of the misalignment. Next, the operation of the second embodiment will be described. Fig. 6 shows |160767.doc •23 201233993 A flowchart of an example of the processing procedure of the in-glass defect measuring system in the embodiment. The same steps as those in the processing shown in FIG. 4 are denoted by the same reference numerals as in FIG. 4, and the description thereof is omitted. The operation is the same as that of the first embodiment. Fig. 7 is an explanatory view showing an example of a glass ribbon reflected in an image. When there is a defect, the image 21 and 22 of the defect are also reflected in the image. In the example shown in Fig. 7, the central portion 2 U ' 22a of the respective images 21 and 22 which are characteristic points of the image also appears in the image as a white region. Further, although the images 21 and 22 are illustrated, The rectangle 23 is connected, but the circumscribed rectangle 23 is not reflected in the image. After step S3, the arithmetic unit 4 counts the number of pixels from the end portion 31 of the glass ribbon in the image to the feature point of the image. That is, the number of pixels of the portion of the symbol c in Fig. 7 is counted. Then, the arithmetic unit 4 multiplies the number of pixels by the distance in the real space per one pixel (step su). The result of the multiplication corresponds to the distance from the end (side) of the glass ribbon in the actual space to the defect. That is, u is calculated in step S11. In the description of the above step S11, the case where the end portion 31 of the broken glass ribbon is reflected in the image is described as an example. When the end portion 31 is not reflected in the image, the distance u can be calculated in the following manner. Since the position of the line camera 3 is fixed, it can be determined in advance from the end of the glass ribbon to the line array. The distance in the actual space from the end of the end of the broken strip in the image taken by the camera 3 (set. Then, the arithmetic unit 4 calculates the portion from the end of the captured image to the image. The distance to the point. In this calculation, for example, as long as The image of the 160767.doc

S -24- 201233993 端之部分至特徵點為止之像素數進行計數，並將該像素數采以母1像素之貫際空間中之距離即可。運算裝置4只要科由將該距離加上利用線陣相機設置位置決定之u〇，而算出自實際空間中之玻璃帶之端部（側面）至缺陷為止之距離u即可。再者，圖7所示之例中，以使用像21之中心部分2U作為特徵點而求出自圖像内之玻璃帶之端部31至中心部分為止之距離的情形為例。作為特徵點，亦可使用另一個像 22之中心部分22a。無論使用哪個中心部分作為特徵點，均可求出自實際空間中之玻璃帶之端部（側面）至缺陷為止之距離u。根據使用中心部分21a、22a中之哪一個作為特徵點，像素數之計數結果有所不同，但其差異微小，距離u 僅包含可,t、視之誤差4，作為特徵點，亦可使用外接矩形23内之特徵性之點（例如，外接矩形23之任一頂點）。於該情形時，距離u亦僅包含可忽視之誤差。於步驟sii之後，作為重疊之2個像21、22所佔區域⑺固像之區域之並集）之面積，運算裝置4對該區域内之像素數 P進行計數（步驟S12)。繼而，運算裝置4藉由將步驟S3、su、S12中所求出之 h_、u、p代入式（4)中，而計算缺陷之直徑中與搬送方向平行^直徑8(步驟S13)e如圖7所示，像之長軸與搬送方向線平行之情形時，該直徑3為缺陷之長軸。如已說明般，於拍攝圖像中’即便像之長抽與搬送方向線不完全平行，亦由於兩者大致平行，故而步驟SI3中計算出之直徑3可看作 160767.doc -25- 201233993 缺陷之長轴。以後之處理與第1實施形態中之步驟S5、S6相同。即，運算裝置4藉由自步驟S3中所算出之h中減去步驟S13中所算出之s，而求出yd(步驟S5)e繼而，運算裝置4利用〜及折射角β，進行式（2)之計算，而計算缺陷之高度方向位置 d 〇第2實施形態中亦可獲得與第丨實施形態相同之效果。又，於第2實施形態中，亦在步驟S13中算出s之值，因此亦可求出缺陷之大小（長軸）。 [實施形態3] 本發明之第3實施形態與第丨實施形態同樣地包含搬送輥 1、光源2、線陣相機3、及運算裝置4(參照圖丨）。光源2及線陣相機3相對於玻璃帶5之位置關係與第丨實施形態相同，而省略說明。於第3實施形態中，運算裝置4亦算出玻璃帶5内之缺陷特徵值’利用該特徵值，計算yd。但是，第2實施形態人算出缺陷之直徑s作為特徵值，而於第3實施形態中，計算缺陷之2個直徑之比。具體而言，運算裝置4求出缺陷之直徑中搬送方向之直徑相對於與搬送方向正交之方向之直徑的比例作為缺陷之特徵值。Μ，若將缺陷之直徑中與送方向正父之方向之直徑設為ΙΊ、搬送方向之直徑設為則计算Γ2/Γι作為特徵值。以下，將r2/r丨記作r。、，再者’於所拍攝之圖像中缺陷之像之長轴與搬送方向線、丁之隋形時，上述ri相當於缺陷之短軸，r2相當於缺陷 160767.docS -24- 201233993 The number of pixels from the end of the end to the feature point is counted, and the number of pixels is taken as the distance in the space of the parent 1 pixel. The arithmetic unit 4 calculates the distance u from the end portion (side surface) of the glass ribbon in the actual space to the defect by adding the distance to the distance determined by the position of the line camera. Further, in the example shown in Fig. 7, the case where the distance from the end portion 31 of the glass ribbon in the image to the center portion is obtained using the central portion 2U of the image 21 as a feature point is taken as an example. As a feature point, the central portion 22a of the other image 22 can also be used. Regardless of which central portion is used as the feature point, the distance u from the end (side) of the glass ribbon in the actual space to the defect can be obtained. According to which one of the central portions 21a, 22a is used as the feature point, the counting result of the number of pixels is different, but the difference is small, the distance u includes only the error, t, and the error 4, and as the feature point, an external connection can also be used. A characteristic point within the rectangle 23 (eg, any apex of the circumscribed rectangle 23). In this case, the distance u also contains only errors that can be ignored. After step sii, the arithmetic unit 4 counts the number of pixels P in the area as the area of the area of the area (7) where the two images 21 and 22 are overlapped (step S12). Then, the arithmetic unit 4 substitutes h_, u, and p obtained in steps S3, su, and S12 into equation (4) to calculate the diameter of the defect parallel to the transport direction, and the diameter 8 (step S13) e. As shown in Fig. 7, when the long axis is parallel to the transport direction line, the diameter 3 is the long axis of the defect. As described above, in the captured image, 'even if the long stroke of the image is not completely parallel with the transport direction line, since the two are substantially parallel, the diameter 3 calculated in step SI3 can be regarded as 160767.doc -25- 201233993 The long axis of the defect. The subsequent processing is the same as steps S5 and S6 in the first embodiment. In other words, the arithmetic unit 4 obtains yd by subtracting the s calculated in step S13 from h calculated in step S3 (step S5) e, and then the arithmetic unit 4 performs the equation using the ? and the refraction angle β ( 2) Calculation, and calculating the height direction position d of the defect 〇 In the second embodiment, the same effect as that of the second embodiment can be obtained. Further, in the second embodiment, since the value of s is also calculated in step S13, the size of the defect (long axis) can be obtained. [Embodiment 3] A third embodiment of the present invention includes a transport roller 1, a light source 2, a line camera 3, and an arithmetic unit 4 (see Fig. 同样) in the same manner as the third embodiment. The positional relationship between the light source 2 and the line lens camera 3 with respect to the glass ribbon 5 is the same as that of the second embodiment, and the description thereof is omitted. In the third embodiment, the arithmetic unit 4 also calculates the defect characteristic value in the glass ribbon 5, and uses the characteristic value to calculate yd. However, in the second embodiment, the diameter s of the defect is calculated as the characteristic value, and in the third embodiment, the ratio of the two diameters of the defect is calculated. Specifically, the arithmetic unit 4 obtains the ratio of the diameter of the defect in the transport direction to the diameter in the direction orthogonal to the transport direction as the feature value of the defect. ΜIf the diameter of the defect in the direction of the normal direction of the feed direction is ΙΊ and the diameter of the transport direction is set, Γ2/Γι is calculated as the eigenvalue. Hereinafter, r2/r丨 is denoted as r. Furthermore, when the long axis of the image of the defect in the captured image and the direction of the transport direction, and the shape of the Ding, the above ri corresponds to the short axis of the defect, and r2 corresponds to the defect 160767.doc

S -26- 201233993 之長軸。即’作為特徵值r，計算「長軸/短軸」。但是，即便於圖像中缺陷之像之長軸與搬送方向線不完全平行之情形時’亦由於兩者大致平行’故而可將上述^看作缺陷之知·轴’將上述!>2看作缺陷之長軸。即，即便於圖像中缺陷之像之長軸與搬送方向線不完全平行之情形時，亦可將作為特徵值而計算之r看作缺陷之「長軸/短軸」。即便如上述般看待，r亦僅包含可忽視之程度之誤差，並不影響缺陷之兩度方向位置之算出。運算裝置4算出r作為缺陷之特徵值後，利用該r，求出 yd(自第1個像被拍攝之位置至第2個像被拍攝之位置為止之缺陷之移動距離）。又，運算裝置4於計算上述特徵值“夺，根據重疊之2個像之位置關係’利用預先決定之算式計算特徵值。用以算出該特徵值ri式作為下述函數被預先決定，該函數將以玻璃帶之端部為基準之像之特徵點所對應之位置的座標、第1實施形態中所說明之h、及2個重疊之像之面積設為變數。用以求出特徵值r之算式能夠以例如以下之式（5)表示。 r=b 丨 u2+b2h2+b3P2+b4uh+bM^ 式(5) 該函數巾之變數U、h、P與第2實施形態中所示之式⑷ 中之變數u、h、p相同。即，「u」為自與搬送方向平行之玻璃帶之側面至缺陷之中心為止之距離。「h」為根據拍攝到缺之圖像，藉由與第j實施形態中之步驟叫目同之計异而獲4之值。P為於拍攝到缺陷之圖像中，2個像所佔區 160767.doc -27- 201233993 域(2個像之區域之並集)之面積，具體而言，以圖像内之像素數表不式(5)中之bpb,❶為係數。由於特徵值犷易於受破帶之寬度方向之缺陷的位置所影響，故而將包含上述變數11之算式（例如為上述式（5))用於!*之計算。式（5)中之係數bl〜bi〇係藉由最小平方法而預先求出。具體而s ’利用成為樣本之缺陷，實際測量犷、u。又，對於包含成為樣本之缺陷之玻璃帶，進行與第i實施形態中所說明之步驟S卜S3相同之處理而獲得h。又，根據當時步驟 S2中所獲仔之圖像’對2個像之並集之區域之像素數p進行 «十數。準備複數個成為樣本之缺陷，對於該等各缺陷，以上述方式獲得r、u、h、P。若獲得複數組r、u、h、p之組，則根據該等r、u、h、p之組，藉由最小平方法，求出式（5)中之係數b丨〜b丨0即可。 r與u、h、p之間存在關聯，藉由最小平方法，可求出式 (5)中之各係數。運算裝置4根據藉由拍攝缺陷之高度方向位置成為測定對象之玻璃帶而獲得之圖像，求出u、h、p，並代入式中，藉此算出r。又，運算裝置4於所拍攝之圖像中將搬送方向線％與通過2個像之辛心之線所成之角設為^時，求出之值。繼而’運算裝置4利用h、u、r、tane，計算yd。運算裝置4利用該yd及折射角β，算出缺陷之高度方向位置。其次’對第3實施形態之動作進行說明。圖8係表示第3 實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流 I60767.doc •28· 201233993 私圖。對於與第1實施形態及第2實施形態相同之處理，標 s主與圖4及圖6相同之符號且省略說明。至步驟S12中求出p為止之動作（步驟S1、s2、S3 ' S11、 S1 2)與第2實施形態相同。於步驟S12之後，運算裝置4藉由將步驟S3、sn ' Sl2中所求出之h、u、p代入式（5)中，而計算Γ(即，缺陷之直徑 . 中搬送方向之直徑之長度相對於與搬送方向正交之方向之直徑之長度的比例）（步驟S21)。圖9係表示圖像内所映出之玻璃帶之例之說明圖。對於與圖7相同之要素，標註與圖7相同之符號且省略說明。於步驟S21之後，運算裝置4對2個重疊之像21、22之外接矩形23之邊中與相當於玻璃帶之搬送方向之方向正交之邊（換言之，為與圖像中之搬送方向線正交之邊）的像素數進行計數。即，對圖9中符號D所示之部分之像素數進行計數。繼而，運算裝置4將該像素數乘以每〗像素之實際空間中之距離（步驟S22)。將其結果所獲得之長度記作w。即， w為對應於圖9中符號D所示之部分之實際空間中的長度。又，運算裝置4求出外接矩形之邊中與相當於玻璃帶之搬送方向之方向平行之邊、與通過2個像21、22之中心部 • 分21a、22a之線所成之角Θ的正切即tan0(步驟S23)。 Θ亦可稱為通過2個像21、22之中心部分2U、22a之線與搬送方向線所成之角。因此，運算裝置4亦可例如預先決定yc(參照圖14)之值，利用已說明之方法計算並進行式（3)之计算’藉此計算tan0。或者，亦可利用其他方法計 160767.doc •29· 201233993 算 tan9。其次，運算裝置4利用至步驟S23為止之處理中之計算結果即h、r、w、tane ’算出yd(步驟S24)。具體而言運算裝置4只要藉由進行以下所示之式（6)之計算而計算&即可。 yd = (h-r.w)/(l-r.tane) 式（6) 運算裝置4利用上述yd及預先決定之折射角β，進行式（2) 之計算，而計算缺陷之高度方向位置<1(步驟S25)。該計算與第1實施形態中之步驟S6相同。再者，亦可於步驟S25中求出缺陷之高度方向位置而作為處理結束。又，與其他實施形態同樣地，於算出與搬送方向平行之缺陷之直徑之長度之情形時，運算裝置4只要藉由自h中減去yd而算出s即可（步驟S26)。於第3實施形態中亦可獲得與第丨實施形態或第2實施形態相同之效果。又，第3實施形態中，作為缺陷之特徵值，亦可獲得缺陷之直徑之比即r。上述各實施形態中，運算裝置4可藉由例如按照程式進行動作之電腦而實現。例如，電腦可按照程式，作為運算裝置4而進行動作。又上述各實施形態十，以缺陷為氣泡之情形為例，但本發明中，作為測定對象之缺陷不限定於氣泡，只要為滿足作為包含特徵點之一定形狀之像被㈣之條件的缺陷即可。作為如上所述之缺陷，除氣泡以外可舉出異物等。產業上之可利用性 160767.doc 201233993 本發明可較佳地應用於玻璃帶㈣之缺陷之高度方向位置等的測定》已詳細地且參照特定之實施態樣對本申請案進行了說明’但業者應明白只要不脫離本發明之精神及範圍便可進行各種變更或修正。本申清案係基於2010年12月9曰申請之曰本專利申請（日本專利特願2010-275048)者，且其内容作為參照而併入本文。【圖式簡單說明】圖1係表示本發明之玻璃帶内缺陷測定系統之構成例之模式圖。圖2A係表示中央線之說明圖。圖2B係表示圖像内之相當於中央線之線之說明圖。圖3係表示玻璃帶内之氣泡之長軸之方向與搬送輥1之搬送方向之關係的說明圖。圖4係表示第1實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流程圖。圖5係表示2個重疊之像之外接矩形之區域的說明圖。圖6係表示第2實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流程圖。圖7係表示圖像内所映出之玻璃帶之例之說明圖。圖8係表示第3實施形態中之玻璃帶内缺陷測定系統之處理過程之例的流程圖。圖9係表示圖像内所映出之玻璃帶之例之說明圖。 160767.doc •31· 201233993 圖10A係模式性地表示第i測定方法之說明圖。圖10B係表示第1測定方法中所拍攝之缺陷之圖像之例的說明圖。圖11A係模式性地表示第2測定方法之說明圖。圖Π β係表示第2測定方法中所拍攝之缺陷之圖像之例的說明圖。圖12 Α係模式性地表示第3測定方法之說明圖。圖12B係表示第3測定方法中所拍攝之缺陷之圖像之例的說明圖。圖13係表示所搬送之玻璃帶内之缺陷被線陣相機拍攝時之位置的說明圖。圖14係拍攝距離yc之說明圖。【主要元件符號說明】 1 搬送輥 2 光源 3 線陣相機 4 運算裝置 5 玻璃帶 8 界面 9 界面 21 像 21a 中心部 22 像 22a 中心部 160767.doc 201233993 23 外接矩形 24 長邊 31 玻璃帶之端部 71 搬送方向 72 氣泡之長軸 81 線陣相機 81a 第1線陣相機 81b 第2線陣相機 82 玻璃帶 83 缺陷 84 像 85 像 86 缺陷之像 91 位置 92 位置 95 中央線 96 搬送方向線 98 第1像 99 第2像 d 高度方向位置 SI 〜S6 步驟 S11-S13 步驟 S21〜S26 步驟 Xcc 偏移量 160767.doc -33- 201233993 yi 最小值 yi . 最大值 yc 拍攝距離 yd 移動距離 a 入射角 β 折射角 160767.doc • 34· sThe long axis of S -26- 201233993. That is, as the feature value r, "long axis/short axis" is calculated. However, even if the long axis of the image of the defect in the image is not completely parallel to the direction of the transport direction, 'there is a fact that the two are substantially parallel, so the above-mentioned ^ can be regarded as the knowledge of the axis.' Think of the long axis of the defect. That is, even when the long axis of the image of the defect in the image is not completely parallel to the direction of the transport direction, r calculated as the feature value can be regarded as the "long axis/short axis" of the defect. Even if viewed as above, r contains only errors that can be ignored, and does not affect the calculation of the position of the defect in the two directions. When the arithmetic unit 4 calculates r as the feature value of the defect, the r is used to obtain yd (the moving distance of the defect from the position where the first image is captured to the position where the second image is imaged). Further, the arithmetic unit 4 calculates a feature value by calculating a predetermined value based on the positional relationship of the two images that are superimposed. The calculation means ri is used to calculate the feature value ri as a function, and the function is determined in advance. The coordinates of the position corresponding to the feature point of the image based on the end of the glass ribbon, the h described in the first embodiment, and the area of the two superimposed images are used as variables. The calculation formula can be expressed, for example, by the following formula (5): r = b 丨 u 2+ b2h 2+ b3 P 2+ b4uh + bM ^ (5) The variables U, h, and P of the function towel are as shown in the second embodiment. The variables u, h, and p in the formula (4) are the same, that is, "u" is the distance from the side of the glass ribbon parallel to the conveyance direction to the center of the defect. "h" is a value obtained by the same method as the step in the jth embodiment, based on the image taken out of the missing image. P is the area of the image in which the two images occupy the defect, 160767.doc -27- 201233993 (the union of the two image regions), specifically, the number of pixels in the image. Not in the bpb in the formula (5), ❶ is a coefficient. Since the characteristic value 犷 is easily affected by the position of the defect in the width direction of the broken tape, the equation including the above-described variable 11 (for example, the above formula (5)) is used for the calculation of !*. The coefficients bl 〜 bi 〇 in the formula (5) are obtained in advance by the least squares method. Specifically, s 'uses the defects of the sample, and actually measures 犷, u. Further, for the glass ribbon containing the defect as a sample, the same processing as the step Sb of S1 described in the i-th embodiment is performed to obtain h. Further, the number of pixels p of the region where the two images are combined is «10 times according to the image obtained in the step S2 at that time. A plurality of defects which become samples are prepared, and for each of the defects, r, u, h, and P are obtained in the above manner. If a group of complex arrays r, u, h, and p is obtained, the coefficients b丨 to b丨0 in equation (5) are obtained by the least squares method according to the groups of such r, u, h, and p. can. There is an association between r and u, h, and p, and the coefficients in equation (5) can be found by the least squares method. The arithmetic unit 4 calculates u from the image obtained by photographing the glass ribbon to be measured in the height direction of the defect, and obtains u, h, and p, and substitutes the equation to calculate r. Further, the arithmetic unit 4 obtains a value when the angle between the transport direction line % and the line passing through the imaginary lines of the two images is set to ^ in the captured image. Then, the arithmetic unit 4 calculates yd using h, u, r, and tane. The arithmetic unit 4 calculates the position in the height direction of the defect by using the yd and the refraction angle β. Next, the operation of the third embodiment will be described. Fig. 8 is a flow chart showing an example of the processing procedure of the in-glass defect measuring system in the third embodiment. I60767.doc • 28· 201233993 Private drawing. The same processes as those in the first embodiment and the second embodiment are denoted by the same reference numerals as those in Figs. 4 and 6 and their description will be omitted. The operation until step p is obtained in step S12 (steps S1, s2, S3 'S11, S1 2) is the same as in the second embodiment. After the step S12, the arithmetic unit 4 calculates the Γ (ie, the diameter of the defect. The diameter of the transport direction) by substituting h, u, and p obtained in the steps S3 and sn 'S1 into the equation (5). The ratio of the length to the length of the diameter in the direction orthogonal to the conveyance direction) (step S21). Fig. 9 is an explanatory view showing an example of a glass ribbon reflected in an image. The same elements as those in Fig. 7 are denoted by the same reference numerals as those in Fig. 7, and the description thereof is omitted. After step S21, the arithmetic unit 4 separates the sides of the two overlapping images 21 and 22 which are orthogonal to the direction of the transport direction of the glass ribbon (in other words, the direction of the transport direction in the image). The number of pixels of the orthogonal side is counted. That is, the number of pixels of the portion indicated by the symbol D in Fig. 9 is counted. Then, the arithmetic unit 4 multiplies the number of pixels by the distance in the real space of each pixel (step S22). The length obtained by the result is referred to as w. That is, w is the length in the real space corresponding to the portion indicated by the symbol D in Fig. 9. Further, the arithmetic unit 4 obtains the angle between the side of the circumscribed rectangle that is parallel to the direction corresponding to the direction in which the glass ribbon is conveyed, and the line that passes through the line of the center portions 21a and 22a of the two images 21 and 22. The tangent is tan0 (step S23). The crucible may also be referred to as an angle formed by a line passing through the center portions 2U and 22a of the two images 21 and 22 and a conveying direction line. Therefore, the arithmetic unit 4 can calculate the value of yc (refer to Fig. 14) in advance, and calculate and perform the calculation of the equation (3) by the method described above, thereby calculating tan0. Alternatively, you can use other methods to calculate tan9. Next, the arithmetic unit 4 calculates yd using h, r, w, and tane', which are the calculation results in the processing up to step S23 (step S24). Specifically, the arithmetic unit 4 can calculate & by simply performing the calculation of the equation (6) shown below. Yd = (hr.w) / (lr.tane) Equation (6) The arithmetic unit 4 calculates the height direction position of the defect by using the above yd and the predetermined refraction angle β to calculate the position of the defect in the height direction <1 ( Step S25). This calculation is the same as step S6 in the first embodiment. Further, the position in the height direction of the defect may be obtained in step S25 as the end of the process. Further, similarly to the other embodiments, when calculating the length of the diameter of the defect parallel to the transport direction, the arithmetic unit 4 calculates s by subtracting yd from h (step S26). Also in the third embodiment, the same effects as those of the second embodiment or the second embodiment can be obtained. Further, in the third embodiment, as the characteristic value of the defect, r which is the ratio of the diameter of the defect can be obtained. In each of the above embodiments, the arithmetic unit 4 can be realized by, for example, a computer that operates in accordance with a program. For example, the computer can operate as the arithmetic unit 4 in accordance with the program. In the above-described tenth embodiment, the case where the defect is a bubble is taken as an example. However, in the present invention, the defect to be measured is not limited to the bubble, and is a defect that satisfies the condition that the image having a certain shape including the feature point is (4). can. As a defect as mentioned above, a foreign material etc. are mentioned except a bubble. Industrial Applicability 160767.doc 201233993 The present invention can be preferably applied to the measurement of the height direction position and the like of the defect of the glass ribbon (4). The present application has been described in detail with reference to specific embodiments. It is to be understood that various changes and modifications may be made without departing from the spirit and scope of the invention. The present application is based on the Japanese Patent Application No. 2010-275048, filed on Dec. 9, 2010, the content of which is hereby incorporated by reference. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic view showing a configuration example of a defect measuring system for a glass ribbon according to the present invention. Fig. 2A is an explanatory view showing a center line. Fig. 2B is an explanatory view showing a line corresponding to a center line in an image. Fig. 3 is an explanatory view showing the relationship between the direction of the long axis of the bubble in the glass ribbon and the conveying direction of the conveying roller 1. Fig. 4 is a flow chart showing an example of a process of the in-glass defect measurement system in the first embodiment. Fig. 5 is an explanatory view showing a region in which two overlapping images are connected to a rectangular shape. Fig. 6 is a flow chart showing an example of a process of the in-glass defect measurement system in the second embodiment. Fig. 7 is an explanatory view showing an example of a glass ribbon reflected in an image. Fig. 8 is a flow chart showing an example of a process of measuring the flaw in the glass ribbon in the third embodiment. Fig. 9 is an explanatory view showing an example of a glass ribbon reflected in an image. 160767.doc • 31·201233993 FIG. 10A is an explanatory diagram schematically showing the i-th measuring method. Fig. 10B is an explanatory view showing an example of an image of a defect captured in the first measurement method. Fig. 11A is an explanatory view schematically showing a second measurement method. Fig. β is an explanatory diagram showing an example of an image of a defect photographed in the second measurement method. Fig. 12 is an explanatory view schematically showing a third measurement method. Fig. 12B is an explanatory view showing an example of an image of a defect photographed in the third measuring method. Fig. 13 is an explanatory view showing a position at which a defect in the conveyed glass ribbon is taken by a line camera. Fig. 14 is an explanatory diagram of the shooting distance yc. [Description of main component symbols] 1 Transport roller 2 Light source 3 Linear camera 4 Arithmetic device 5 Glass ribbon 8 Interface 9 Interface 21 Image 21a Center 22 Image 22a Center 160767.doc 201233993 23 External rectangle 24 Long side 31 Glass end Part 71 Transport direction 72 Long axis of the bubble 81 Line camera 81a First line camera 81b Second line camera 82 Glass band 83 Defect 84 Image 85 Image 86 Defect image 91 Position 92 Position 95 Center line 96 Transport direction line 98 First image 99 Second image d Height direction position SI to S6 Step S11-S13 Step S21 to S26 Step Xcc Offset 160767.doc -33- 201233993 yi Minimum value yi . Maximum value yc Shooting distance yd Moving distance a Incident angle β refraction angle 160767.doc • 34· s

Claims

201233993 七、申請專利範圍： 1. 一種玻璃帶内缺陷測定方法，其特徵在於包括：拍攝步驟’其係自光源對所搬送之玻璃帶照射光，藉由配置於以上述玻璃帶反射之光所到達之位置上的拍攝機構拍攝.上述玻璃帶；及運算步驟，其係根據由上述拍攝機構所拍攝之圖像内 • 之起因於上述玻璃帶之同一缺陷之2個重疊的像且上述缺陷之種類所固有之形狀之2個像的位置關係，算出上述玻璃帶内之上述缺陷之高度方向位置。 2·如請求項1之玻璃帶内缺陷測定方法，其中於上述運算步驟中，計算自上述2個重疊之像中之一個像的拍攝位置至另一個像之拍攝位置為止之缺陷的移動距離；藉由計算出之上述移動距離、及上述玻璃帶内之光之折射角，而算出上述玻璃帶内之上述缺陷之高度方向位置。 3. 如請求項2之玻璃帶内缺陷測定方法，其中於上述運算步驟中，藉由自上述2個重疊之像之外接矩形中的與相當於上述玻璃帶之搬送方向之方向平行之邊的像素數所 ‘ 對應之實際”中之長度中’減去與上述搬送方向平行之缺陷之直徑之長度，而算出上述移動距離。 4. 如請求項2之玻璃帶内缺陷測定方法，其中於上述運算步驟中，根據上述2個重疊之像之位置關係，利用包含上述二璃帶之寬度方向上之像之位置作為變數的預先決定之算式，算出上述缺陷之特徵值，利用該特徵值，算 160767.doc 201233993 出上述移動距離》 5.如請求項4之玻璃帶内缺陷測定方法，其中上述特徵值為與上述玻璃帶之搬送方向平行之缺陷之直徑的長度，藉由自上述2個重疊之像之外接矩形中的與相當於上述搬送方向之方向平行之邊的像素數所對應之實際空間中之長度中’減去上述直徑之長度，而算出上述移動距離。 6_如請求項4之玻璃帶内缺陷測定方法，其中上述特徵值為上述缺陷之2個直徑之比，藉由相當於上述拍攝機構之正面方向之拍攝位置的圖像内之線與通過上述2個像之各中心之線所成的角、及上述比，而算出上述移動距離。 7· 種玻璃帶内缺陷測定系統，其特徵在於包含：搬送機構，其搬送缺陷之咼度方向位置成為測定對象之玻璃帶；光源’其對上述玻璃帶照射光；拍攝機構，其配置於以上述玻璃帶反射之光所到達之位置上’拍攝上述玻璃帶；及運算機構，其根據由上述拍攝機構所拍攝之圖像内之起因於上述玻璃帶之同一缺陷之2個重疊的像且上述缺陷之種類所固有之形狀之2個像的位置關係，算出上述玻璃帶内之上述缺陷之高度方向位置。 π求項7之玻璃帶内缺陷測定系統，其中運算機構計算自上述2個重疊之像中之一個像的拍攝位置至另一個 I60767.doc 201233993 像之拍攝位置為止之缺陷的移動距離，且藉由計算出之上述移動距離、及上述玻璃帶内之光之折射角，而算出上述玻璃帶内之上述缺陷之高度方向位置。 9·如請求項7之玻璃帶内缺陷測定系統，其中運算機構藉由自上述2個重叠之像之外接矩形中的與相當於上述玻璃帶之搬送方向之方向平行之邊的像素數所對應之實際空間中之長度中’減去與上述搬送方向平行之缺陷之直徑之長度，而算出上述移動距離。 10.如請求項7之玻璃帶内缺陷測定系統，其中運算機構根據上述2個重疊之像之位置關係，利用包含上述玻璃帶之寬度方向上之像之位置作為變數的預先決定之算式，算出上述缺陷之特徵值，利用該特徵值，算出上述移動距離。 160767.doc201233993 VII. Patent application scope: 1. A method for measuring defects in a glass ribbon, comprising: a photographing step of: irradiating light to a glass ribbon conveyed from a light source, and arranging light reflected by the glass ribbon The photographing mechanism at the position of the photographing unit captures the glass ribbon; and the calculating step of the two overlapping images caused by the same defect of the glass ribbon in the image captured by the photographing mechanism and the defect The positional relationship between the two images of the shape intrinsic to the type is calculated in the height direction of the defect in the glass ribbon. 2. The method of measuring a defect in a glass ribbon according to claim 1, wherein in the calculating step, calculating a moving distance of a defect from a shooting position of one of the two overlapping images to a shooting position of the other image; The height direction position of the defect in the glass ribbon is calculated by calculating the moving distance and the refraction angle of the light in the glass ribbon. 3. The method for measuring a defect in a glass ribbon according to claim 2, wherein in the calculating step, the side parallel to the direction corresponding to the conveying direction of the glass ribbon is connected from the two overlapping images The length of the diameter of the defect corresponding to the transport direction is subtracted from the length of the corresponding number of pixels, and the moving distance is calculated. 4. The method for measuring the defect in the glass ribbon according to claim 2, wherein In the calculation step, based on the positional relationship of the two superimposed images, the feature value of the defect is calculated by using a predetermined equation including the position of the image in the width direction of the two glass ribbon as a variable, and the feature value is used to calculate The above-mentioned moving distance is as follows: 5. The method for measuring a defect in a glass ribbon according to claim 4, wherein the characteristic value is a length of a diameter of a defect parallel to a conveying direction of the glass ribbon, by overlapping from the above two In the length of the real space corresponding to the number of pixels in the rectangle parallel to the direction corresponding to the direction of the transport direction The method of measuring the defect in the glass ribbon according to claim 4, wherein the characteristic value is a ratio of two diameters of the defect, which is equivalent to the photographing mechanism The moving distance is calculated by the angle between the line in the image of the image capturing position in the front direction and the line passing through the respective centers of the two images, and the above-described ratio. The glass tape inner defect measuring system is characterized in that The conveying mechanism includes: a glass ribbon to be measured in a direction in which the conveyance defect is in a position; a light source that emits light to the glass ribbon; and an imaging mechanism that is disposed at a position where the light reflected by the glass ribbon reaches a position The glass ribbon; and the arithmetic unit, wherein the two images of the shape of the same defect caused by the same defect of the glass ribbon in the image captured by the imaging means and the shape of the defect are different The relationship between the height direction directions of the defects in the glass ribbon is calculated. The glass ribbon inner defect measuring system of π item 7 wherein the computing machine Calculating the moving distance of the defect from the shooting position of one of the two overlapping images to the shooting position of another I60767.doc 201233993, and calculating the moving distance and the light in the glass ribbon The refractive index is calculated to calculate the height direction position of the defect in the glass ribbon. The glass ribbon inner defect measuring system of claim 7, wherein the computing mechanism is connected to the rectangle from the two overlapping images The length of the diameter in the actual space corresponding to the number of pixels in which the direction of the direction in which the glass ribbon is conveyed is the same as the length of the defect parallel to the transport direction, and the moving distance is calculated. The glass ribbon inner defect measuring system according to Item 7, wherein the calculating means calculates the characteristics of the defect by using a predetermined formula including the position of the image in the width direction of the glass ribbon as a variable based on the positional relationship of the two overlapping images. The value is used to calculate the moving distance using the feature value. 160767.doc