201226888 六、發明說明： 【發明所屬之技術領域】 本發明係關於一種玻璃基板。 【先則技術】 先前提出有各種測定玻璃基板内之缺陷之高度方向位置 等之方法。 作為測定玻璃基板内之缺陷之高度方向位置 法’存在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。方便起見將該方法稱為第丨測定方法。圖 13 A係模式性地表示第丨測定方法之說明圖。第丨測定方法 中，如圖13A所示，於使光通過玻璃基板82之狀態下搬送 玻璃基板82。而且，以線型相機81拍攝所搬送之玻璃基板 82之内部。只要於玻璃基板82之内部存在缺陷以，缺陷 就會被拍攝到。圖133係表示所拍攝之缺陷之圖像之例。 圖13A中，將缺陷83模式性地以長方形表示，圖nB中亦 以長方形表示玻璃基板之圖像内所表現之缺陷之像％，但 缺陷之形狀並不限定於長方形 '然而，以下所示之圖 圖14B H15A、圖15B、圖16、圖π中亦將缺陷模 式地以長方形表卜再者，圖13B所示之f頭為玻璃基板 82之搬送方向。於以線型相機81拍攝玻璃基板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 以測疋自線型相機81至缺陷為止之絕對距離，並根據該距 離計算缺陷之高度方向位置。作為㈣相機之焦點使缺陷 之存在位置與相機之焦點—致之方法，存在卿⑽㈣ 160610.doc 201226888 from Focus，聚焦測深）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缺陷之高度方向位置之方法及裝 置例如記載於專利文獻1〜3等中。 又，作為測定玻璃基板内之缺陷之高度方向位置之其他 普通方法而存在如下方法’即利用人射至玻璃基板之光之 反射光在兩個位置拍攝同一缺陷，根據其結果所獲得之兩 個像之位置關係而測定缺陷之高度方向位置。方便起見將 該方法記作第2測定方法。圖14Α係模式性地表示第2測定 方法之說明圖。第2測定方法中，例如圖14Α所示，在玻璃 基板82中，使光入射至與線型相機81相同之側，其反射光 到達線型相機81。而且，搬送玻璃基板82，以線型相機81 拍攝玻璃基板82之内部》至於玻璃基板内之光之路徑，參 照圖16之上段所示之側視圖稍後闡述。缺陷83與玻璃基板 82之搬送一併移動，於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，在線型相機81中作為像而被捕 獲。其結果為，即便有一個缺陷83亦於所拍攝之圖像中顯 現出2個像。圖14Β係以第2測定方法所拍攝之圖像之例。 如圖14Β所示，對同一缺陷顯現出2個像84、85。第2測定 方法中，根據圖14Β所例示之圖像中之2個像之位置關係計 算缺陷83之高度方向位置。又，至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再者，圖14Β所 示之箭頭係玻璃基板82之搬送方向。 利用入射至透明基板等之光之反射光在兩個位置拍攝同 160610.doc 201226888 一缺陷，根據兩個像之位置關係測定缺陷之高度方向位置 之方法及裝置例如記載於專利文獻4〜6、8等。 又，存在如下方法，即於玻璃基板之兩表面，與第2測 定方法相同地拍攝圖像，根據於玻璃基板之各個面上所拍 攝之圖像中之像之位置關係而測定缺陷之高度方向位置。 方便起見將該方法記作第3測定方法。圖丨5 A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15A所 示，於玻璃基板82中，使光入射至與第^線型相機8込相同 之侧，其反射光到達第1線型相機81a。同樣地，使光入射 至與第2線型相機81 同之側，其反射光到達第2線型相 機81b。而且，搬送玻璃基板82，以第丨線型相機81a及第2 線型相機81b分別拍攝玻璃基板82之内部。如此一來，於 第1線型相機8la中與第2測定方法之情形相同地捕獲到兩 個像。又’於第2線型相機81b中亦捕獲到兩個像。圖15B 係以第3測定方法所拍攝之圖像之例。第3測定方法中，如 圖15B所示，獲得一方之線型相機自玻璃基板之上側拍攝 之圖像、與另一方之線型相機自玻璃基板之下側拍攝之圖 像。於各圖像中，分別顯現出兩個像。第3測定方法中， 根據自玻璃基板之上側及下側拍攝之各圖像中之像的位置 關係而計算缺陷83之高度方向位置。再者，圖丨5B中，例 示自上側拍攝之圖像中像重疊之情形。又，至於缺陷之尺 寸’對所拍攝之圖像進行圖像處理而測定缺陷之尺寸。再 者’圖15B所示之箭頭為玻璃基板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 I60610.doc 201226888 置之方法例如記載於專利文獻7。 第2測定方法及第3測定方法中，以同一缺陷之像於圖像 内未重疊作為條件而計算缺陷之高度方向位置。再者，第 3測定方法中’如圖15B所例示於一方之圖像中像重疊之情 形時’使用另一方之圖像計算缺陷之高度方向位置即可。 以下’表示第2測定方法中根據所拍攝之圖像中之兩個 像之位置關係而測定缺陷之高度方向位置之具體例。圖16 係表示所搬送之玻璃基板内之缺陷被線型相機拍攝時之位 置的說明圖。圖16之上段所示之圖為玻璃基板之側視圖， 圖16之下段之左側所示之圖為與圖丨6之上段所示之侧視圖 對應之俯視圖。又，圖16之下段之右侧所示之圖表示拍攝 所搬送之玻璃基板82内之一個缺陷83時所獲得之圖像。 圖16所示之側視圊及俯視圖内所示之長方形表示玻璃基 板82内之缺陷83。本例中，缺陷為丨個^ 1個缺陷83與所搬 送之玻璃基板82 —併移動。圖16所示之側視圖及俯視圖 中’分別圖示移動至位置91時之缺陷83與移動至位置92時 之缺陷83。於圖16所示之側視圖及俯視圖中，缺陷本身並 非存在兩個。 如圖1 6之上段之侧視圖所示，到達線型相機8丨之光自玻 璃基板82中之線型相機側之面入射至所搬送之玻璃基板 82。而且’若所入射之光到達玻璃基板82中之入射側之相 反側之界面，則於該界面反射，通過入射側之界面而到達 線型相機81 ^到達線型相機81之光之入射角α依賴於線型 相機81之設置位置。藉由固定線型相機8丨之設置位置，可 I60610.doc 201226888 決定入射角α為固定值。又，光之折射角β依賴於光之入射 角α及玻璃基板82之折射率η而決定。此處，入射角α及折 射率η為既知，折射角β亦被決定為固定值。折射率11、入 射角α及折射角β成立式（1)之關係。 n=sina/sinp 式⑴ 因此’只要入射角α及折射率n為既知，則藉由關於p解 出式（1)而求出折射角β » 又，於圖16所示之例中，自玻璃基板μ中之與線型相機 8 1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。 線型相機81持續拍攝玻璃基板82之内部。缺陷83與玻璃 基板82—併於搬送方向移動。而且，若缺陷以移動至與入 射至玻璃基板82並於界面反射之後到達線型相機81之光之 路徑的最初之交又位置91，則線型相機81拍攝第丨個像（以 下記作第1像）作為缺陷83之像。進而，若缺陷幻移動至第 2-人與光之路徑父叉位置92，則線型相機8丨拍攝第2個像 (以下s己作第2像）作為缺陷83之像。其結果為，如圖16之下 段之右側所示，於所拍攝之圖像中出現第“㈣及第2像 99 » 再者’於缺陷83為透光性之情形時，透過缺陷83之光到 達線型相機81而作為像而被捕獲。於缺陷83為遮光性之缺 陷之情形時’缺陷83作為黑色之像而顯現於圖像。缺陷Μ 不涂是否為遮光性 捕獲。 於移動至位置91、92時均作為像而被 160610.doc 201226888 又’如圖16所示，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yd。又，將線型相 機81之正面方向之拍攝位置之連線稱為中央線95»更具體 而言’將線型相機81之正面方向之拍攝位置之連線正投影 至玻璃基板82之界面而獲得之直線為中央線95。yd可根據 於所拍攝之圖像（參照圖16之下段之右側）中第1像％及第2 像99正投影至相當於中央線95之圖像内之線96時的像98、 99之距離來測定。 根據圖像。若測定出yd之值，則使用折射角β，計算如 下所示之式（2) ’藉此可求出缺陷83之高度方向位置 d=yd/(2 · tanP) 式（2) 又’將使自線型相機81朝向第1像之拍攝位置91之直線 正投影至玻璃基板之界面而成的直線與中央線95所成的角 設為Θ。此時’所拍攝之圖像（參照圖丨6之下段之右側）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，tane可以如下方式算出。以下，對圖16 之下#又之左側之俯視圖所示之yc進行說明，並對tanQ之計 算進行說明。 圖16中’表示缺陷83自線型相機81之正面偏移之情形。 如圖17所示，於假設缺陷83存在於線型相機81之正面之情 形時，將使拍攝有第2像之位置92正投影至玻璃基板82之 界面之位置、與使線型相機81之透鏡部分正投影至玻璃基 板82之界面之位置的距離稱為拍攝距離ye。其中，拍攝距 離yc根據缺陷83之高度方向位置d而變化。使d為最大時， 160610.doc 201226888 拍攝距離設為最小值yi ’使d為最小時，拍攝距離ye設為最 大值丫2(參照圖1 7之上段所示之側視圖）。即，y i $ yc $ y2。 如此，嚴格而言ye依賴於d ’但yc例如亦可於$ yc ^ y2之 範圍預先決定。y。即便不為準確之值，只要為於 y 1 S yc S y2之範圍之值，tane中只能包含可無視之誤差。 又，將缺陷83自線型相機之正面方向之偏移記作Xec(參 照圖16之下段之左側）。Xcc可根據所拍攝之圖像（參照圖i 6 之下段之右側）中自相當於中央線95之線96至第2像99為止 之距離而確定。即，於圖像内，計數相當於自線96至第2 像99為止之距離之像素數。由於線型相機81之位置固定， 故而每個像素之實際空間中之距離亦作為固定值而決定。 藉由用每個像素之實際空間中之距離乘以相當於自線96至 第2像99為止之距離之像素數，可算出Xec之長度。 此處’ t_可使用及Xcc以如下之式（3)所示以近似式表 示即，tan0可使用yc及xcc藉由式（3)之計算而求出。 [數1] tanG： yd+yc yc 式（3) 又，專利文獻8中記載有如下方法，即一邊使玻璃板移 動-邊使光人射至玻璃板，由該人射光及反射光檢測缺 陷，從而運算缺陷之高度方向位置。專利文獻8中所記载 之方法中’於檢測缺陷之圖案之情形時，當於破璃板之移 動方向不存在幾乎㈣之大小之圖案時，即於靠近破璃板 之背面存在缺陷之情形或缺陷較大之情形時，將該缺陷之 J60610.doc 201226888 高度方向位置判斷為〇。因此，專利文獻8中所記載之方法 中，於上述之情形時，無法準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2〇〇4_361384號公報 [專利文獻3]曰本專利特開2〇〇8·76〇71號公報 [專利文獻4]日本專利第292〇〇56號公報 [專利文獻5]日本專利特開平9_61139號公報 [專利文獻6]曰本特表2003_508786號公報 [專利文獻7]國際公開編號w〇 2006/057125 [專利文獻8]日本專利特開2010_8177號公報 【發明内容】 [發明所欲解決之問題] 較佳為於玻璃基板之表面不存在缺陷所導致之隆起。例 如’可列舉氣泡作為玻螭基板内之缺陷之例。若氣泡位於 玻璃基板之表面之附近，則會有於玻璃基板之表面產生隆 起之類的問題。 例如，若將如此於表面存在隆起之玻璃基板用作液晶顯 示面板中之透明基板，則由於該隆起而導致單元間隙不均 尤其於顯不立體圖像（三維圖像）之液晶顯不面板之情 形時，由於處理左眼用圖像與右眼用圖像兩種圖像，故而 與顯示二維圖像之液晶顯示面板相比，所處理之影像資訊 160610.doc •10· 201226888 量為2倍。而且，必需高速切換左眼用圖像與右眼用圖 像’從而必需將單元間隙設為狹窄間隙。因此，於顯示立 體圖像（三維圖像）之液晶顯示面板之情形時，更嚴格追求 單元間隙之均一性，亦不允許先前所允許之存在於玻璃基 板之表面附近之氣泡所引起之玻璃基板表面之輕微隆起。 又’若表面之隆起到達某極限以上，則重疊玻璃基板 時’負重集中於該隆起部分而成為斷裂之原因。 因此’較佳為於液晶顯示面板所使用之玻璃基板中，於 至少一方之側之表面（液晶側之表面）不存在隆起。 因此，本發明之目的在於提供一種至少一方之表面不隆 起之玻璃基板。 [解決問題之技術手段] 本發明之玻璃基板之特徵在於，將玻璃基板之板厚設為 Τ(μπι) ’將自該玻璃基板之表面至存在於該玻璃基板内之 氣泡為止之距離設為ϋ(μιη)，將氣泡之球換算直徑設為 e(gm)時’存在於至少自一方之表面起在τ/2(μηι)以内之層 中之氣泡之球換算直徑e滿足eg 0.01χΕ>ι 6+15。本發明之 玻璃基板之板厚Τ(μιη)並無特別限定，但於在玻璃基板内 存在氣泡之情形時’玻璃基板之板厚Τ(μπι)越薄，自玻璃 基板之表面至存在於玻璃基板内之氣泡為止之距離Dbm) 越小’玻璃基板表面隆起之可能性越高，因此較佳為i 〇 μιη以上700 μιη以下’更佳為1〇 μηι以上4〇〇 μηι以下，進而 佳為10 μιη以上100 μιη以下，尤佳為1〇 μιη以上50 μηι以 下。 例如，本發明之玻璃基板藉由如下玻璃基板檢查方法而 1606I0.doc 201226888 判定為存在於至少自表面起在Τ/2(μπι)以内之層中之氣泡 之球換算直徑e滿足e$〇.〇lxDi‘6+15，該玻璃基板檢查方 法包含：拍攝步驟’自光源（例如光源2)向沿著條紋方向而 搬送之玻璃基板照射光，藉由配置於在上述玻璃基板反射 之光所到達之位置之拍攝機構（例如線型相機3)而拍攝上述 玻璃基板；運算步驟，根據由拍攝機構拍攝之圖像内之玻 璃基板内之同一氣泡所引起之兩個重合之橢圓形之像的位 置關係而算出玻璃基板内之氣泡之高度方向位置；球換算 直徑算出步驟’算出氣泡之球換算直徑e ;及判定步驟， 判定於藉由氣泡之高度方向位置而決定之玻璃基板之表面 至氣泡為止之距離D與氣泡之球換算直徑e之間，是否滿足 eg 0.01 xD1 6+15。 又’例如本發明之玻璃基板藉由如下玻璃基板檢查方法 而判定為存在於至少自表面起在Τ/2(μπι)以内之層中之氣 泡之球換算直徑6滿足6$〇.〇1><〇1.6+15，該玻璃基板檢查 方法於運算步驟中，計算自同一氣泡所引起之兩個重合之 像（例如像21、22)之外切矩形中之與相當於玻璃基板之搬 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h)減去與搬送方向平行之氣泡之直徑之長度（例如 s)而得之值，藉由所計算之該值與玻璃基板内之光之折射 角算出玻璃基板内之氣泡之高度方向位置，且包含·藉由由 拍攝機構拍攝之圖像而算出與搬送方向正交之方向之氣泡 之直徑之長度之步騾，於球換算直徑算出步驟，將與搬送 方向平行之氣泡之直徑之長度設為3(μιη)，將與搬送方向正 160610.doc -12- 201226888 交之方向之氣泡之直徑之長度設為t(pm)時計算（sxt2)l/3， 藉此算出該氣泡之球換算直徑e，於判定步驟，判定藉由 氣泡之高度方向位置而決定之自玻璃基板之表面至氣泡為 止之距離D與氣泡之球換算直徑e之間是否滿足 eg 0.01 xD16+15。 又’例如本發明之玻璃基板藉由如下玻璃基板檢查方法 而判定為存在於至少自表面起在Τ/2(μπι)以内之層中之氣 /包之球換算直徑e滿足XD16+i5，該玻璃基板檢查 方法於運算步驟’根據同一氣泡所引起之兩個重合之，像之 位置關係，使用包含將與搬送方向正交之玻璃基板之寬度 方向上之像之位置作為變數（例如變數u)之預先決定之計算 式（例如式（6)或式（7))，算出氣泡之特徵量（例如8或r)，使 用該特徵量而計算自兩個重合之像之外切矩形中之與相當 於玻璃基板之搬送方向之方向平行之邊的像素數所對應之 貫際空間中之長度減去與搬送方向平行之氣泡之直徑之長 度的值，藉由所計算之該值與上述玻璃基板内之光之折射 角而算出玻璃基板内之氣泡之高度方向位置，且包含藉由 由拍攝機構拍攝之圖像而算出與搬送方向正交之方向之氣 泡之直徑之長度之步驟，於球換算直徑算出步驟，將與搬 送方向平行之氣泡之直控之長度設為s(^m)，將與搬送方 向正交之方向之氣泡之直徑之長度設為Κμιη)時計算 (s t ) ’藉此算出該氣泡之球換算直徑e，於判定步驟， 判定藉由氣泡之高度方向位置而決定之自玻璃基板之表面 至氣泡為止之距離D與氣泡之球換算直徑e之間是否滿足 160610.doc -13- 201226888 e$ 〇.〇 1 xd1.6. 15 〇 又例如本發明之玻璃基板藉由如下玻璃基板檢查方法 而判定為存在於至少自表面起在τ/2(μηι)以内之層中之氣 /包之球換算直徑6滿足e$〇 〇ΐχ〇1$+ΐ5，該玻璃基板檢查 方法於運算步驟，根據同一氣泡所引起之兩個重合之像之 位置關係’使用預先決定之計算式（例如式⑹），計算與搬 送方向平行之氣泡之直徑之長度（例如s)作為特徵量，計算 自兩個重合之像之外切矩形中之與相當於搬送方向之方向 平仃之邊之像素數所對應之實際空間中之長度減去該直徑 之長度而得之值，藉由所計算之該值與玻璃基板内之光之 折射角而算出玻璃基板内之氣泡之高度方向位置，且包含 藉由由拍攝機構拍攝之圖像而算出與搬送方向正交之方向 之氣泡之直徑之長度之步驟，於球換算直徑算出步驟，將 與搬送方向平行之氣泡之直徑之長度設為s(pm)，將與搬 送方向正交之方向之氣泡之直徑之長度設為ί(μιη)時計算 (sxt ) /3 ’藉此算出該氣泡之球換算直徑^，於判定步驟， 判定藉由氣泡之高度方向位置而決定之自玻璃基板之表面 至氣泡為止之距離D與氣泡之球換算直徑e之間是否滿足 eg 0.01xD16+15。 又’例如本發明之玻璃基板藉由如下玻璃基板檢查方法 而判定為存在於至少自表面起在Τ/2(μϊη)以内之層中之氣泡 之球換算直徑e滿足e$〇.〇lxD16+15，該玻璃基板檢查方法 於運算步驟，使用預先決定之計算式（例如式（7))，根據同 一氣泡所引起之兩個重合之像之位置關係，算出氣泡之 160610.doc •14- 201226888 兩個直徑之比（例如!)作為特徵量，藉由相當於拍攝機構之 正面方向之拍攝位置之圖像内之線與通過兩個像之各中心 之線所成之角及上述比，計算自兩個重合之像之外切矩形 中之與相當於玻璃基板之搬送方向之方向平行之邊之像素 數所對應的實際空間中之長度減去與搬送方向平行之氣泡 之直徑之長度而得之值’藉由所計算之上述值與玻璃基板 内之光之折射角而算出玻璃基板内之氣泡之高度方向位 置，且包含藉由由拍攝機構拍攝之圖像而算出與搬送方向 正交之方向之氣泡之直徑之長度之步驟，於球換算直徑算 出步驟，將與搬送方向平行之氣泡之直徑之長度設為 s(pm) ’將與搬送方向正交之方向之氣泡之直徑之長度設 為Κμηι)時計算（Sxt2)1/3，藉此算出該氣泡之球換算直徑e， 於判定步驟，判定藉由氣泡之高度方向位置而決定之自玻 璃基板之表面至氣泡為止之距離D與氣泡之球換算直徑6之 間是否滿足eg 0.01 xD16+15。 又，例如本發明之玻璃基板亦可為自以浮式法製造之玻 璃帶進行板狀裁切而成之玻璃基板，自相當於玻璃帶之底 面之表面起存在於Τ/2(μηι)以内之層中之氣泡之球換算直 徑 e滿足 〇.〇ixd16+15。 又’例如本發明之玻璃基板亦可為液晶顯示面板之玻璃 基板，自朝向液晶側之表面起存在於Τ/2(μηι)以内之層中 之氣泡之球換算直徑e滿足〇.〇1xD16+15。 [發明之效果] 根據本發明之玻璃基板，可防止至少一方之側之表面中 160610.doc 15 201226888 之隆起。 【實施方式】 以下，參照圖式對本發明之實施形態進行說明。 圖1係表示本發明之玻璃基板之側視圖之例之說明圖。 本發明之玻璃基板5 1為滿足以下條件之玻璃基板。即，本 發明之玻璃基板51中，於將玻璃基板之板厚度設為 Τ(μπι)，將自該玻璃基板之表面至存在於該玻璃基板内之 氣泡為止之距離設為D(pm) ’將該氣泡之球換算直徑設為 ε(μπι)時’自玻璃基板之兩個表面中之至少一方之表面起 存在於Τ/2(μπι)以内之層中之氣泡之球換算直徑6滿足如下 條件，即滿足如下所示之式（4)。 0.01 xD1,6+15 式（4) 此處，DST/2。具體而言，自玻璃基板之表面至氣泡為 止之距離係將接近該氣泡之表面定為基準者。圖1所示之 例中，所謂自玻璃之表面至氣泡57為止之距離D，係將成 為玻璃基板之主面之兩個表面52、53中之接近氣泡57之表 面52作為基準之情形之自表面52至氣泡57為止之距離。 又’氧泡為玻璃基板或玻璃帶内之缺陷之一種。圖i 中，為了表示球換算直徑e而方便起見將氣泡57以球形圖 示，但貫際之氣泡為接近以橢圓之長軸為中心使橢圓旋轉 之旋轉橢圓體的形狀》因此，玻璃内之氣泡可看作以橢圓 之長軸為中心使橢圓旋轉之旋轉橢圓體。又，將該橢圓之 短軸之長度設為t(pm)，長軸之長度設為“#…。圖2係表示 此種氣泡之形狀之說明圖。又，圖3係表示自上方觀察此 160610.doc •16· 201226888 種氣泡之狀態之說明圖。如圖2所示，氣泡之高度與氣泡 之寬度可看作共通之值而均為卜又’氣泡之長度為與橢 圓之長軸相等之值而為S。 若將該氣泡之球換算直徑設為6(μιη)，則球換算直徑e藉 由如下所示之式（5)之計算而求出。 e=(Sxt2),/3 式（5) 即，球換算直徑e為（sxt2)之立方根。 例如，玻璃基板51為用作液晶顯示面板之透明基板之玻 璃基板。於此情形時，將成為玻璃基板51之主面之兩個表 面中之至少朝向液晶側之表面作為基準，並將自其表面至 氣泡為止之距離設為〇(μιη)Β，於該氣泡之球換算直徑e與 距離D之間’只要式（4)成立即可。其中，該氣泡係存在於 自朝向液晶側之表面Τ/2(μιη)以内之層的氣泡，τ/2。 再者，即便於將玻璃基板之另一方之表面作為基準之情形 時，於球換算直徑e與距離D之間相同之關係亦可成立。再 者，所謂成為玻璃基板之主面之兩個表面中之朝向液晶側 之表面，亦可說為例如配置有透明電極之面。 因此，圖1所示之玻璃基板51係用於液晶顯示面板之玻 璃基板’若表面52為朝向液晶側之面，則只要將表面52作 為基準而測定自表面至氣泡為止之距離D即可。 又，自以浮式法製造之玻璃帶對玻璃基板進行板狀裁切 而製造用於液晶顯示面板之玻璃基板之情形時，研磨相當 於玻璃帶之底面之面，作為將該面朝向液晶側之構成而製 造液晶顯示面板。因此，自以浮式法製造之玻璃帶對玻璃 160610,doc •17- 201226888 基板進行板狀裁切而製造用於液晶顯示面板之玻璃基板之 情形時’將成為玻璃基板51之主面之兩個表面中之至少相 當於玻璃帶之底面之表面作為基準，將自該底面至氣泡為 止之距離設為D(pm)時，於該氣泡之球換算直徑e與距離d 之間，只要式（4)成立即可。其中，該氣泡為存在於相當於 自底面之側之表面Τ/2(μιη)以内之層中之氣泡，Dg T/2。 再者，即便於將相當於玻璃帶之上部面之面作為基準之情 形時，於球換算直徑e與距離D之間相同之關係亦可成立。 再者，將以浮式法製造之玻璃帶之下側之面稱為底面，將 上側之面稱為上部面。 因此，圖1所示之玻璃基板51為自以浮式法製造之玻璃 帶進行板狀裁切而成之玻璃基板，若表面52為相當於底面 之面’則只要將表面52作為基準而測定自表面至氣泡為止 之距離D即可。 有關自玻璃基板之表面（此處設為圖1所示之表面52)起 存在於Τ/2(μιη)以内之層中之氣泡，自表面至氣泡為止之 距離D與氣泡之球換算直徑e之間式（4)成立係指越接近表 面52之氣泡球換算直徑越小。換言之，於表面之附近， 不存在球換算直徑之較大之氣泡。因此，可防止氣泡之影 曰所引起之表面52之隆起，從而可提高玻璃基板之品質。 又，本發明之玻璃基板51中’由於可如此防止表面52之隆 起，故而於用作液晶顯示面板中之透明基板之情形時可 使單元間隙均一。 再者，於將自以浮式法製造之玻璃帶板狀裁切而成之玻 160610.doc 201226888 璃基板用作液晶顯示面板之情形時，研磨相當於玻璃帶之 底面之面’但本發明之玻璃基板51係亦可為研磨之前之表 面52作為基準、於自表面52至氣泡為止之距離D(其中 D $ T/2)與氣泡之球換算直徑e之間式（4)成立之玻璃基板。 又’於藉由浮式法等而製造之玻璃帶或自該玻璃帶狀裁 切而成之玻璃基板’產生沿著玻璃帶之主要延伸方向之條 紋。作為玻璃帶之主要延伸方向’並非藉由導引構件所實 現之向玻璃帶之寬度方向之延伸，而係指沿著玻璃帶之前 進方向之延伸之方向。以下，將玻璃帶之主要延伸方向僅 記作玻璃帶之延伸方向。所謂條紋係指由於與玻璃帶之延 伸方向垂直之方向上之板厚之變動及起伏而於玻璃帶之延 伸方向所產生之條紋。於自玻璃帶進行板狀裁切而成之玻 璃基板亦產生條紋。又，由於玻璃帶之延伸方向與自玻璃 f製造裝置（未圖示）將玻璃帶送出之前進方向相同，故而 製造時送出之玻璃帶之前進方向條紋方向與玻璃帶之延伸 方向均為相同之方向。 以下，對測定玻璃基板中之自表面至氣泡為止之距離 D，或算出氣泡之球換算直徑e，檢查式（4)是否成立之玻 璃基板檢查方法十所使用之檢查系統之例或玻璃基板檢查 方法進行說明。若為藉由該玻璃基板檢查方法而判定為式 (4)成立之玻璃基板，則符合本發明之玻璃基板。 该玻璃基板檢查方法令，將玻璃基板之各種缺陷中之氣 泡作為對象，算出自玻璃基板之表面至氣泡為止之距離 D。玻璃帶或玻璃基板内之氣泡為橢圓體。因此，拍攝玻 160610.doc •19· 201226888 璃基板内之氣泡而獲得之圖像中’氣泡之像成為橢圓。 又，作為圖像而拍攝之氣泡之像（橢圓形之像）中，中心部 分為白色。因此，顯現於圖像之氣泡之像之中心部分可作 為特徵性之點（以下記作特徵點）而利用。 首先，關於玻璃基板，對檢查式（4)是否成立之破璃基 板檢查方法中之第1玻璃基板檢查方法進行說明。圖4係檢 查玻璃基板中之自表面至氣泡為止之距離D與氣泡之球換 算直徑e之間式（4)是否成立的檢查系統之構成例之模式 圖。該檢查系統包含搬送輥1、光源2、線型相機3及運算 裝置4。 搬送輥1支撐成為檢查對象之玻璃基板5，將玻璃基板5 向固定方向以固定速度搬送。將玻璃基板5向沿著玻璃基 板5本身之條紋方向之方向搬送。因此，搬送輥1之玻璃基 板5之搬送方向為與玻璃基板5之條紋方向相同之方向。 又’本例中以下述情形為例，即使玻璃基板之兩個表面中 之成為決定至氣泡為止之距離之基準之表面（設為圖1中之 表面52)朝向光源2及線型相機3之相反側，將玻璃基板5支 揮於搬送輥1。例如’玻璃基板5為自以浮式法製造之玻璃 帶板狀裁切而成之玻璃基板之情形時，只要使相當於玻璃 帶之底面之面朝向光源2及線型相機3之相反側，使玻璃基 板5支標於搬送輥1即可。又，若玻璃基板$用作液晶顯示 面板中之透明基板，則只要使朝向液晶側之面朝向光源2 及線型相機3之相反側，使玻璃基板5支撐於搬送輥1即 可°而且’該玻璃基板檢查方法中，測定自玻璃基板5中 160610.doc .20· 201226888 之搬送輥1側之表面52至氣泡為止之高度方向位置（距離）。 此處，高度方向位置為自搬送輥丨側之表面至氣泡為止之 距離。因此，使成為基準之表面朝向搬送輥1側之情形 時，高度方向位置之測定值意味著自成為基準之表面至氣 泡為止之距離D。 再者，亦可使成為決定至氣泡為止之距離之基準之表面 52朝向搬送報1之相反側而使玻璃基板5支撐於搬送輥1。 於此情形時’自成為基準之表面至氣泡為止之距離D為由 玻璃基板5之板厚T減去高度方向位置之測定值而成之值。 板厚T為既知，不論使成為決定至氣泡為止之距離之基準 之表面52朝向哪一側之情形時，自成為基準之表面至氣泡 為止之距離D均根據氣泡之高度方向位置之測定值而決 定。 如已說明般’此處以如下情形為例，即使成為決定至氣 泡為止之距離之基準之表面52朝向光源2及線型相機3之相 反側（即搬送輥1側）而使玻璃基板5支撐於搬送輥1。 光源2配置於玻璃基板5之兩個面中之一面之側，並朝向 玻璃基板5照射光。該光自界面8入射至玻璃基板5，通過 玻璃基板内而由入射側之相反側之面52反射。經反射之光 通過入射側之界面8而到達線型相機3。再者，圖4中將光 之路徑簡化而表示，如圖16之上段之側視圖所示，光之路 徑於光入射至界面8時及界面52中之反射後通過界面8時， 分別折射。 線型相機3配置於自光源2照射而由玻璃基板5反射之光 I606I0.doc •21 - 201226888 所到達之位置。具體而言，將玻璃基板5作為基準，.配置 於與光源2相同之側。又，例如線型相機3係將光源2作為 基準，配置於玻璃基板5之搬送方向。而且，線型相機3拍 攝玻璃基板5之内部’生成圖像作為拍攝結果。 藉由決定光源2及線型相機3之配置位置，光之路彳⑥中入 射角α(參照圖16之上段）亦決定為固定值。進而，玻璃基 板5之折射率η亦為既知，藉由解式〇)，將自光源2至線$ 相機3為止之光之路徑中之折射角ρ之值亦決定為固定值。· 玻璃基板5被搬送，線型相機3於固定位置繼續進行玻璃 基板5之拍攝。因此，隨著時間經過而於玻璃基板5中所拍 攝之部位發生變化。因此，若將線型相機3之正面方向之 拍攝位置之連線正投影至玻璃基板5之界面8，則表示為直 線。將該直線稱為巾央線。圖5八係表示中央線之說明圖， 圖5Β係表示相當於圖像内之t央線之線之說明圖。圖μ 係玻璃基板5之俯視圖。伴隨玻璃基板5之搬送，線型相機 3之正面之拍攝位置發生變化，將向該連線之界面之正投 影圖示為中央線95。又’圖5B表示藉由線型相機3而拍攝 之圖像。圖像内中，相當於中央線95之線96以 '點鍵線表 示。該線96可為與線型相機3之正面方向之拍攝位置對應 之像素之連線。又，中央線95與玻璃基板5之搬送方向平 行，相當於中央線95之圖像内之線96可表示圖像内之相當 於玻璃基板5之搬送方向之方向。再者，由於玻璃基心沿 著其條紋方向搬送，圖像内之線96亦可表示相當於條纹方 向之方向。將相當於中央線95之圖像内之線％記作搬送方 160610.doc -22· 201226888 向線再者，圖中’為了說明而圖示搬送方向線96，但 貫際之拍攝圖像中搬送方向線96並不顯現於圖像内。 於在玻璃基板5内存在氣泡之情形時，由於丨個氣泡而於 線型相機3所拍攝之圖像内該氣泡之像顯現兩個。又，拍 攝有氣泡之圖像内所顯現之氣泡之像為橢圓形，其中心部 為白色。 運算裝置4參照藉由線型相機3而拍攝之圖像來測定氣泡 之高度方向位置》該氣泡之高度方向位置為圖16之上段之 側視圖中表示為「d」之長度。即，於玻璃基板5中，為自 光源2之相反側之表面52至氣泡為止之距離。運算裝置4於 拍攝共通之氣泡而獲得之成對之橢圓形之像重合之情形 時，根據該成對之橢圓形之位置關係，算出玻璃基板5内 之氣泡之高度方向位置。具體而t，運算裝置4計算自圖 像内中重合之兩簡圓形之像之外切矩形中之與相當於玻 璃基板之搬送方向之方向平行之邊之像素數所對應之實際 空間中的距離’減去氣泡之直徑中之與搬送方向平行之直 徑之長度的值。再者，力圖像内，與相當於玻璃基板之搬 送方向之方向平行，係指與搬送方向線％(參照圖5b)平 灯運异裝置4由藉由上述減法運算求出之值與玻璃基板$ 中之折射角β而算出氣泡之高度方向位置。關於該計算， 參照圖8稍後闡述。 又’沿著玻璃基板之條紋方向搬送玻璃基板之情形時， 玻璃基板内之氣泡之長軸與搬送輥】之搬送方向(換言之玻 璃基板5之條时向）錢平行。如，氣泡之長軸 1606I0.doc -23· 201226888 72之方向與搬送輥1之玻璃基板5之搬送方向71之偏移最大 為10 °如此’氣泡之長軸72與搬送輥1之搬送方向71大致 平行’因此於線型相機3所拍攝之圖像中，表示為橢圓形 之氣泡之像之長軸與搬送方向線96(參照圖5B)亦大致平 行以下’以所拍攝之圖像中氣泡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’於成對之像未重合之情形時，只要運算裝置4藉 由公知之方法算出氣泡之高度方向位置即可。 線型相機3之配置位置被固定。因此，線型相機3所 拍攝之圖像中之丨個像素所對應之實際空間中之距離亦決 定為固定值。圖像中之像素所對應之實際空間中之距離為 既知者。 圖7係表示關於存在於玻璃基板之氣泡檢查是否滿足式 (4)之條件之玻璃基板檢查方法中之第丨玻璃基板檢查方法 之例的流程圖。 首先，光源2對檢查對象之玻璃基板5間始照射光（步驟 S1)。 而且，搬送輥1將配置於搬送輥丨上之玻璃基板5向固定 方向搬送，線型相機3繼續所搬送之玻璃基板5之内部之拍 攝。此時，玻璃基板5以玻璃基板5本身之條紋方向與搬送 方向相同之方式配置於搬送輥】上，並沿著條紋方向搬 送。而且，線型相機3生成圖像作為拍攝結果（步驟S2)。 線型相機3將藉由拍攝而獲得之圖像發送至運算裝置4。 於在玻璃基板5之内部存在氣泡之情形時，於由步驟s2 J60610.doc -24- 201226888 所獲知之圖像中包含氣泡之像。具體而言，於圖像内顯現 出橢圓形之像作為氣泡之像。又，如圖16所說明般，於氣 泡移動至與反射前之光之路徑重疊之位置(圖16之上段之 側視圖中所示之位置91)時、及於氣泡移動至與反射後之 光之路徑重疊之位置（圖16之上段之侧視时所示之位置 92)時，分別作為像而顯現於圖像。因此，於存在1個氣泡 之情形時，於圖像2中顯現出兩個像。又，於氣泡較大之 凊形時、或氣泡存在於玻璃基板5之表面52(參照圖句之附 近之情形時，該兩個像重合。 運算裝置4若接收由步驟S2生成之圖像，則自圖像中檢 測兩個重合之像之外切矩形之區域。而且，計數該外切矩 形之邊令之於圖像内與相當於玻璃基板5之搬送方向之方 向平行之邊（即與圖像内之搬送方向線96平行之邊）之像素 數。而且，運算裝置4用每個像素之實際空間中之距離乘 ，X邊之像素數，藉此算出與該邊之像素數對應之實際空 間中之長度（步驟S3)。 圖8係表示兩個重合之像之外切矩形之區域之說明圖。 如圖8所示，決定圖8所示之外切矩形以作為重合之兩個像 21、22之外切矩形。像21、22為橢圓，可看作全等。圖8 所示之例中，外切矩形23之長邊與搬送方向線96(參照圖 5B)平行。於此情形時，運算裝置4計數像21、22之外切矩 形23之長邊24之像素數，用每個像素之實際空間中之距離 乘以該像素數。將該長邊24所對應之實際空間中之長度以 「hJ表示。此處，h之單位為μηι。 160610.doc •25· 201226888 又，於缺陷為氣泡之情形時，像21之中心部21a於圖像 上為白色。該中心部21a為像21之特徵點。運算裝置4 十數 自一方之像21之中心部21a至外切矩形23之短邊中之較近 之短邊為止的像素數。即’計數圖8中符號a所示之部分之 像素數。運算裝置4用每個像素之實際空間中之距離乘以 該像素數。該乘法運算結果為相當於圖8所示之八之部分所 對應之實際空間中之長度’具體而言，為與搬送方向平行 之氣泡之直徑（氣泡之直徑中與搬送方向平行之直經）之1/2 之長度。圖8所示之例中，該直徑為氣泡之長軸。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 之氣泡之直徑之長度（步驟S4) ^該氣泡之直徑之長度相當 於圖3所不之s。此處，s之單位為μπι。與實際空間中之sq 之長度對應之圖像内之部位為圖8中符號Α所示之部分。 又，兩個像21、22可看作全等，因此於圖8中，可看成 A=A’。 再者，此處，以使用像21之中心部21a計算s之情形為例 進行說明，但亦可使用像22之中心部而計算s。 又，圖8中，以氣泡之像之長軸與搬送方向線平行之情 形為例進行㈣，但亦存在氣泡之像之長軸與搬送方向線 並不完全平行之情形^然而’玻璃基板内之氣泡之長軸盥 玻璃基板之搬送方向之偏移最大僅為1〇。(參照圖6)。因 此’即便氣泡之像之長㈣搬送方向線不完全平行以可 看作兩者平行而與上述步驟幻、s4同樣地計算h、s。即， 於求出h時，只要計數重合之兩個像之外切矩形之長邊之 160610.doc -26- 201226888 像素數’用每個像素之實際空間中之距離乘以該像素” 可又，求出s時，只要計數自一方之像之中心部至外切 矩形之短邊中之較近之短邊為止之像素數，用每個像辛之 實際空間中之距離乘以該像素數，將該乘法計算結果乘以 2即可。即便氣泡之像之長軸與搬送方向線不完全平行， 如上所述冲算h、s，使用該h、s計算氣泡之高度方向位置 亦僅包含可無視之程度之誤差。X，即便於此情形時，亦 可將S看作氣泡之長軸。 其-人，運算裝置4由在步驟83中所#出之h減去於步驟S4 令所算出之s(步驟S5)。將該講法運算結果設為^為自 拍攝有第1像之位置至拍攝有第2像之位置為止之氣泡之移 動距離gp纟步驟S5算出之力為拍攝有氣泡之像之兩點 間之距離。再老，盘音机& 一貫際二間中之yd之長度對應之圖像内 之部位為圖8中符號B所示之部分。 運算裝置4使用由步驟S5算出之力與預先決定之折射角β 來進行式（2)之計算，從而計算氣泡之高度方向位置^。 即，计算yd/(2 . tanp)，將其計算結果設為d(步驟S6卜氣 泡之高度方向位置d為自玻璃基板5 _之搬送輥1側之表面 52(參照圖句至氣泡為止之距離。 /、-人，運算裝置4根據氣泡之高度方向位置d來決定自成 為基準之表面52至氣泡為止之距離D(步驟S7)。如本例 般，於將成為決定至氣泡為止之距離之基準之表面朝向搬 达輥1側而配置玻璃基板之情形時，自其表面至氣泡為止 之距離D與由步驟36算出之氣泡之高度方向位置廿相等。因 I60610.doc •27- 201226888 此，只要將氣泡之高度方向位置d之值設為自成為基準之 表面52至氣泡為止之距離D即可。即，運算裝置4中，只要 設為D=d，從而決定距離d之值即可。 再者，於將作為決定至氣泡為止之距離之基準之表面朝 向搬送輥1之相反側而配置玻璃基板之情形時，自其表面 至氣泡為止之距離D可藉由由玻璃基板之板厚τ(μιη)減去 氣泡之高度方向位置d之值而獲得。即，於此情形時’運 算裝置4中，只要設為D=T_d而決定距離d之值即可。其 中，玻璃基板之板厚T為既知。 步驟S7之後，運算裝置4根據由步驟S2所拍攝之圖像， 算出與搬送方向正交之方向之氣泡之直徑之長度（步驟 S8)。步驟S8中，利用由步驟S3所檢測之兩個重合像之外 切矩形之區域（參照圖8)。具體而言，運算裝置只要計數自 重。之兩個像t彳之像之中心部至外切矩形之長邊中之 較近之長邊為止的像素數’用每個像素之實際空間中之距 離乘以該像素數，並將該乘法運算結果乘以2即可。該值 為氣泡之寬度’相當於阁1 ^^一 祁田於圖2所不之t。此處，t之單位為 μπ-又，由於氣泡為旋轉橢圓體，故而氣泡之高度與氣 泡之寬度相同為ί(μηι)。 其次’運算裝置4使用由+明 •用由步驟S4算出之與搬送方向平行 氣泡之直徑之長度s、與由步驟以算出之與搬送方向正交 之方向之孔泡之直杈之長度t而算出該氣泡之球換算直徑 e(步驟S9)。運算裝置4 0 二 ，、要藉由進盯式（5)之計算而算出球 換舁直徑e即可。即，谨曾 P運异裝置4藉由計算（sxt2)之立方根 160610.doc •28· 201226888 而算出球換算直徑e。此處，e之單位為μιη。 再者，步驟S3〜S9係分別針對成對之橢圓形之像之組的 每一個而進行。 其次，運算裝置4檢測自玻璃基板5之表面52起之距離1) 為172以下之氣泡。T為玻璃基板5之板厚，而且，運算裝 置4依序選擇該氣泡，對所選擇之氣泡判定所算出之距離D 及球換异直徑e之間式（4)是否成立（步驟si〇)。運算裝置針 對成對之橢圓形之像之組之每一個計算自表面52起之距離 D及球換算直徑e。步驟sl〇中，運算裝置4只要針對距離d 為T/2以下之橢圓形之組之每一個判定為存在丨個氣泡，從 而檢測自表面52起之距離D為T/2以下之氣泡即可。而且， 所依序選擇各個所檢測之氣泡，對所選擇之氣泡判定所算 出之距離D及球換算直徑e之間「eg〇〇lxDi.6+15」之關係 是否成立。對於自表面52起之距離〇為172以下之氣泡之各 者，「eg〇.〇ixDi.6+15」之關係成立之玻璃基板符合本發 明之玻璃基板。另一方面，於自表面52起之距離]〇為丁/2以 下之氣泡中，於存在「eS〇.01xDi.6+15」之關係不成立之 氣泡之情形時’該玻璃基板不符合本發明之玻璃基板。 因此’對本發明之玻璃基板51(參照圖丨）進行上述玻璃 基板檢查方法（圖7所示之步驟S1〜S10)之情形時，對於自 表面52起之距離d為T/2以下之氣泡之各者判定為 「e$0.01xD16+15」之關係成立。 拍攝存在於搬送輥1(參照圖4)側之表面52之附近之氣泡 而成之圖像中’該氣泡所引起之兩個像重合出現。又，即 160610.doc -29- 201226888 便於氣泡較大之情形時’拍攝該氣泡而成之圖像中，兩個 像亦重合出現。參照圖丨4 a而說明之第2測定方法中，同一 缺陷所引起之兩個像重疊之情形時，無法測定該像之高度 方向位置。又，參照圖15 A而說明之第3測定方法中，由於 如圖1 5B所示拍攝自上側起之圖像與自下側起之圖像，只 要因此任意一方之圖像中像不重疊，則可測定缺陷之高度 方向位置。然而，於缺陷較大之情形時，由兩個線型相機 拍攝之各圖像中，像有時亦重疊，於此情形時，無法測定 缺陷之高度方向位置。對此’上述玻璃基板檢查方法（圖7 所示之步驟S1〜S10)中，即便同一氣泡所引起之像重疊亦 可計算氣泡之高度方向位置。因此，可決定自表面52起之 距離D，關於自表面52起之距離1)為172以下之氣泡，從而 可判定該距離D與氣泡之球換算直徑e之間式是否成 立。 又’參照圖14 A而說明之第1測定方法中，缺陷之高度方 向位置之測定結果容易受到所搬送之玻璃基板之上下振動 之影響，但上述步驟S1〜S10所示之玻璃基板檢查方法中， 難以受到此種影響’從而可精度較佳地計算氣泡之高度方 向位置。其結果為，關於自表面52起之距離D為172以下之 氣泡，可精度較佳地判定該距離D與氣泡之球換算直徑e之 間式（4)是否成立。 關於自玻璃基板中之成為基準之表面52起之距離D為τ/2 以下之氣泡’檢查該距離D與氣泡之球換算直徑e之間式 (4)疋否成立之玻璃基板檢查方法並不限定於圖7所示之方 160610.doc •30· 201226888 法（步雜〜S1〇)。以下，對進行相同之檢查之第2玻璃基 板檢查方法及第3玻璃基板檢查方法進行說明。任一情形 時，例如均可使用圖4所例示之檢查系統來檢查。相對於 成為檢查對象之玻璃基板5的光源2及線型相機3之位置關 係與第1玻璃基板檢查方法相同，省略說明。然而，利用 運算裝置4之氣泡之高度方向位置d之測定方法與糾玻璃 基板檢查方法不同。 再者，第2玻璃基板檢查方法及第3玻璃基板檢查方法 中，玻璃基板以向沿著玻璃基板本身之條紋方向之方向搬 送之方式而於搬送輥1配置並搬送。 第2玻璃基板檢查方法中，運算裝置4算出檢查對象之玻 璃基板5内之氣泡之特徵量。而且，運算裝置4使用該特徵 量進行如下計算，即由重合之兩個像之外切矩形中之與相 當於玻璃基板之搬送方向之方向平行之邊之像素數所對應 之實際空間中之長度減去與玻璃基板之搬送方向平行之氣 泡之直徑（氣泡之直徑中之與搬送方向平行之徑）的長度。 又，運算裝置4計算上述特徵量時，根據重合之兩個像之 位置關係使用預先決定之計算式計算特徵量。 又，第2玻璃基板檢查方法巾，計算與玻璃基板之搬送 方向平行之氧泡之直徑之長度作為特徵量。 用以算出上述特徵量之式預先決定作為使將玻璃基板之 端部作為基準之像之特徵點對應的位置之座標、於第！玻 璃基板檢查方法所說明之h、及兩個重之合之像之面積為 變數的函數。用以決定該特徵量（氣泡之直徑令之與搬送 160610.doc -31 - 201226888 方向平行之直徑）之計算式例如可以如下之式（6)表示。 S=aiu2+a2h2+a3p2+a4uh+a5hp+a6up+a7u+a8h+a9p+a10 式（6) 式（6)中’「u」為與將玻璃基板之端作為基準之像之特 徵點對應之位置的座標，具體而言，為自與搬送方向平行 之玻璃基板之側面至氣泡之中心為止之距離。此處，u之 單位為mm»「h」為根據拍攝氣泡而成之圖像藉由與第丄玻 璃基板檢查方法中之步驟S3相同之計算而獲得之值。此 處，h之單位為μπι。P為於拍攝氣泡而成之圖像中兩個像 所占之區域（兩個像之區域之合集）之面積，具體而言由 圖像内之像素數表示。式（6)中之a广ai〇為係數。又，式（6) 中之s為與玻璃基板之搬送方向平行之氣泡之直徑。 成為特徵量之直徑s由於玻璃帶中之寬度方向之氣泡之 位置而容易受到影響。而且，自帶狀之玻璃帶對玻璃基板 進行板狀裁切之位置一般於玻璃帶之寬度方向上固定。例 如，一般若將自玻璃帶之側部至玻璃基板之板狀裁切位置 為止之距離設為X，則使X固定，對玻璃基板群依次進行 板狀裁切。由此，可以說成為特徵量之直徑s由於玻璃基 板中之與搬送方向垂直方向上之氣泡之位置（換言之，玻 璃基板中之與條紋方向垂直之方向上之氣泡之位置）亦容 易觉到影響。因此，將包含上述變數u之計算式（例如上述 式（6))用於s之計算。 又，所拍攝之圖像中’氣泡之像之長軸與搬送方向線平 行之It形時，上述s相當於氣泡之長軸。然而，即便於圖 像中，氣泡之像之長軸與搬送方向線並不完全之情形時， 160610.doc -32· 201226888 由於兩者大致平行， 击 上述特徵量S可看作氣泡之長 軸。即便如此將3看 轧泡之長 程戶之誤# π 長軸，亦可僅包含可無視之 誤差，/不影響氣泡之高度方向位置之算出。 ▲ / ()中之係數…、。藉由最小平方法預先求出。具體而 吕，使用成為樣品之氣泡n w σ p A L u。又，對包含成為樣 °日〇之氣泡之玻璃基板進行與第1破璃基板檢查方法中所說 明之步㈣训目同之處理而獲得h。又，自此時由步驟S2 所獲知之圖像6十數成為兩個像之合集之區域之像素數P。 準備複數個成為樣品之氣泡，對該等各氣泡如此獲得S、 u、h、P。若獲得複數組s、u、h、p之組，則自該等s、 ^”之組中藉由最小平方法求出式⑹中之係數^即 *5J" 〇 S與u、h、p之間具有關聯，藉由最小平方法可求出式（6) 中之各係數。 運异裝置4根據藉由拍攝成為氣泡之高度方向位置之測 定對象之玻璃基板而獲得之圖像求出u、h、p，藉由代入 至式（6)中算出s。而且，運算裝置4計算h_s(=yd)，使用該 計算結果與折射角β算出氣泡之高度方向位置。 圖9係表示第2玻璃基板檢查方法之例之流程圖。關於與 第1玻璃基板檢查方法相同之處理，標註與圖7相同之符 號，並省略說明。 直至步驟S3算出h為止之動作與第1玻璃基板檢查方法相 同。 圖10係表示顯現於圖像内之玻璃基板之例之說明圖。於 160610.doc -33- 201226888 存在氣泡之情形時，於圖像内亦顯現氣泡之像21、22 ^ 又，圖ίο所不之例中，作為像之特徵點，各像21、22之中 心部分2 1 a、22a亦作為白色區域顯現於圖像内。再者，雖 圖示像21、22之外切矩形23 ,但外切矩形23未必顯現於圖 像内。 步驟S3之後，運算裝置4計數圖像内之自玻璃基板之端 部3 1至像之特徵點為止之像素數。即，計數圖丨〇中符號匸 所示之部分之像素數。而且，運算裝置4使用每個像素之 實際空間中之距離乘以該像素數（步驟s丨丨）。該乘法運算結 果相當於實際空間中之自玻璃基板之端部（側面）至氣泡為 止之距離u »即’步驟S11中，算出u。 其中，上述步驟S11之說明中’為了簡化說明，以玻璃 基板之端部3 1映在圖像内之情形為例進行說明。於玻璃基 板之端部3 1不映在圖像内之情形時，只要如下所述計算距 離u即可。因線型相機3之設置位置固定，故而可預先求出 自玻璃基板之端部至藉由線型相機3所拍攝之圖像内之玻 璃基板端部側之端為止的實際空間中之距離（設為u〇)。而 且，運算裝置4 s十舁所拍攝之圖像中之自該端之部分至像 之特徵點為止之距離。該計算中’例如只要計數圖像中之 自該端之部分至特徵點為止之像素數，使用每個像素之實 際空間中之距離乘以該像素數即可。運算裝置4只要藉由 對該距離加上藉由線型相機設置位置而決定之u〇而算出實 際空間中之自玻璃基板之端部（側面）至缺陷為止之距離u即 〇J~ 〇 160610.doc •34· 201226888 再者’圖10所示之例中，以使用像21之中心部分21“乍 為特徵點而求出圖像内之自玻璃基板之端㈣至中心部分 21為止之距離之情形為例。作為特徵點，亦可使用另一 像22之中〜。p分22a。不論使用哪一個中心部分作為 ㈣點’均可求出實際空間中之自玻璃基板之端部（側面） 至氣泡為止之距離u。根據使用中心部分& ' %中之哪一 個作為特徵點而像素數之計數結果不同，但該差甚小，距 離u中僅包含可無視之誤差。又，作為特徵點，亦可使用 外切矩形23内之特徵性之點(例如外切矩形23中任一頂 點）°於此情形時，距離11亦僅包含可無視之誤差。 步驟S11之後，運算裝置4計數該區域内之像素數p作為 重合之兩個像21、22所占之區域（兩個像之區域之合集）之 面積（步驟S12)» 而且，運算裝置4藉由將由步驟S3、Sll、S12求出之h、 P代入式（6)中而什算氣泡之直徑中之與搬送方向平行之 直Μ步驟Sl3)。如圖1〇所示像之長軸與搬送方向線平行 形時。亥直s為氣泡之長軸。如已說明般，即便於 拍攝圖像中’像之長軸與搬送方向線並不完全平行，但由 於兩者幾乎平行’故而步驟⑴所計算之直 之長軸。 项f札心 …吏之處理與第1玻璃基板檢查方法中之步驟S5〜S10相 同。 即，運_算裝置4藉由由在步驟Μ中所算出之匕減去步驟 令所算出之S來求出％(步驟S5)。而且，運算裝置4使用 160610.doc -35- 201226888 yd與折射角β來進行式（2)之計算，從而計算氣泡之高度方 向位置d(步驟S6)。 進而’運算裝置4根據氣泡之高度方向位置d決定自成為 基準之表面52至氣泡為止之距離D(步驟S7) ^於使成為決 定至氣泡為止之距離之基準之表面朝向搬送輥1側而配置 玻璃基板之情形時，只要D=d即可。又，於使成為決定至 氣泡為止之距離之基準之表面朝向搬送輥1之相反側而配 置玻璃基板之情形時，運算裝置4只要設為D=T_d而決定距 離D之值即可。 而且，運算裝置4根據步驟S2所拍攝之圖像算出與搬送 方向正交之方向之氣泡之直徑之長度t(步驟S8)。該t之算 出方法可與第1玻璃基板檢查方法中之步驟S8相同。繼 而’運算裝置4藉由計算（Sxt2)之立方根而算出氣泡之球換 算直徑e(步驟S9)。 再者，步驟S3〜S9之處理係針對成對之橢圓形之像組之 每一個而進行。 進而’運算裝置4檢測自玻璃基板5之表面52起之距離D 為丁/2以下之氣泡。而且，運算裝置4依序選擇該氣泡，對 所選擇之氣泡判定所算出之距離D及球換算直徑e之間式 (4)之關係（即〇 〇lxD16+i5)是否成立（步驟S10)。 於對本發明之玻璃基板51(參照圖1)進行第2玻璃基板檢 查方法（圖9所示之步驟S1〜S 10)之情形時，對於自表面52 起之距離D為T/2以下之氣泡各個亦判定為 「e$〇.〇lxDi.6+15」之關係成立。 160610.doc -36- 201226888 又，圖9所示之第2玻璃基板檢查方法中，與第丨玻璃基 板檢查方法（參照圖7)同樣地即便同—氣泡之像重疊亦可計 算氣泡之咼度方向位置。又，難以受到所搬送之玻璃基板 之上下振動之影響，關於自表面52起之距離D為τ/2以下之 氣泡，可精度較佳地判定該距離D與氣泡之球換算直徑6之 間式（4)是否成立。 其次，對第3玻璃基板檢查方法進行說明。第3玻璃基板 檢查方法中，相對於成為檢查對象之玻璃基板5的光源2及 線型相機3(參照圖4)之位置關係亦與第丨玻璃基板檢查方法 相同，而省略說明。 第3玻璃基板檢查方法中，運算裝置4算出玻璃基板5内 之氣泡之特徵量，使用該特徵量計算〜。然而，第2破璃 基板檢查方法中，算出氣泡之直徑s作為特徵量，但第地 璃基板檢查方法中’計算氣泡之兩個直徑之比。具體而 言’運算裝置4求出氣泡之直徑中之相對於與搬送方向正 父之方向之直徑的搬送方向之直徑之比例作為氣泡之特徵 量。即，若將氣泡之直徑中之與搬送方向正交之方向之直 徑設為r! ’將搬送方向之直徑設為”，則計算咏作為特徵 量。以下將Γ2/Γ1記作r。 再者，所拍攝之圖像中，氣泡之像之長轴與搬送 平行之情形時，上述u相當於氣泡之短徑，_當於氣 泡之長轴。即，計算「長轴/短徑」作為特徵量卜 ， 即便於圖像中，氣泡之像之長軸與搬送方向線並不入巫 行之情形時，由㈣者大致平行，故而可將上氣 1606l0.doc •37- 201226888 泡之短杈，將上述h看作氣泡之長軸。即，即便於圖像中 氣泡之像之長軸與搬送方向線並不完全平行之情形時，亦 可將作為特徵量而計算之r看作氣泡之「長轴/短徑」。即便 如此，r中亦僅包含可無視之程度之誤差，並不影響氣泡 之高度方向位置d之算出。 運算裝置4於算出Γ作為氣泡之特徵量之後，使用該r求 出yd(自拍攝有第丨個像之位置至拍攝有第2個像之位置為止 之氣泡的移動距離）。 又’運算裝置4於計算上述特徵量γ時根據重合之兩個像 之位置關係使用預先決定之計算式計算特徵量。 用以計算該特徵量r之式預先決定作為使將玻璃基板之 端4作為基準之像之特徵點對應之位置之座標、於第i玻 璃基板檢查方法所說明之h、及兩個重合之像之面積為變 數的函數。用以決定特徵量r之計算式例如可以如下之式 (7)表示。 r=blU2+b2h2+b3p2+b4uh+b5hp+b6up+b7u+b8h+b9p+bl〇 式⑺ 該函數中之變數u、h、p與第2玻璃基板檢查方法所示之 式(6)中之變數u、h、p相同。即’「u」為自與搬送方向平 行之玻璃基板之側面至氣泡之中心為止之距離。「心為根 據拍攝氣泡而成之圖像藉由與第丨玻璃基板檢查方^中之 步驟S3相同之計算而獲得之值。P為於拍攝氣^成之圖 像中兩個像所占之區域（兩個像之區域 σ果面積，具 體而言，以圖像内之像素數表示。式 、 數。 之卜〜、為係 J60610.doc •38· 201226888 特徵量r由於玻璃帶中之寬度方向之氣泡之位置而容易 受到影響。而且，如已說明般，自帶狀之玻璃帶對玻璃基 板進行板狀裁切之位置一般於玻璃帶之寬度方向上固定。 由此’可以說特徵量r由於玻璃基板中之與搬送方向垂直 之方向上之氣泡之位置（換言之，玻璃基板中之與條紋方 向垂直之方向上之氣泡之位置）亦容易受到影響。因此， 將包含上述變數u之計算式（例如上述式用於[之計算。 式（7)中之係數bl〜b10藉由最小平方法預先求出。具體而 5，使用成為樣品之氣泡來實測r、u。又，對包含成為樣 之氣泡之玻璃基板進行與第〗玻璃基板檢查方法中所說 明之步驟S1〜S3相同之處理而獲得he又，自此時由步驟“ 獲得之圖像計數成為兩個像之合集之區域之像素數p。準 備複數個成為樣品之氣泡，對該等各氣泡如此獲得r、u、 h P若獲得複數組Γ、u ' h、p之組，則自該等『、u、匕、 p之組中藉由最小平方法求出式⑺中之係數即可。 r’、u h、P之間具有關聯，藉由最小平方法可求出式（？） 中之各係數。 運算裝置4根據藉由拍攝成為氣泡之高度方向位置之測 定對象之玻璃基板而獲得之圖像求出《、h、p，藉由代入 式（7)中算出r。 又’運算裝置4於所拍攝之圖料將搬送方向線96與通 過兩個像之中心之線所稱之角設為叫，求出㈣之值。而 且’運算農置4使用h、u'r、tane計算％。運算裝置4使用 該〜與折射角β而算出氣泡之高度方向位置d。 160610.doc -39- 201226888 圖11係表示第3玻璃基板檢查方法之例之流程圖。關於 與第1玻璃基板檢查方法或第2玻璃基板檢查方法相同之處 理’標註與圖7或圖9相同之符號，並省略說明。 直至由步驟S12求出p為止之動作（步驟S1、S2、S3、 Sll、S12)與第2玻璃基板檢查方法。 步驟S12之後，運算裝置4藉由將由步驟S3、S1丨、si2求 出之h、u、p代入式（7)中而計算Γ(即氣泡之直徑中之相對 於與搬送方向正交之方向之直徑之長度的搬送方向之直徑 之長度的比例）（步驟S21)。 圖12係表示顯現於圖像内之玻璃基板之例之說明圖。對 與圖10相同之要素標註與圖10相同之符號，並省略說明。 步驟S21之後，運算裝置4計數兩個重合之像21 、22之夕卜 切矩形23之邊中之與相當於玻璃基板之搬送方向之方向正 交之邊（換&之，圖像中之與搬送方向線正交之邊）之像素 數。即，計數圖1 2中符號D所示之部分之像素數。而且， 運算裝置4使用每個像素之實際空間中之距離乘以該像素 數（步驟S22)。將其結果所獲得之長度記作w。即，w為與 圖12中符號D所示之部分對應之實際空間中之長度。 又，運算裝置4求出外切矩形之邊中之與相當於玻璃基 板之搬送方向之方向平行的邊、及通過兩個像21、22之中 〜口P刀21a、22a之線所成的角0之正切即tan0(步驟S23)。 Θ亦可稱為通過兩個像21 ' 22之中心部分2U、22a之線與 搬送方向線所成之角。因此，亦可為運算裝置4例如預先 決定yc(參照圖17)之值，以已說明之方法計算Xec，進行式 160610.doc -40- 201226888 (3)之計算，藉此計算tane。或者，亦可以其他方法計算 tan9 ° 其次，運算裝置4使用於直至步驟823為止之處理完成算 出之h、r、w、tane而算出yd(步驟S24)。具體而言，只要 運算裝置4藉由進行如下所示之式（8)之計算而計算yd即 〇 yd=(h-r · w)/(l-r · tan0) 式（8) 運算裝置4使用上述yd與預先決定之折射角p來進行式（2) 之计鼻，從而計算氣泡之高度方向位置d(步驟S25)。該計 算與第1玻璃基板檢查方法中之步驟S6相同。 步驟S25之後，運算裝置4根據氣泡之高度方向位置 疋自成為基準之表面52至氣泡為止之距離D(步驟S7”使 成為決定至氣泡為止之距離之基準之表面朝向搬送輥㈠則 而配置玻璃基板之情形時，只要0=(1即可。又，於使成為 決定至氣泡為止之距離之基準之表面朝向搬送輥丨之相反 側而配置玻璃基板之情形時，運算要置4只要為D=T_d而決 定距離D之值即可。該處理與第丨玻璃基板檢查方法（參照 圖7)或第2玻璃檢查方法（參照圖9)中之步驟S7(參照圖9)相 同。 繼而，運算裝置4算出與搬送方向平行之氣泡之直徑之 長度s(步驟S4)。與搬送方向平行之氣泡之直徑之長度3之 算出亦可以與第1玻璃基板檢查方法（參照圖7)中之步驟以 相同之方法進行《或者，亦可藉由進行與第2玻璃檢查方 法（參照圖9)中之步驟S13之處理相同之處理而算出與搬送 160610.doc •41- 201226888 方向平行之氣泡之直徑之長度S。 以後之處理與第丨玻璃基板檢查方法及第2玻璃基板檢查 方法中之步驟S8〜S10相同。 即’運算裝置4根據由步驟S2所拍攝之圖像算出與搬送 方向正交之方向之氣泡之直徑之長度t(步驟S8)。繼而，運 算裝置4藉由計算（sxt2)之立方根而算出氣泡之球換算直徑 e(步驟S9)。 再者’步驟S3〜S9之處理係針對成對之橢圓形之像之組 之每一個分別進行。 進而’運算裝置4檢測自玻璃基板5之表面52起之距離D 為T/2以下之氣泡《而且，運算裝置4依序選擇該氣泡，對 所選擇之氣泡判定所算出之距離D及球換算直徑e之間式 (4)之關係（即0.01)^16+15)是否成立（步驟S10)。 於對本發明之玻璃基板51(參照圖1)進行第3玻璃基板檢 查方法（圖11所示之步驟S1~S10)之情形時，亦分別對於自 表面52起之距離D為T/2以下之氣泡，判定為「e^o.oix D16+15」之關係成立。 圖11所示之第3玻璃基板檢查方法中，即便與第1玻璃基 板檢查方法（參照圖7)或第2玻璃基板檢查方法（參照圖9)同 樣地由同一氣泡所引起之亦可計算像重疊氣泡之高度方向 位置。又’難以受到所搬送之玻璃基板之上下振動之影 響’關於自表面52起之距離D為T / 2以下之氣泡，可精度較 佳地判定其距離D與氣泡之球換算直徑e之間式（4)是否成 立。 160610.doc -42- 201226888 上述各玻璃基板檢查方法中，運算裝置藉由例如追隨程 式而動作之電腦而實現。例如電腦亦可追隨程式作為運算 裝置4而動作。 其次，對本發明之玻璃基板51(參照圖丨）之製造方法進 行說明。圖1所示之本發明之玻璃基板5丨例如只要對以浮 式法製造之玻璃帶應用上述第丨玻璃基板檢查方法（參照圖 7)、第2玻璃基板檢查方法（參照圖9)、及第3玻璃基板檢查 方法（參照圖11)中之任意一個方法，關於自底面起之距離 D為T/2以下之氣泡分別選出該距離D與氣泡之球換算直徑e 之間式（4)成立的蜂璃帶’自該玻璃帶切割出玻璃基板即 可。於對玻璃帶應用第丨玻璃基板檢查方法、第2玻璃基板 k查方法及第3玻璃基板檢查方法之任一方法之情形時， 可使用玻璃帶代替上述玻璃基板。此時，採用上述第 璃基板檢查方法、第2玻璃基板檢查方法及第3玻璃基板檢 查方法之任一者之情形均將底面作為基準，算出自底面至 氣泡為止之距離D，並且算出氣泡之球換算直徑e，關於距 離D為T/2以下之氣泡只要分別選出距離D與氣泡之球換算 直徑e之間式（4)成立之玻璃帶即可。 自如此分別選出之玻璃帶進行板狀裁切而成之玻璃基板 中’可防止相當於底面之面的隆起。 又，於將如此進行板狀裁切玻璃基板用作液晶顯示面板 之透明基板之情形時，只要研磨玻璃帶中之相當於底面之 面’該面朝向液晶側而製造液晶顯示面板即可。其結果為 可製造單元間隙均—之液晶顯示面板。 160610.doc •43- 201226888 又’因可防止氣泡引起之表面之隆起，故而本發明之玻 璃基板可較佳地利用在顯示立體圖像之液晶顯示面板。 因可防止表面之隆起，故而與重疊玻璃基板之情形時， 可防止負重集中於破璃基板之一部分。因此，即便於重疊 玻璃基板之情形時亦可防止玻璃基板之斷裂。 再者，作為玻璃帶之製造方法存在熔融法。關於以熔融 法製造之玻璃帶，將任一表面作為用以決定距離D之基準 面均可。而且，若算出自成為基準面之表面至氣泡為止之 距離D，並且算出氣泡之球換算直徑e，關於距離d為T/2以 下之氣泡分別選出距離D與氣泡之球換算直徑e之間式（4) 成立之玻璃帶，與上述同樣地對玻璃基板進行板狀裁切， 則獲得本發明之玻璃基板51(參照圖丨）。於將自以熔融法製 造之玻璃帶進行板狀裁切而成之玻璃基板用作液晶顯示面 板之透明基板之情形時，亦只要使作為用以決定距離d之 基準面之面朝向液晶側而製造液晶顯示面板即可。於此情 形時，亦可製造單元間隙均一之液晶顯示面板。再者，於 將自以熔融法製造之玻璃帶進行板狀裁切而成之玻璃基板 用作液晶顯示面板之透明基板之情形時，亦可不進行研 磨。 [實施例] 表1係分別表示例1〜例丨〇作為本發明之實施例及例i i〜例 20作為比較例。 160610.doc -44· 201226888201226888 VI. Description of the Invention: TECHNICAL FIELD OF THE INVENTION The present invention relates to a glass substrate. [Prior Art] Various methods for measuring the position in the height direction of defects in the glass substrate have been proposed. As a method of measuring the height direction of a defect in a glass substrate, there is a method of measuring the focus of the camera and detecting the height direction of the defect when shooting a defect. This method is referred to as a third measurement method for convenience. Fig. 13A is an explanatory view schematically showing the second measurement method. In the second measurement method, as shown in Fig. 13A, the glass substrate 82 is conveyed while passing light through the glass substrate 82. Further, the inside of the conveyed glass substrate 82 is taken by the line camera 81. As long as there is a defect inside the glass substrate 82, the defect is captured. Figure 133 is an example of an image showing a defect detected. In Fig. 13A, the defect 83 is schematically represented by a rectangle, and in Fig. nB, the image % of the defect expressed in the image of the glass substrate is also indicated by a rectangle, but the shape of the defect is not limited to the rectangle. However, the following is shown. 14B, FIG. 15B, FIG. 16, and FIG. π also show the defect pattern in a rectangular shape, and the f-head shown in FIG. 13B is the transport direction of the glass substrate 82. When the inside of the glass substrate 82 is taken 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 to measure the absolute distance from the line camera 81 to the defect, and the height direction of the defect is calculated based on the distance. position. As a method of (4) the focus of the camera to make the location of the defect and the focus of the camera, there are methods such as Qing (10) (four) 160610.doc 201226888 from Focus, focus sounding). 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 a device for measuring the position in the height direction of a defect by adjusting the focus of the camera are described, for example, in Patent Documents 1 to 3. Further, as another common method for measuring the position in the height direction of the defect in the glass substrate, there is a method in which the same defect is captured at two positions by the reflected light of the light that is incident on the glass substrate by the human, and the two obtained according to the result are obtained. The height direction position of the defect is measured like the positional relationship. This method is referred to as the second measurement method for convenience. Fig. 14 is a schematic view showing the second measurement method. In the second measurement method, for example, as shown in Fig. 14A, in the glass substrate 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 substrate 82 is conveyed, and the inside of the glass substrate 82 is taken by the line camera 81. The path of the light in the glass substrate is described later with reference to the side view shown in the upper portion of Fig. 16. The defect 83 moves together with the conveyance of the glass substrate 82, and when it overlaps with the path of the light before reflection and overlaps with the path of the reflected light, it is captured as an image in the linear camera 81. As a result, even if there is a defect 83, two images are displayed in the captured image. Fig. 14 is an example of an image taken by the second measurement method. As shown in Fig. 14A, two images 84, 85 appear on the same defect. In the second measurement method, the position in the height direction of the defect 83 is calculated from the positional relationship of the two images in the image illustrated in Fig. 14A. Further, as for the size of the defect, the image taken is subjected to image processing to measure the size of the defect. Further, the arrow shown in Fig. 14A is the conveying direction of the glass substrate 82. A method and apparatus for measuring a position in a height direction of a defect based on a positional relationship between two images by using a reflected light of light incident on a transparent substrate or the like at two positions, for example, are described in Patent Documents 4 to 6, 8 and so on. Further, there is a method of capturing an image on the both surfaces of the glass substrate in the same manner as in the second measurement method, and measuring the height direction of the defect based on the positional relationship of the image in the image captured on each surface of the glass substrate. position. This method is referred to as the third measurement method for convenience. Fig. 5 A shows an explanatory diagram of the third measurement method in a schematic manner. In the third measurement method, for example, as shown in Fig. 15A, light is incident on the same side as the second type camera 8A in the glass substrate 82, and the reflected light reaches the first line type camera 81a. Similarly, the light is incident on the side opposite to the second line type camera 81, and the reflected light reaches the second line type camera 81b. Then, the glass substrate 82 is conveyed, and the inside of the glass substrate 82 is imaged by each of the first line type camera 81a and the second line type camera 81b. In this way, two images are captured in the first linear camera 81a in the same manner as in the second measurement method. Further, two images are captured in the second line camera 81b. Fig. 15B is an example of an image taken by the third measurement method. In the third measurement method, as shown in Fig. 15B, an image taken from the upper side of the glass substrate of one of the line type cameras and an image taken from the lower side of the glass substrate by the other line type camera are obtained. In each image, two images are respectively displayed. In the third measurement method, the position in the height direction of the defect 83 is calculated based on the positional relationship of the image in each of the images taken from the upper side and the lower side of the glass substrate. Further, in Fig. 5B, the case where the images are superimposed from the image taken from the upper side is exemplified. Further, as for the size of the defect, the image taken is subjected to image processing to measure the size of the defect. Further, the arrow shown in Fig. 15B is the conveying direction of the glass substrate 82. An image is taken from both sides of a transparent substrate or the like to obtain a height direction of a defect. I60610.doc 201226888 A method is described in Patent Document 7, for example. In the second measurement method and the third measurement method, the height direction position of the defect is calculated by the fact that the image of the same defect does not overlap in the image. Further, in the third measurement method, when the image is superimposed on one of the images as illustrated in Fig. 15B, the image in the height direction of the defect may be calculated using the other image. The following description shows 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 captured image in the second measurement method. Fig. 16 is an explanatory view showing a position at which a defect in the conveyed glass substrate is taken by a line camera. The figure shown in the upper part of Fig. 16 is a side view of the glass substrate, and the figure shown on the left side of the lower part of Fig. 16 is a plan view corresponding to the side view shown in the upper part of Fig. 6. Further, the graph shown on the right side of the lower portion of Fig. 16 indicates an image obtained when a defect 83 in the conveyed glass substrate 82 is taken. The side view shown in Fig. 16 and the rectangle shown in the top view show the defect 83 in the glass substrate 82. In this example, the defect is that one defect 83 is moved along with the glass substrate 82 being conveyed. The side view and the plan view shown in Fig. 16 respectively show the defect 83 when moving to the position 91 and the defect 83 when moving to the position 92. In the side view and the plan view shown in Fig. 16, the defect itself does not exist in two. As shown in the side view of the upper portion of Fig. 16, the light reaching the linear camera 8 is incident on the transported glass substrate 82 from the surface of the glass substrate 82 on the linear camera side. Further, if the incident light reaches the interface on the opposite side of the incident side of the glass substrate 82, it is reflected at the interface, and the incident angle α of the light reaching the linear camera 81 through the interface of the incident side reaches the linear camera 81 depends on The setting position of the line camera 81. By fixing the position of the line camera 8 I, I60610.doc 201226888 determines that the incident angle α is a fixed value. Further, the light refraction angle β is determined depending on the incident angle α of the light and the refractive index η of the glass substrate 82. Here, the incident angle α and the refractive index η are known, and the refraction angle β is also determined to be a fixed value. The refractive index 11, the incident angle α, and the refraction angle β satisfy the relationship of the formula (1). n=sina/sinp (1) Therefore, as long as the incident angle α and the refractive index n are known, the refraction angle β » is obtained by solving the equation (1) with respect to p, and in the example shown in Fig. 16, The position d in the height direction of the glass substrate μ from the surface on the opposite side to the defect 83 of the line type camera 81 is a measurement target. The line camera 81 continues to capture the inside of the glass substrate 82. The defect 83 and the glass substrate 82 are moved in the transport direction. Further, if the defect moves to the first position 91 which is the path of the light which is incident on the glass substrate 82 and is reflected by the interface and reaches the line camera 81, the line camera 81 captures the first image (hereinafter referred to as the first image). ) as an image of defect 83. Further, when the defect illusion moves to the second person and the light path parent fork position 92, the line type camera 8 丨 captures the second image (hereinafter referred to as the second image) as the image of the defect 83. As a result, as shown on the right side of the lower part of Fig. 16, the "(4) and the second image 99 » in the captured image appear, the light transmitted through the defect 83 when the defect 83 is translucent. The linear camera 81 is reached and captured as an image. When the defect 83 is a defect of light blocking property, the defect 83 appears as an image of black. The defect Μ is not coated as a light-shielding capture. In addition, as shown in FIG. 16, 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 yd. The line connecting the photographing positions in the front direction of the line type camera 81 is referred to as a center line 95», more specifically, a line obtained by orthographically projecting a line connecting the photographing positions in the front direction of the line type camera 81 to the interface of the glass substrate 82 is obtained as a straight line. The center line 95.yd can be imaged based on the first image % and the second image 99 in the captured image (see the right side of the lower portion of FIG. 16) to the line 96 corresponding to the image in the center line 95. The distance between 98 and 99 is determined. According to the image, if yd is determined Then, using the refraction angle β, the equation (2) shown below is calculated. [The height direction position d=yd/(2 · tanP) of the defect 83 can be obtained. Equation (2) Further, the line camera 81 will be oriented. The angle between the line formed by the line of the first image capturing position 91 and the interface of the glass substrate and the center line 95 is set to Θ. At this time, the image taken (see the right side of the lower part of Fig. 6) The angle between the straight line passing through the centers of the first image 98 and the second image 99 and the line 96 is also Θ. In addition, at this time, tane can be calculated as follows. Hereinafter, the figure below is shown in Fig. 16. The yc shown in the top view of the left side will be described, and the calculation of tanQ will be described. In Fig. 16, 'the case where the defect 83 is offset from the front side of the line type camera 81. As shown in Fig. 17, it is assumed that the defect 83 exists in the line type camera. In the case of the front side of 81, the distance between the position at which the position 92 of the second image is projected to the interface of the glass substrate 82 and the position at which the lens portion of the line camera 81 is projected to the interface of the glass substrate 82 is called The shooting distance ye, wherein the shooting distance yc is based on the height direction of the defect 83 When d is the maximum, 160610.doc 201226888 The shooting distance is set to the minimum value yi 'When d is the minimum, the shooting distance ye is set to the maximum value 丫2 (refer to the side view shown in the upper part of Fig. 17) That is, yi $ yc $ y2. Thus, strictly speaking, ye depends on d ' but yc can also be predetermined in the range of $ yc ^ y2. y. Even if it is not accurate, as long as y 1 S yc The value of the range of S y2 can only contain errors that can be ignored in the tane. Also, the offset of the defect 83 from the front direction of the line camera is referred to as Xec (refer to the left side of the lower part of Fig. 16). Xcc can be determined from 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. i6). That is, in the image, the number of pixels corresponding to the distance from the line 96 to the second image 99 is counted. Since the position of the line camera 81 is fixed, the distance in the actual space of each pixel is also determined as a fixed value. The length of Xec can be calculated by multiplying the distance in the actual space of each pixel by the number of pixels corresponding to the distance from the line 96 to the second image 99. Here, "t_ can be used and Xcc is expressed by an approximate expression as shown by the following formula (3), that is, tan0 can be obtained by calculation of the formula (3) using yc and xcc. [Equation 1] tanG: yd+yc yc (3) Further, Patent Document 8 describes a method of detecting a defect by the person emitting light and reflecting light while moving the glass plate while causing the light person to hit the glass plate. , thereby calculating the height direction position of the defect. In the method described in Patent Document 8, when the pattern of the defect is detected, when there is no pattern of almost the size of (4) in the moving direction of the glass sheet, there is a defect near the back side of the glass sheet. In the case of a large defect, the position of the height direction of the defect J60610.doc 201226888 is judged as 〇. Therefore, in the method described in Patent Document 8, in the above case, the position in the height direction of the defect cannot be accurately obtained. [Patent Document 1] [Patent Document 1] Japanese Laid-Open Patent Publication No. 2001-305072 [Patent Document 2] Japanese Patent Laid-Open Publication No. Hei No. Hei. [Patent Document 4] Japanese Patent Publication No. 292-56 [Patent Document 5] Japanese Patent Laid-Open Publication No. Hei 9-61139 (Patent Document 6) 曰本特表 2003_508786号 [Patent Document [Problem to be Solved by the Invention] It is preferable that the surface of the glass substrate is free from defects due to defects. For example, a bubble may be cited as an example of a defect in a glass substrate. If the bubble is located in the vicinity of the surface of the glass substrate, there is a problem that a bulge occurs on the surface of the glass substrate. For example, if a glass substrate having such a bulge on the surface is used as a transparent substrate in a liquid crystal display panel, the cell gap is uneven due to the bulging, particularly in the case of a liquid crystal display panel in which a stereoscopic image (three-dimensional image) is displayed. In the case, since the image for the left eye and the image for the right eye are processed, the processed image information is compared with the liquid crystal display panel displaying the two-dimensional image. The amount of image information is 160610.doc •10· 201226888 is 2 Times. Further, it is necessary to switch the image for the left eye and the image for the right eye at high speed, and it is necessary to set the cell gap to a narrow gap. Therefore, in the case of a liquid crystal display panel displaying a stereoscopic image (three-dimensional image), the uniformity of the cell gap is more strictly pursued, and the glass substrate caused by the bubble which is previously allowed to exist near the surface of the glass substrate is not allowed. A slight bulge on the surface. Further, when the ridge of the surface reaches a certain limit or more, when the glass substrate is stacked, the load is concentrated on the ridge portion and becomes a cause of the fracture. Therefore, it is preferable that the glass substrate used in the liquid crystal display panel has no ridge on the surface (the surface on the liquid crystal side) on at least one side. Accordingly, it is an object of the present invention to provide a glass substrate in which at least one of the surfaces is not raised. [Means for Solving the Problems] The glass substrate of the present invention is characterized in that the thickness of the glass substrate is Τ (μπι) 'the distance from the surface of the glass substrate to the bubbles existing in the glass substrate is set to ϋ(μιη), when the diameter of the ball of the bubble is set to e (gm), the ball-converted diameter e of the bubble existing in the layer of at least τ/2 (μηι) from the surface of one of the bubbles satisfies eg 0.01 χΕ> ι 6+15. The thickness Τ (μιη) of the glass substrate of the present invention is not particularly limited, but the thinner the plate thickness μ (μπι) of the glass substrate is from the surface of the glass substrate to the glass present in the case where bubbles are present in the glass substrate. The smaller the distance Dbm) in the substrate, the higher the possibility that the surface of the glass substrate is raised. Therefore, it is preferably i 〇μηη or more and 700 μm or less, and more preferably 1 〇μηι or more and 4 〇〇μηι or less. 10 μηη or more and 100 μηη or less, and particularly preferably 1 μm μη or more and 50 μηι or less. For example, the glass substrate of the present invention is determined by the following glass substrate inspection method, 1606I0.doc 201226888, to determine that the ball-converted diameter e of the bubble existing in the layer at least 表面/2 (μπι) from the surface satisfies e$〇. 〇lxDi'6+15, the glass substrate inspection method includes: an imaging step of irradiating light from a light source (for example, the light source 2) to a glass substrate conveyed in a stripe direction, and arranging light reflected on the glass substrate Positioning the glass substrate at a position of a photographing mechanism (for example, a line camera 3); and calculating a positional relationship between two coincident elliptical images caused by the same bubble in the glass substrate in the image captured by the photographing mechanism And calculating the height direction position of the bubble in the glass substrate; the ball conversion diameter calculation step 'calculating the bubble conversion diameter e of the bubble; and the determination step, determining the surface of the glass substrate determined by the position of the height direction of the bubble to the bubble Whether the distance D and the bubble ball diameter e are equal to eg 0.01 xD1 6+15. Further, for example, the glass substrate of the present invention is determined by the glass substrate inspection method as follows: the ball-converted diameter 6 of the bubble existing in at least Τ/2 (μπι) from the surface satisfies 6$〇.〇1><〇1. 6+15, the glass substrate inspection method is calculated in the calculation step, and the two overlapping images (for example, 21, 22) caused by the same bubble are calculated in parallel with the direction corresponding to the transport direction of the glass substrate. The length in the actual space corresponding to the number of pixels on the side (for example, h) minus the length (for example, s) of the diameter of the bubble parallel to the transport direction, by calculating the value and the value in the glass substrate The light refraction angle calculates the height direction position of the bubble in the glass substrate, and includes a step of calculating the length of the diameter of the bubble in the direction orthogonal to the transport direction by the image captured by the imaging means, and the diameter in the sphere In the calculation step, the length of the diameter of the bubble parallel to the transport direction is set to 3 (μιη), and the transport direction is 160610. Doc -12- 201226888 When the length of the diameter of the bubble in the direction of intersection is t (pm), calculate (sxt2) l/3, and calculate the ball diameter e of the bubble. In the determination step, determine the height by the bubble. Whether the distance D from the surface of the glass substrate to the bubble and the sphere conversion diameter e of the bubble satisfy the eg 0. 01 xD16+15. Further, for example, the glass substrate of the present invention is determined by the glass substrate inspection method as follows: the gas/package ball diameter e that exists in the layer at least Τ/2 (μπι) from the surface satisfies XD16+i5, In the glass substrate inspection method, the position of the image in the width direction orthogonal to the conveyance direction is used as a variable (for example, the variable u) in accordance with the positional relationship between the two bubbles due to the same bubble. a predetermined calculation formula (for example, equation (6) or equation (7)), calculating a feature quantity of the bubble (for example, 8 or r), and using the feature quantity to calculate the difference from the outer rectangle of the two overlapping images a value corresponding to the length of the diameter of the bubble parallel to the transport direction, which corresponds to the length of the pixel corresponding to the number of pixels in the direction parallel to the direction in which the glass substrate is transported, and the calculated value and the glass substrate Calculating the height direction of the bubble in the glass substrate by the angle of refraction of the light inside, and calculating the length of the bubble diameter in the direction orthogonal to the transport direction by the image captured by the imaging mechanism In the ball conversion diameter calculation step, the length of the direct control of the bubble parallel to the conveyance direction is s (^m), and the length of the diameter of the bubble in the direction orthogonal to the conveyance direction is Κμιη) (st) 'The ball diameter converted by the bubble e is calculated, and in the determination step, the distance D from the surface of the glass substrate to the bubble determined by the position of the height direction of the bubble and the ball-converted diameter e of the bubble are determined. Whether it meets 160610. Doc -13- 201226888 e$ 〇. 〇 1 xd1. 6. Further, for example, the glass substrate of the present invention is determined by the glass substrate inspection method as follows: the gas/package ball diameter 6 which exists in the layer at least within the range of τ/2 (μηι) from the surface satisfies e$〇〇 Ϊ́χ〇1$+ΐ5, the glass substrate inspection method is calculated in the calculation step, and based on the positional relationship of the two coincident images caused by the same bubble, a bubble parallel to the transport direction is calculated using a predetermined calculation formula (for example, equation (6)) The length of the diameter (for example, s) is used as the feature quantity, and the length in the actual space corresponding to the number of pixels in the outer rectangle of the two overlapping images and the side corresponding to the direction of the transport direction is calculated. The value obtained by the length of the diameter is calculated from the calculated angle between the value and the angle of refraction of the light in the glass substrate, and the position of the bubble in the glass substrate is calculated in the height direction, and is calculated by the image captured by the imaging mechanism. In the ball conversion diameter calculation step, the length of the diameter of the bubble parallel to the conveyance direction is s (pm), and the conveyance direction is the step of calculating the length of the diameter of the bubble in the direction in which the conveyance direction is orthogonal. When the length of the diameter of the bubble in the direction of intersection is ί (μιη), the sphere conversion diameter ^ of the bubble is calculated, and in the determination step, the position determined by the height direction of the bubble is determined. Whether the distance D between the surface of the glass substrate and the bubble and the diameter of the sphere of the bubble are equal to eg 0. 01xD16+15. Further, for example, the glass substrate of the present invention is determined by the following glass substrate inspection method to determine that the ball-converted diameter e of the bubble existing in at least Τ/2 (μϊη) from the surface satisfies e$〇. 〇lxD16+15, the glass substrate inspection method is used in the calculation step, using a predetermined calculation formula (for example, the formula (7)), based on the positional relationship of the two coincident images caused by the same bubble, the bubble 160610 is calculated. Doc •14- 201226888 The ratio of the two diameters (for example, !) is the feature quantity, which is formed by the line in the image corresponding to the shooting position of the front direction of the shooting mechanism and the line passing through the centers of the two images. And the ratio of the above-mentioned ratio, the length of the actual space corresponding to the number of pixels in the side of the tangent rectangle which is parallel to the direction of the transport direction of the glass substrate, and the bubble parallel to the transport direction The value obtained by the length of the diameter 'calculates the height direction position of the bubble in the glass substrate by the calculated value and the angle of refraction of the light in the glass substrate, and includes the image captured by the imaging mechanism. In the ball-converted diameter calculation step, the length of the diameter of the bubble parallel to the conveyance direction is s (pm) 'the bubble in the direction orthogonal to the conveyance direction, the step of calculating the length of the diameter of the bubble in the direction in which the conveyance direction is orthogonal When the length of the diameter is Κμηι), (Sxt2) 1/3 is calculated, and the ball-converted diameter e of the bubble is calculated, and in the determination step, the position determined by the height direction of the bubble is determined. Glass substrate surface until the bubbles and the bubble distance D between the diameter of the ball 6 in terms of whether or not satisfy eg 0. 01 xD16+15. Further, for example, the glass substrate of the present invention may be a glass substrate which is cut into a sheet shape from a glass ribbon produced by a float method, and is present within Τ/2 (μηι) from the surface corresponding to the bottom surface of the glass ribbon. The ball of the bubble in the layer is converted to a diameter e that satisfies 〇. 〇ixd16+15. Further, for example, the glass substrate of the present invention may be a glass substrate of a liquid crystal display panel, and the ball-converted diameter e of the bubbles existing in the layer of Τ/2 (μηι) from the surface facing the liquid crystal side satisfies 〇. 〇1xD16+15. [Effects of the Invention] According to the glass substrate of the present invention, it is possible to prevent the surface of at least one side from being 160610. Doc 15 The uplift of 201226888. [Embodiment] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Fig. 1 is an explanatory view showing an example of a side view of a glass substrate of the present invention. The glass substrate 51 of the present invention is a glass substrate that satisfies the following conditions. In other words, in the glass substrate 51 of the present invention, the thickness of the glass substrate is Τ (μπι), and the distance from the surface of the glass substrate to the bubbles existing in the glass substrate is D (pm) ' When the diameter of the bubble ball is ε (μπι), the ball-converted diameter 6 of the bubble existing in the layer of Τ/2 (μπι) from the surface of at least one of the two surfaces of the glass substrate satisfies the following The condition satisfies the formula (4) shown below. 0. 01 xD1, 6+15 Equation (4) Here, DST/2. Specifically, the distance from the surface of the glass substrate to the bubble is determined to be close to the surface of the bubble. In the example shown in Fig. 1, the distance D from the surface of the glass to the bubble 57 is the case where the surface 52 close to the bubble 57 among the two surfaces 52, 53 of the main surface of the glass substrate is used as a reference. The distance from the surface 52 to the bubble 57. Further, the oxygen bubble is one of defects in the glass substrate or the glass ribbon. In Fig. i, the bubble 57 is shown as a sphere for the sake of convenience in indicating the diameter of the sphere e, but the continuous bubble is a shape close to the ellipsoid which rotates the ellipse around the major axis of the ellipse. The bubble can be regarded as a spheroid that rotates the ellipse centering on the long axis of the ellipse. Further, the length of the minor axis of the ellipse is t (pm), and the length of the major axis is "#." Fig. 2 is an explanatory view showing the shape of such a bubble. Fig. 3 is a view showing this from above. 160610. Doc •16· 201226888 An illustration of the state of the bubble. As shown in Fig. 2, the height of the bubble and the width of the bubble can be regarded as a common value and both are the same. The length of the bubble is equal to the long axis of the ellipse and is S. When the ball diameter of the bubble is set to 6 (μιη), the ball-converted diameter e is obtained by calculation of the formula (5) shown below. e=(Sxt2), /3 Equation (5) That is, the sphere conversion diameter e is the cube root of (sxt2). For example, the glass substrate 51 is a glass substrate used as a transparent substrate of a liquid crystal display panel. In this case, at least the surface on the liquid crystal side of the two surfaces of the main surface of the glass substrate 51 is used as a reference, and the distance from the surface to the bubble is set to 〇(μιη)Β, and the bubble is The ball is converted between the diameter e and the distance D as long as the formula (4) is established. Here, the bubble is present in the layer of the layer Τ/2 (μιη) from the surface of the liquid crystal side, τ/2. Further, even when the other surface of the glass substrate is used as a reference, the relationship between the sphere-converted diameter e and the distance D can be established. Further, the surface of the two surfaces which are the main surfaces of the glass substrate facing the liquid crystal side may be, for example, a surface on which a transparent electrode is disposed. Therefore, the glass substrate 51 shown in Fig. 1 is used for the glass substrate of the liquid crystal display panel. If the surface 52 faces the liquid crystal side, the distance D from the surface to the bubble can be measured using the surface 52 as a reference. In the case where the glass substrate manufactured by the floating method is plate-cut to produce a glass substrate for a liquid crystal display panel, the surface corresponding to the bottom surface of the glass ribbon is polished to face the liquid crystal side. The liquid crystal display panel is manufactured. Therefore, when the glass ribbon manufactured by the floating method is plate-cut to the glass substrate 160610, doc • 17-201226888, and the glass substrate for the liquid crystal display panel is manufactured, it will become the main surface of the glass substrate 51. When at least the surface of the bottom surface of the glass ribbon is used as a reference, and the distance from the bottom surface to the bubble is D (pm), the sphere is converted between the diameter e and the distance d, as long as 4) It can be established. Here, the bubble is a bubble present in a layer corresponding to a surface Τ/2 (μιη) corresponding to the side from the bottom surface, Dg T/2. Further, even when the surface corresponding to the upper surface of the glass ribbon is used as a reference, the relationship between the sphere-converted diameter e and the distance D can be established. Further, the surface on the lower side of the glass ribbon manufactured by the floating method is referred to as a bottom surface, and the upper surface is referred to as an upper surface. Therefore, the glass substrate 51 shown in FIG. 1 is a glass substrate which is cut into a plate shape from a glass ribbon manufactured by a floating method, and if the surface 52 is a surface corresponding to the bottom surface, the surface 52 is used as a reference. The distance D from the surface to the bubble is sufficient. Regarding the surface of the glass substrate (herein, the surface 52 shown in FIG. 1), the bubble existing in the layer within Τ/2 (μιη), the distance D from the surface to the bubble, and the diameter of the sphere of the bubble are e The equation (4) is established to mean that the smaller the diameter of the bubble sphere is, the closer it is to the surface 52. In other words, in the vicinity of the surface, there is no bubble having a larger diameter of the sphere. Therefore, the bulging of the surface 52 caused by the influence of the air bubbles can be prevented, so that the quality of the glass substrate can be improved. Further, in the glass substrate 51 of the present invention, since the bulging of the surface 52 can be prevented as such, the cell gap can be made uniform when used as a transparent substrate in a liquid crystal display panel. In addition, the glass ribbon is cut from the glass plate manufactured by the floating method. Doc 201226888 When the glass substrate is used as a liquid crystal display panel, the surface corresponding to the bottom surface of the glass ribbon is polished. However, the glass substrate 51 of the present invention may be the surface 52 before polishing as a reference, from the surface 52 to the bubble. The glass substrate in which the formula (4) is established between the distance D (where D $ T/2) and the sphere of the bubble are converted to e. Further, a glass ribbon manufactured by a float method or the like, or a glass substrate cut from the glass ribbon shape, produces a streak along the main direction of the glass ribbon. The main direction of extension of the glass ribbon 'is not the direction of the width of the glass ribbon achieved by the guiding member, but the direction along which the glass ribbon extends in the forward direction. Hereinafter, the main direction of extension of the glass ribbon is simply referred to as the direction in which the glass ribbon extends. The term "stripe" refers to a stripe which is generated in the direction in which the glass ribbon is stretched due to fluctuations and undulations in the direction perpendicular to the direction in which the glass ribbon extends. The glass substrate cut from the glass ribbon is also striped. Further, since the extending direction of the glass ribbon is the same as the direction in which the glass ribbon is fed from the glass f manufacturing apparatus (not shown), the direction of the stripe in the forward direction of the glass ribbon fed during manufacture is the same as the direction in which the glass ribbon extends. direction. In the following, an example of an inspection system or a glass substrate inspection used for measuring the distance D from the surface to the bubble in the glass substrate, or calculating the sphere-converted diameter e of the bubble, and whether or not the inspection method (4) is established The method is explained. In the case of the glass substrate in which the formula (4) is established by the glass substrate inspection method, the glass substrate according to the present invention is obtained. In the glass substrate inspection method, the distance D from the surface of the glass substrate to the bubble was calculated for the bubble among the various defects of the glass substrate. The bubbles in the glass ribbon or the glass substrate are ellipsoidal. Therefore, shooting glass 160610. Doc •19· 201226888 In the image obtained by the bubble in the glass substrate, the image of the bubble becomes an ellipse. Further, in the image of the bubble (the image of the ellipse) taken as an image, the center portion is divided into white. Therefore, the central portion of the image of the bubble appearing in the image can be utilized as a characteristic point (hereinafter referred to as a feature point). First, regarding the glass substrate, the first glass substrate inspection method in the method of inspecting the glass substrate in which the inspection formula (4) is established will be described. Fig. 4 is a schematic view showing a configuration example of an inspection system in which the equation (4) is established between the distance D from the surface to the bubble in the glass substrate and the ball-equivalent diameter e of the bubble. This inspection system includes a conveying roller 1, a light source 2, a line type camera 3, and an arithmetic unit 4. The conveyance roller 1 supports the glass substrate 5 to be inspected, and conveys the glass substrate 5 at a fixed speed in the fixed direction. The glass substrate 5 is conveyed in the direction along the stripe direction of the glass substrate 5 itself. Therefore, the conveying direction of the glass substrate 5 of the conveying roller 1 is the same direction as the stripe direction of the glass substrate 5. Further, in the present example, the surface of the two surfaces of the glass substrate which is the reference for determining the distance to the bubble (the surface 52 in FIG. 1) is opposite to the light source 2 and the line type camera 3, as an example. On the side, the glass substrate 5 is supported on the conveying roller 1. For example, when the glass substrate 5 is a glass substrate cut from a glass ribbon plate produced by a floating method, the surface corresponding to the bottom surface of the glass ribbon is directed to the opposite side of the light source 2 and the linear camera 3, The glass substrate 5 may be attached to the conveying roller 1. Moreover, when the glass substrate $ is used as a transparent substrate in a liquid crystal display panel, the glass substrate 5 can be supported by the conveyance roller 1 as long as the surface facing the liquid crystal side faces the opposite side of the light source 2 and the linear camera 3, and In the glass substrate inspection method, it is measured from the glass substrate 5 160610. Doc . 20· 201226888 The position (distance) in the height direction from the surface 52 on the side of the conveying roller 1 to the bubble. Here, the position in the height direction is the distance from the surface on the side of the conveying roller to the bubble. Therefore, when the surface to be referenced faces the conveyance roller 1 side, the measured value in the height direction position means the distance D from the surface to be the reference to the bubble. In addition, the surface 52 which is the reference for determining the distance to the bubble may be directed to the conveyance roller 1 so as to face the side opposite to the conveyance report 1. In this case, the distance D from the surface to be the bubble to the bubble is a value obtained by subtracting the measured value of the position in the height direction from the thickness T of the glass substrate 5. The plate thickness T is known, and the distance D from the reference surface to the bubble is based on the measured value of the height direction of the bubble, regardless of which side of the surface 52 which is the reference for determining the distance to the bubble. Decide. As described above, the surface of the surface 52 which is the reference for determining the distance to the bubble is directed to the opposite side of the light source 2 and the linear camera 3 (that is, the side of the transport roller 1), and the glass substrate 5 is supported and transported. Roller 1. The light source 2 is disposed on one side of one of the two faces of the glass substrate 5, and emits light toward the glass substrate 5. This light enters the glass substrate 5 from the interface 8, passes through the inside of the glass substrate, and is reflected by the surface 52 on the opposite side of the incident side. The reflected light reaches the line camera 3 through the interface 8 on the incident side. Further, in Fig. 4, the path of the light is simplified, and as shown in the side view of the upper portion of Fig. 16, the path of the light is refracted when the light enters the interface 8 and the reflection in the interface 52 passes through the interface 8. The line type camera 3 is disposed on the light that is irradiated from the light source 2 and reflected by the glass substrate 5 I606I0. Doc •21 - 201226888 The location reached. Specifically, the glass substrate 5 is used as a reference. It 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 substrate 5 with the light source 2 as a reference. Further, the line type camera 3 captures the inside of the glass substrate 5 to generate an image as a result of the shooting. By determining the arrangement positions of the light source 2 and the line type camera 3, the incident angle α (refer to the upper portion of Fig. 16) in the optical path 亦6 is also determined to be a fixed value. Further, the refractive index η of the glass substrate 5 is also known, and the value of the refraction angle ρ in the path of the light from the light source 2 to the line 3 to the camera 3 is also determined to be a fixed value. The glass substrate 5 is transported, and the linear camera 3 continues the photographing of the glass substrate 5 at a fixed position. Therefore, the portion photographed in the glass substrate 5 changes as time passes. Therefore, if the line connecting the photographing positions in the front direction of the line camera 3 is projected to the interface 8 of the glass substrate 5, it is indicated as a straight line. This line is called the towel line. Fig. 5 is an explanatory view showing a center line, and Fig. 5 is an explanatory view showing a line corresponding to a central line in the image. Fig. 5 is a plan view of the glass substrate 5. With the conveyance of the glass substrate 5, the photographing position of the front side of the line type camera 3 changes, and the front projection of the interface to the line is shown as the center line 95. Further, Fig. 5B shows an image taken by the line camera 3. In the image, a line 96 corresponding to the center line 95 is indicated by a 'dot key line. The line 96 may be a line connecting pixels corresponding to the shooting position of the line camera 3 in the front direction. Further, the center line 95 is parallel to the transport direction of the glass substrate 5, and the line 96 corresponding to the image of the center line 95 indicates the direction in the image corresponding to the transport direction of the glass substrate 5. Further, since the glass base is conveyed in the stripe direction, the line 96 in the image can also indicate the direction corresponding to the stripe direction. The line % corresponding to the image in the center line 95 is recorded as the transporter 160610. Doc -22· 201226888 Further to the line, in the figure, the direction line 96 is shown for illustration, but the conveyance direction line 96 does not appear in the image in the captured image. When there is a bubble in the glass substrate 5, the image of the bubble appears in the image taken by the line camera 3 due to the single bubble. Further, the image of the bubble appearing in the image in which the bubble is captured is an elliptical shape, and the center portion thereof is white. The arithmetic unit 4 refers to the image captured by the line camera 3 to measure the position in the height direction of the bubble. The position in the height direction of the bubble is the length indicated as "d" in the side view of the upper portion of Fig. 16. That is, in the glass substrate 5, the distance from the surface 52 on the opposite side of the light source 2 to the bubble is obtained. When the arithmetic unit 4 overlaps the pair of elliptical images obtained by capturing the common bubbles, the position of the bubbles in the glass substrate 5 in the height direction is calculated based on the positional relationship of the pair of elliptical shapes. Specifically, the arithmetic unit 4 calculates the actual space corresponding to the number of pixels in the outer rectangle of the image in which the image is overlapped with the direction parallel to the direction in which the glass substrate is transported. The distance is 'the value of the length of the diameter parallel to the transport direction among the diameters of the bubbles. Further, in the force image, the direction parallel to the direction in which the glass substrate is conveyed is the value of the light source transfer device 4 (see FIG. 5b) by the subtraction calculation and the glass. The position of the height direction of the bubble is calculated by the refraction angle β in the substrate $. This calculation will be described later with reference to FIG. 8. Further, when the glass substrate is conveyed in the stripe direction of the glass substrate, the long axis of the bubble in the glass substrate is parallel to the transport direction of the transport roller (in other words, the direction of the glass substrate 5). For example, the long axis of the bubble 1606I0. The direction of the doc -23·201226888 72 is offset from the transport direction 71 of the glass substrate 5 of the transport roller 1 by a maximum of 10°. Thus, the long axis 72 of the bubble is substantially parallel to the transport direction 71 of the transport roller 1. Therefore, the linear camera 3 is used. In the captured image, the long axis of the image of the bubble shown as an ellipse is substantially parallel to the transport direction line 96 (see FIG. 5B), and the long axis and the transport direction line of the image of the bubble in the captured image. The case of 96 parallel is taken as an example for explanation. Further, when the paired images are not overlapped, the arithmetic unit 4 may calculate the height direction of the bubble by a known method. The arrangement position of the line type camera 3 is fixed. Therefore, the distance in the actual space corresponding to the pixels in the image captured by the line camera 3 is also determined to be a fixed value. The distance in the actual space corresponding to the pixels in the image is known. Fig. 7 is a flow chart showing an example of a second glass substrate inspection method in the glass substrate inspection method in which the bubble inspection of the glass substrate satisfies the condition of the formula (4). First, the light source 2 starts to irradiate light between the glass substrates 5 to be inspected (step S1). Further, the transport roller 1 transports the glass substrate 5 disposed on the transport roller 向 in the fixed direction, and the linear camera 3 continues the photographing of the inside of the transported glass substrate 5. At this time, the glass substrate 5 is placed on the conveying roller so that the stripe direction of the glass substrate 5 itself is the same as the conveying direction, and is conveyed in the stripe direction. Moreover, the line type camera 3 generates an image as a shooting result (step S2). The line type camera 3 transmits an image obtained by shooting to the arithmetic unit 4. When there is a bubble inside the glass substrate 5, in step s2 J60610. Doc -24- 201226888 The image obtained contains images of bubbles. Specifically, an image of an ellipse appears as an image of a bubble in the image. Further, as illustrated in Fig. 16, when the bubble 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. 16), and the bubble moves to the light after the reflection When the position where the paths overlap (the position 92 shown in the side view of the upper part of Fig. 16), the image appears as an image. Therefore, in the case where there is one bubble, two images appear in the image 2. Further, when the bubble is large, or when the bubble exists on the surface 52 of the glass substrate 5 (refer to the vicinity of the picture, the two images overlap. The arithmetic device 4 receives the image generated by the step S2, Then, the area of the rectangular rectangle is detected from the two overlapping images in the image, and the side of the circumscribed rectangle is counted so as to be parallel to the direction corresponding to the direction of the transport of the glass substrate 5 (ie, The number of pixels in the parallel direction of the transport direction line 96 in the image. Further, the arithmetic unit 4 multiplies the distance in the real space of each pixel by the number of pixels on the X side, thereby calculating the number of pixels corresponding to the side. The length in the actual space (step S3) Fig. 8 is an explanatory view showing the area of the two overlapping images outside the rectangular rectangle. As shown in Fig. 8, the outer rectangle shown in Fig. 8 is determined as the coincident two. The rectangles are cut out like 21 and 22. The angles 21 and 22 are elliptical and can be regarded as congruent. In the example shown in Fig. 8, the long side of the circumscribed rectangle 23 is parallel to the transport direction line 96 (see Fig. 5B). In this case, the arithmetic device 4 counts the image of the long side 24 of the rectangle 23 outside the images 21, 22. And multiplying the actual number of pixels in the spatial distance of each pixel. The actual length of the long side of the space 24 corresponding to the "hJ FIG. Here, h is the unit μηι. 160610. Doc •25· 201226888 Further, when the defect is a bubble, the central portion 21a 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 of one to the short side of the short side of the circumscribed rectangle 23. That is, the number of pixels of the portion indicated by the symbol a in Fig. 8 is counted. The arithmetic unit 4 multiplies the number of pixels by the distance in the real space of each pixel. The result of the multiplication is a length corresponding to the actual space corresponding to the portion of the eight shown in Fig. 8. Specifically, the diameter of the bubble parallel to the transport direction (the diameter of the bubble is parallel to the transport direction) The length of 1/2. In the example shown in Fig. 8, the diameter is the long axis of the bubble. The arithmetic unit 4 calculates the length of the diameter of the bubble parallel to the transport direction by multiplying the result of the multiplication by 2 (step S4). The length of the diameter of the bubble is equivalent to that of Fig. 3. Here, the unit of s is μπι. The portion in the image corresponding to the length of sq in the real space is the portion indicated by the symbol Α in Fig. 8. Further, the two images 21, 22 can be regarded as congruent, and therefore, in Fig. 8, it can be regarded as A = A'. Here, the case where s is calculated using the center portion 21a of the image 21 will be described as an example, but s may be calculated using the center portion of the image 22. Further, in Fig. 8, the case where the long axis of the image of the bubble is parallel to the direction of the transport direction is taken as an example (4), but the long axis of the image of the bubble is not completely parallel with the direction of the transport direction. The long axis of the bubble is offset by a maximum of 1 搬 in the direction of transport of the glass substrate. (Refer to Figure 6). Therefore, even if the length of the image of the bubble (4) is not completely parallel to the conveyance direction line, it can be regarded that the two are parallel, and h and s are calculated in the same manner as the above-described steps and s4. That is, when h is found, as long as the two images of the coincidence are counted, the long side of the rectangle is 160610. Doc -26- 201226888 The number of pixels 'multiply the distance in the actual space of each pixel by the pixel". When s is obtained, it is only counted from the center of one image to the short side of the circumscribed rectangle. The number of pixels near the short side is multiplied by the number of pixels in the real space of each image, and the multiplication result is multiplied by 2. Even if the long axis of the image of the bubble is incomplete with the direction of the transport direction Parallel, calculate h and s as described above, and calculate the height direction of the bubble using the h and s. It only contains the error that can be ignored. X, even in this case, S can be regarded as the long axis of the bubble. The human-to-person computing device 4 subtracts the s calculated in step S4 from the hash of # in step 83 (step S5). The result of the lecture is set to ^ from the position where the first image is captured. The moving distance of the bubble until the position where the second image is captured is the distance between the two points in which the image of the bubble is captured. The older, the sounder & The portion in the image corresponding to the length is the portion indicated by the symbol B in Fig. 8. The arithmetic device 4 makes Step S5 is calculated by the force of the predetermined refraction angle β to the formula (2) of the calculation, thereby calculating the position of the height direction of the bubble ^ That is, calculate yd / (2. Tanp), the calculation result is d (step S6, the height direction position d of the bubble is the surface 52 from the glass substrate 5_ on the side of the conveyance roller 1 (refer to the distance from the picture to the bubble. /, - person, operation The device 4 determines the distance D from the reference surface 52 to the bubble based on the position d of the bubble in the height direction (step S7). As in this example, the surface of the reference to the distance determined by the bubble is directed toward the transfer roller. When the glass substrate is placed on one side, the distance D from the surface to the bubble is equal to the height direction position 气泡 of the bubble calculated in step 36. Since I60610. Doc • 27- 201226888 Therefore, the value of the height direction position d of the bubble can be set to the distance D from the reference surface 52 to the bubble. In other words, in the arithmetic unit 4, the value of the distance d may be determined by setting D = d. In the case where the glass substrate is placed on the side opposite to the transport roller 1 as the reference for determining the distance to the bubble, the distance D from the surface to the bubble can be made by the thickness τ of the glass substrate. (μιη) obtained by subtracting the value of the position d of the height direction of the bubble. In other words, in the case of the operation device 4, the value of the distance d may be determined by setting D = T_d. Among them, the thickness T of the glass substrate is known. After step S7, the arithmetic unit 4 calculates the length of the diameter of the bubble in the direction orthogonal to the transport direction based on the image captured in step S2 (step S8). In step S8, the area of the rectangle is circumscribed by the two coincident images detected in step S3 (see Fig. 8). Specifically, the arithmetic unit only needs to count its own weight. The number of pixels from the center of the image of the t彳 to the longest side of the long side of the circumscribed rectangle 'multiplied by the distance in the real space of each pixel, and multiplied by the number of pixels Multiply the result of the operation by 2. The value is the width of the bubble' equivalent to the cabinet 1 ^ ^ one of the fields in Figure 2 is not t. Here, the unit of t is μπ- and since the bubble is a spheroid, the height of the bubble is the same as the width of the bubble ί(μηι). Next, the 'computing device 4' uses the length s of the diameter of the bubble parallel to the transport direction calculated by the step S4 and the length t of the straight hole of the bubble orthogonal to the direction of the transport calculated by the step. The ball conversion diameter e of the bubble is calculated (step S9). The arithmetic unit 4 0 has to calculate the ball diameter e by the calculation of the heading type (5). That is, I have tried to calculate the (sxt2) cube root 160610. Doc •28· 201226888 Calculate the ball conversion diameter e. Here, the unit of e is μιη. Further, steps S3 to S9 are performed for each of a group of pairs of elliptical images. Next, the arithmetic unit 4 detects air bubbles having a distance 1) of 172 or less from the surface 52 of the glass substrate 5. T is the thickness of the glass substrate 5, and the arithmetic unit 4 selects the bubbles in order, and determines whether or not the formula (4) between the distance D calculated by the selected bubble and the ball-differential diameter e is satisfied (step si〇). . The arithmetic unit calculates the distance D from the surface 52 and the sphere-converted diameter e for each of the groups of the pair of elliptical images. In the step S1, the arithmetic unit 4 determines that there is one bubble for each of the elliptical groups having the distance d of T/2 or less, and detects the bubble having a distance D from the surface 52 of T/2 or less. . Further, each of the detected bubbles is sequentially selected, and between the distance D calculated by the selected bubble and the ball-converted diameter e, "eg〇〇lxDi. Whether the relationship of 6+15" is established. For each of the bubbles 172 or less from the surface 52, "eg〇. 〇ixDi. The glass substrate in which the relationship of 6+15" is established conforms to the glass substrate of the present invention. On the other hand, the distance from the surface 52 is 气泡 in the bubble below 1/2, in the presence of "eS〇. 01xDi. When the relationship of 6+15" is not satisfied, the glass substrate does not conform to the glass substrate of the present invention. Therefore, when the glass substrate 51 (see FIG. 7) of the present invention is subjected to the above-described method of inspecting the glass substrate (steps S1 to S10 shown in FIG. 7), the distance d from the surface 52 is equal to or less than T/2. Each is judged as "e$0. The relationship of 01xD16+15" was established. In the image obtained by the bubble existing in the vicinity of the surface 52 on the side of the conveyance roller 1 (see Fig. 4), the two images caused by the bubble appear to coincide. Again, that is 160610. Doc -29- 201226888 When the bubble is large, the two images appear in the image taken by the bubble. In the second measurement method described with reference to Fig. 4a, when the two images caused by the same defect overlap, the position in the height direction of the image cannot be measured. Further, in the third measurement method described with reference to FIG. 15A, since the image from the upper side and the image from the lower side are imaged as shown in FIG. 15B, the image does not overlap in any one of the images. The position of the height direction of the defect can be determined. However, in the case where the defect is large, the images captured by the two line cameras sometimes overlap, and in this case, the position in the height direction of the defect cannot be measured. In the glass substrate inspection method (steps S1 to S10 shown in Fig. 7), the position of the bubble in the height direction can be calculated even if the images caused by the same bubble overlap. Therefore, the distance D from the surface 52 can be determined, and the distance 1) from the surface 52 is 172 or less, and it can be determined whether or not the distance between the distance D and the ball-converted diameter e of the bubble is established. In the first measurement method described with reference to FIG. 14A, the measurement result of the position in the height direction of the defect is easily affected by the vibration of the glass substrate being conveyed, but in the glass substrate inspection method shown in the above steps S1 to S10 It is difficult to be affected by this effect', so that the height direction position of the bubble can be calculated with high precision. As a result, it is possible to accurately determine whether or not the equation (4) between the distance D and the ball-converted diameter e of the bubble is satisfied with respect to the bubble having the distance D from the surface 52 of 172 or less. The glass substrate inspection method in which the distance D from the reference surface 52 in the glass substrate is τ/2 or less is inspected between the distance D and the ball-converted diameter e of the bubble. Limited to the square 160610 shown in Figure 7. Doc •30· 201226888 Law (steps ~ S1〇). Hereinafter, the second glass substrate inspection method and the third glass substrate inspection method for performing the same inspection will be described. In either case, for example, the inspection system illustrated in Fig. 4 can be used for inspection. The positional relationship between the light source 2 and the linear camera 3 with respect to the glass substrate 5 to be inspected is the same as that of the first glass substrate, and the description thereof will be omitted. However, the measurement method using the height direction position d of the bubble of the arithmetic unit 4 is different from the glass-correcting substrate inspection method. In the second glass substrate inspection method and the third glass substrate inspection method, the glass substrate is placed on the conveyance roller 1 and conveyed so as to be conveyed in the direction along the stripe direction of the glass substrate itself. In the second glass substrate inspection method, the arithmetic unit 4 calculates the feature amount of the air bubbles in the glass substrate 5 to be inspected. Further, the arithmetic unit 4 calculates the length in the real space corresponding to the number of pixels of the side of the two rectangles which are parallel to the direction parallel to the direction in which the glass substrate is conveyed, by using the feature amount. The length of the diameter of the bubble (the diameter of the bubble which is parallel to the transport direction) parallel to the transport direction of the glass substrate is subtracted. Further, when calculating the feature amount, the arithmetic unit 4 calculates the feature amount based on the positional relationship between the two images that overlap each other using a predetermined calculation formula. Further, in the second glass substrate inspection method towel, the length of the diameter of the oxygen bubble parallel to the direction in which the glass substrate is conveyed is calculated as a feature amount. The equation for calculating the above-described feature amount is determined in advance as a coordinate of a position corresponding to a feature point of the image on which the end portion of the glass substrate is used as a reference. The area of h and the image of the two combined images described in the glass substrate inspection method are functions of variables. Used to determine the amount of the feature (the diameter of the bubble is made to transport 160610. The calculation formula of doc - 31 - 201226888 diameter parallel) can be expressed, for example, by the following formula (6). S=aiu2+a2h2+a3p2+a4uh+a5hp+a6up+a7u+a8h+a9p+a10 (6) In the formula (6), 'u' corresponds to a feature point of the image on which the end of the glass substrate is used as a reference. The coordinates of the position are specifically the distance from the side of the glass substrate parallel to the transport direction to the center of the bubble. Here, the unit of u is mm»"h" is a value obtained by the same calculation as that of step S3 in the second glass substrate inspection method based on the image of the bubble. Here, the unit of h is μπι. P is the area of the area occupied by the two images in the image taken by the bubble (the collection of the two image areas), specifically, the number of pixels in the image. In the formula (6), a broad ai is a coefficient. Further, s in the formula (6) is a diameter of a bubble parallel to the conveying direction of the glass substrate. The diameter s which becomes the characteristic amount is easily affected by the position of the bubble in the width direction in the glass ribbon. Further, the position at which the self-contained glass ribbon is plate-cut to the glass substrate is generally fixed in the width direction of the glass ribbon. For example, generally, when the distance from the side of the glass ribbon to the plate-shaped cutting position of the glass substrate is X, X is fixed, and the glass substrate group is sequentially cut into a plate shape. Therefore, it can be said that the diameter s of the feature amount is also easily affected by the position of the bubble in the direction perpendicular to the transport direction in the glass substrate (in other words, the position of the bubble in the direction perpendicular to the stripe direction in the glass substrate). . Therefore, a calculation formula containing the above-described variable u (for example, the above formula (6)) is used for the calculation of s. Further, in the captured image, when the long axis of the image of the bubble is in the It-shape parallel to the direction of the transport direction, the above s corresponds to the long axis of the bubble. However, even in the image, when the long axis of the image of the bubble and the direction of the transport direction are not complete, 160610. Doc -32· 201226888 Since the two are roughly parallel, the above characteristic amount S can be regarded as the long axis of the bubble. Even if you look at the long-term fault of the long-term π π long axis, you can only include the error that can be ignored, and the calculation of the height direction of the bubble is not affected. ▲ / () coefficient ...,. It is determined in advance by the least squares method. Specifically, LV is used as a bubble n w σ p A L u of the sample. Further, the glass substrate containing the bubbles as the sample was subjected to the same treatment as the step (4) described in the first glass substrate inspection method to obtain h. Further, the number of pixels 6 of the image 6 obtained from the step S2 at this time becomes the pixel number P of the region in which the two images are combined. A plurality of bubbles which become samples are prepared, and S, u, h, and P are obtained for each of the bubbles. If a group of complex arrays s, u, h, and p is obtained, the coefficient in the equation (6) is obtained by the least squares method from the group of s, ^", ie, *5J" 〇S and u, h, p The coefficients in the equation (6) can be obtained by the least square method. The transport device 4 finds an image obtained by photographing the glass substrate to be measured in the height direction of the bubble. And h, p, and calculate s by substituting into equation (6). Further, the arithmetic unit 4 calculates h_s (= yd), and calculates the height direction position of the bubble using the calculation result and the refraction angle β. Fig. 9 shows the second position. The flow chart of the glass substrate inspection method is the same as that of the first glass substrate inspection method, and the same reference numerals are given to those in Fig. 7. The operation until the calculation of h in step S3 is the same as the first glass substrate inspection method. Fig. 10 is an explanatory view showing an example of a glass substrate which is displayed in an image. Doc -33- 201226888 When there is a bubble, the image of the bubble 21 and 22 are also displayed in the image. In the example of the image, as the feature point of the image, the central portion 2 of each image 21, 22 a, 22a also appear as white areas in the image. Further, although the rectangle 23 is cut out like the 21 and 22, the circumscribed rectangle 23 does not necessarily appear in the image. After the step S3, the arithmetic unit 4 counts the number of pixels from the end portion 31 of the glass substrate to the feature point of the image in the image. That is, the number of pixels of the portion indicated by the symbol 匸 in the figure 计数 is counted. Moreover, the arithmetic unit 4 multiplies the distance in the real space of each pixel by the number of pixels (step s丨丨). This multiplication result corresponds to the distance u » from the end (side) of the glass substrate in the actual space to the bubble. In step S11, u is calculated. In the above description of the step S11, the case where the end portion 31 of the glass substrate is projected in the image will be described as an example for simplification of description. In the case where the end portion 31 of the glass substrate is not reflected in the image, the distance u may be calculated as described below. Since the installation position of the line type camera 3 is fixed, the distance from the end of the glass substrate to the end of the glass substrate end side in the image captured by the line camera 3 can be obtained in advance (set to U〇). Further, the arithmetic unit 4 s sets the distance from the portion of the image taken to the feature point of the image. In this calculation, for example, as long as the number of pixels from the end portion to the feature point in the image is counted, the distance in the real space of each pixel is multiplied by the number of pixels. The arithmetic unit 4 calculates the distance u from the end (side) of the glass substrate to the defect in the actual space by adding the position determined by the position of the line camera to the distance, i.e., 〇J~ 〇 160610. Doc •34· 201226888 Furthermore, in the example shown in FIG. 10, the distance from the end (four) of the glass substrate to the central portion 21 in the image is obtained by using the central portion 21 of the image 21 as the feature point. The case is an example. As a feature point, ~.p of 22a in another image 22 can be used. The end portion (side) of the glass substrate in real space can be obtained regardless of which central portion is used as the (four) point '. The distance u to the bubble. According to which one of the center part & '% is used as the feature point, the counting result of the number of pixels is different, but the difference is very small, and the distance u contains only the error that can be ignored. Point, it is also possible to use the characteristic point in the circumscribed rectangle 23 (for example, any vertices in the circumscribed rectangle 23). In this case, the distance 11 also contains only errors that can be ignored. After the step S11, the arithmetic device 4 counts The number of pixels p in the area is the area of the area occupied by the two overlapping images 21, 22 (the collection of the areas of the two images) (step S12) » Moreover, the arithmetic device 4 will be operated by steps S3, S11, S12 Find the h and P into the equation (6) Step S3) of the diameter of the bubble parallel to the transport direction. As shown in Fig. 1A, the long axis of the image is parallel to the transport direction line. The straight s is the long axis of the bubble. In the captured image, the long axis of the image and the transport direction line are not completely parallel, but since the two are almost parallel, the straight axis calculated by the step (1) is the straight line of the straight line. In the method, steps S5 to S10 are the same. That is, the operation unit 4 obtains % by subtracting the S calculated by the step 算出 calculated in step ( (step S5). 160610. Doc -35 - 201226888 yd calculates the equation (2) with the refraction angle β, thereby calculating the height direction position d of the bubble (step S6). Furthermore, the calculation device 4 determines the distance D from the reference surface 52 to the bubble in accordance with the position d of the bubble in the height direction (step S7). The surface is placed on the side of the transport roller 1 so that the surface to be determined as the distance to the bubble is disposed. In the case of a glass substrate, D=d may be sufficient. In the case where the glass substrate is disposed on the side opposite to the transport roller 1 on the surface on which the distance to the bubble is determined, the arithmetic unit 4 may determine the value of the distance D by setting D = T_d. Then, the arithmetic unit 4 calculates the length t of the diameter of the bubble in the direction orthogonal to the transport direction based on the image captured in step S2 (step S8). The calculation method of this t can be the same as step S8 in the first glass substrate inspection method. Then, the arithmetic unit 4 calculates the bubble ball diameter e of the bubble by calculating the cube root of (Sxt2) (step S9). Further, the processing of steps S3 to S9 is performed for each of the pair of elliptical image groups. Further, the arithmetic unit 4 detects bubbles having a distance D from the surface 52 of the glass substrate 5 of 1/3 or less. Further, the arithmetic unit 4 sequentially selects the bubble, and determines whether or not the relationship (i.e., 〇 x lxD16 + i5) between the distance D calculated by the selected bubble and the ball-converted diameter e (i.e., 〇 x lxD16 + i5) is satisfied (step S10). When the second glass substrate inspection method (steps S1 to S10 shown in FIG. 9) is performed on the glass substrate 51 (see FIG. 1) of the present invention, the bubble having a distance D from the surface 52 of T/2 or less is used. Each is also judged as "e$〇. 〇lxDi. The relationship of 6+15" was established. 160610. Doc-36-201226888 In the second glass substrate inspection method shown in FIG. 9, similarly to the second glass substrate inspection method (see FIG. 7), even if the image of the same bubble overlaps, the position of the bubble in the twist direction can be calculated. . Further, it is difficult to receive the influence of the vibration of the glass substrate to be conveyed, and the distance D from the surface 52 is τ/2 or less, and it is possible to accurately determine the distance between the distance D and the ball-converted diameter 6 of the bubble. (4) Whether it is established. Next, a method of inspecting the third glass substrate will be described. In the third glass substrate inspection method, the positional relationship between the light source 2 and the linear camera 3 (see Fig. 4) of the glass substrate 5 to be inspected is the same as that of the second glass substrate inspection method, and the description thereof is omitted. In the third glass substrate inspection method, the arithmetic unit 4 calculates the feature amount of the bubble in the glass substrate 5, and calculates the value by using the feature amount. However, in the second method of inspecting the glass substrate, the diameter s of the bubble is calculated as the feature amount, but in the method of inspecting the glass substrate, the ratio of the two diameters of the bubble is calculated. Specifically, the arithmetic unit 4 determines the ratio of the diameter of the bubble to the diameter of the transport direction in the direction of the direction of the parent in the transport direction as the characteristic amount of the bubble. In other words, when the diameter of the bubble diameter in the direction orthogonal to the conveyance direction is r! 'The diameter of the conveyance direction is "," 咏 is calculated as the feature amount. Hereinafter, Γ2/Γ1 is denoted as r. In the captured image, when the long axis of the image of the bubble is parallel to the transport, the u corresponds to the short diameter of the bubble, and _ is the long axis of the bubble. That is, the "long axis/short path" is calculated as a feature. The amount of cloth, that is, in the image, when the long axis of the image of the bubble and the direction of the transport direction are not in the situation of the witch, the (4) is substantially parallel, so the upper air can be 1606l0. Doc •37- 201226888 The shortness of the bubble, the above h is regarded as the long axis of the bubble. In other words, even when the long axis of the image of the bubble in the image is not completely parallel to the direction of the transport direction, the r calculated as the feature amount can be regarded as the "long axis/short path" of the bubble. Even so, r contains only errors that can be ignored, and does not affect the calculation of the position d of the bubble in the height direction. After calculating the characteristic amount of Γ as a bubble, the arithmetic unit 4 uses y to obtain yd (the moving distance of the bubble from the position where the first image is captured to the position where the second image is captured). Further, when calculating the feature amount γ, the arithmetic unit 4 calculates the feature amount based on the positional relationship of the two images that overlap each other using a predetermined calculation formula. The equation for calculating the feature amount r is determined in advance as a coordinate at a position corresponding to a feature point of the image on which the end 4 of the glass substrate is used as a reference, h described in the second glass substrate inspection method, and two overlapping images. The area is a function of the variable. The calculation formula for determining the feature amount r can be expressed, for example, by the following formula (7). r=blU2+b2h2+b3p2+b4uh+b5hp+b6up+b7u+b8h+b9p+bl〇(7) The variables u, h, p in the function and the formula (6) shown in the second glass substrate inspection method The variables u, h, and p are the same. That is, 'u' is the distance from the side of the glass substrate parallel to the transport direction to the center of the bubble. "The heart is the value obtained by the same calculation as the step S3 of the second glass substrate inspection method based on the image of the bubble. P is the image of the two images in the image captured by the film. Area (the area of the two image areas, specifically, the number of pixels in the image. Formula, number. Doc •38· 201226888 The characteristic amount r is easily affected by the position of the bubble in the width direction of the glass ribbon. Further, as described above, the position at which the self-contained glass ribbon is plate-cut to the glass substrate is generally fixed in the width direction of the glass ribbon. Therefore, it can be said that the characteristic amount r is also easily affected by the position of the bubble in the direction perpendicular to the transport direction in the glass substrate (in other words, the position of the bubble in the direction perpendicular to the stripe direction in the glass substrate). Therefore, the calculation formula including the above-described variable u (for example, the above formula is used for [calculation. The coefficients bl to b10 in the equation (7) are obtained in advance by the least square method. Specifically, 5, the bubble is used as a sample to measure r, u. Further, the glass substrate including the bubble as the sample is subjected to the same processing as the steps S1 to S3 described in the method of inspecting the glass substrate to obtain he again, and the image count obtained from the step "from now on" The number of pixels p in the area where the two images are combined. Prepare a plurality of bubbles that become samples, and obtain the r, u, and h P for each of the bubbles, and obtain a group of complex arrays u, u ' h, and p. In the group of ", u, 匕, p, the coefficient in the equation (7) can be obtained by the least square method. There is an association between r', uh, and P, and the equation (?) can be obtained by the least square method. The arithmetic unit 4 calculates ", h, p from the image obtained by photographing the glass substrate to be measured in the height direction of the bubble, and calculates r by substituting the equation (7). The computing device 4 will transfer the direction line 96 and pass the two in the captured image. The angle of the line of the center of the image is called the value of (4), and the calculation unit 4 calculates the % using h, u'r, and tane. The arithmetic unit 4 calculates the bubble using the angle ～ with the refraction angle β. The height direction position d. 160610. Doc-39-201226888 Fig. 11 is a flow chart showing an example of a third glass substrate inspection method. The same reference numerals as in the case of the first glass substrate inspection method or the second glass substrate inspection method are denoted by the same reference numerals as those in Fig. 7 or Fig. 9, and the description thereof is omitted. The operation until the p is obtained in step S12 (steps S1, S2, S3, S11, and S12) and the second glass substrate inspection method. After step S12, the arithmetic unit 4 calculates Γ by the substitution of h, u, and p obtained in steps S3, S1, and si2 into equation (7) (i.e., the direction of the bubble is orthogonal to the direction of transport) The ratio of the length of the diameter of the diameter in the transport direction is determined (step S21). Fig. 12 is an explanatory view showing an example of a glass substrate which is displayed in an image. The same elements as those in Fig. 10 are denoted by the same reference numerals as those in Fig. 10, and description thereof will be omitted. After step S21, the arithmetic unit 4 counts the sides of the sides of the two overlapping images 21, 22 which are orthogonal to the direction of the transport direction of the glass substrate (for & The number of pixels in the side orthogonal to the transport direction line. That is, the number of pixels of the portion indicated by the symbol D in Fig. 12 is counted. Moreover, the arithmetic unit 4 multiplies the distance in the real space of each pixel by the number of pixels (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. 12. Further, the arithmetic unit 4 obtains a side parallel to the direction corresponding to the transport direction of the glass substrate among the sides of the circumscribed rectangle, and a line formed by the lines of the pair of P blades 21a and 22a among the two images 21 and 22. The tangent of the angle 0 is tan0 (step S23). The crucible may also be referred to as an angle formed by the line connecting the center portions 2U, 22a of the two 21' 22 and the conveying direction line. Therefore, for example, the arithmetic unit 4 may determine the value of yc (refer to Fig. 17) in advance, and calculate Xec by the method described, and perform the equation 160610. Doc -40- 201226888 (3) Calculation, by which tane is calculated. Alternatively, tan9 ° may be calculated by another method. Next, the arithmetic unit 4 calculates yd using h, r, w, and tane calculated by the processing up to step 823 (step S24). Specifically, the arithmetic unit 4 calculates yd by performing the calculation of the equation (8) as shown below, that is, 〇yd=(hr · w)/(lr · tan0). (8) The arithmetic unit 4 uses the above yd and The predetermined angle of refraction p is used to calculate the height direction position d of the bubble (step S25). This calculation is the same as step S6 in the first glass substrate inspection method. After the step S25, the arithmetic unit 4 arranges the glass in the height direction of the bubble from the surface 52 which is the reference surface to the distance D from the bubble (step S7), and the surface which is the reference for determining the distance to the bubble faces the transport roller (one) In the case of the substrate, as long as 0 = (1), when the glass substrate is placed on the opposite side of the surface on which the distance to the bubble is determined, the surface is placed on the opposite side of the transport roller ,. =T_d determines the value of the distance D. This processing is the same as step S7 (see Fig. 9) in the second glass substrate inspection method (see Fig. 7) or the second glass inspection method (see Fig. 9). The device 4 calculates the length s of the diameter of the bubble parallel to the transport direction (step S4). The calculation of the length 3 of the diameter of the bubble parallel to the transport direction may be performed in accordance with the steps in the first glass substrate inspection method (see FIG. 7). The same method is performed. Alternatively, the same processing as that of the step S13 in the second glass inspection method (see FIG. 9) may be performed to calculate and transfer 160610. Doc •41- 201226888 The length S of the diameter of the bubble parallel to the direction. The subsequent processing is the same as steps S8 to S10 in the second glass substrate inspection method and the second glass substrate inspection method. In other words, the arithmetic unit 4 calculates the length t of the diameter of the bubble in the direction orthogonal to the transport direction based on the image captured in step S2 (step S8). Then, the arithmetic unit 4 calculates the bubble-converted diameter e of the bubble by calculating the cube root of (sxt2) (step S9). Further, the processing of steps S3 to S9 is performed for each of the groups of the pair of elliptical images. Further, the arithmetic unit 4 detects a bubble having a distance D from the surface 52 of the glass substrate 5 of T/2 or less. Further, the arithmetic unit 4 sequentially selects the bubble, and calculates the distance D and the ball conversion calculated for the selected bubble. The relationship between the diameter e and the formula (4) (ie 0. 01)^16+15) is established (step S10). When the third glass substrate inspection method (steps S1 to S10 shown in FIG. 11) is performed on the glass substrate 51 (see FIG. 1) of the present invention, the distance D from the surface 52 is also T/2 or less. Bubble, judged as "e^o. The relationship between oix D16+15" was established. In the third glass substrate inspection method shown in FIG. 11 , even if it is the same as the first glass substrate inspection method (see FIG. 7 ) or the second glass substrate inspection method (see FIG. 9 ), the image can be calculated by the same bubble. The position of the height direction of the overlapping bubbles. Further, it is difficult to receive the influence of the vibration of the glass substrate on the surface of the glass substrate. The distance D from the surface 52 is equal to or less than T / 2, and the distance between the distance D and the ball-converted diameter e of the bubble can be accurately determined. (4) Whether it is established. 160610. Doc-42-201226888 In each of the glass substrate inspection methods described above, the arithmetic unit is realized by, for example, a computer that operates in accordance with the program. For example, the computer can also follow the program as the arithmetic unit 4. Next, a method of manufacturing the glass substrate 51 (see Fig. 本) of the present invention will be described. The glass substrate 5 of the present invention shown in FIG. 1 is used, for example, by applying the above-described second glass substrate inspection method (see FIG. 7), the second glass substrate inspection method (see FIG. 9), and the glass ribbon manufactured by the floating method. In any one of the third glass substrate inspection methods (see FIG. 11), the distance D between the distance D and the bubble from the bottom surface is selected to be equal to or smaller than the ball diameter of the bubble, and the equation (4) is established. The bee ribbon 'can cut the glass substrate from the glass ribbon. In the case where any of the methods of the second glass substrate inspection method, the second glass substrate k inspection method, and the third glass substrate inspection method is applied to the glass ribbon, a glass ribbon may be used instead of the glass substrate. In the case of any of the above-described glass substrate inspection method, the second glass substrate inspection method, and the third glass substrate inspection method, the distance D from the bottom surface to the bubble is calculated based on the bottom surface, and the bubble is calculated. The ball is converted to the diameter e, and the bubble having the distance D of T/2 or less may be selected as the glass ribbon in which the formula (4) is established between the distance D and the ball-converted diameter e of the bubble. The glazing corresponding to the surface of the bottom surface can be prevented from the glass substrate in which the glass ribbons thus selected are cut in a sheet shape. In the case where the plate-shaped cut glass substrate is used as the transparent substrate of the liquid crystal display panel, the liquid crystal display panel can be manufactured by polishing the surface corresponding to the bottom surface of the glass ribbon toward the liquid crystal side. As a result, it is possible to manufacture a liquid crystal display panel in which the cell gap is uniform. 160610. Doc •43-201226888 In addition, the glass substrate of the present invention can be preferably used for a liquid crystal display panel that displays a stereoscopic image because the surface of the glass can be prevented from rising. Since the surface can be prevented from rising, it is possible to prevent the load from being concentrated on one portion of the glass substrate when the glass substrate is overlapped. Therefore, the breakage of the glass substrate can be prevented even in the case of overlapping the glass substrate. Further, as a method of producing a glass ribbon, there is a melting method. Regarding the glass ribbon produced by the melt method, any surface may be used as a reference surface for determining the distance D. Further, when calculating the distance D from the surface of the reference surface to the bubble, and calculating the ball-converted diameter e of the bubble, the distance between the distance D and the ball-converted diameter e of the bubble is selected for the bubble having the distance d of T/2 or less. (4) The glass ribbon to be formed is subjected to sheet-shaped cutting of the glass substrate in the same manner as described above, and the glass substrate 51 of the present invention (see FIG. When a glass substrate obtained by cutting a glass ribbon produced by a melt method into a transparent substrate of a liquid crystal display panel is used, the surface of the reference surface for determining the distance d is directed toward the liquid crystal side. It is enough to manufacture a liquid crystal display panel. In this case, a liquid crystal display panel having a uniform cell gap can also be manufactured. Further, in the case where a glass substrate obtained by cutting a glass ribbon produced by a melt method into a transparent substrate of a liquid crystal display panel is used, the polishing may not be performed. [Examples] Table 1 shows examples 1 to exemplified as examples of the present invention and examples i i to 20 as comparative examples. 160610. Doc -44· 201226888
[表l] Ο(μηι) f(D)(—) ε(μιη) 表面之隆起量 (μιη) 評估 例1 15.35 15.79 9.32 N.D. 〇 例2 25.69 16.80 16.64 N.D. 0 例3 43.02 19.11 11.92 N.D. 〇 例4 47.88 19.88 17.24 N.D. 〇 例5 70.07 23.97 15.50 N.D. 〇 例6 75.39 25.09 16.94 N.D. 〇 例7 100.47 30.97 24.85 N.D. 〇 例8 134.98 40.61 33.07 N.D. 〇 例9 157.93 47.93 35.85 N.D. 0 例10 175.10 53.84 45.01 N.D. 〇 例11 3.34 15.07 20.35 3.1 X 例12 6.69 15.21 23.91 3.0 X 例13 24.32 16.65 31.03 1.5 X 例14 28.12 17.08 51.30 2.3 X 例15 44.23 19.30 53.48 1.6 X 例16 52.74 20.69 85.92 3.5 X 例17 59.74 21.95 145.14 6.2 X 例18 74.94 24.99 171.89 3.7 X 例19 78.43 25.74 141.41 2.1 X 例20 100.78 31.05 166.68 1.9 X 例1〜20之試料係使用藉由浮式法成形後切斷而成之玻璃 基板（旭硝子公司製造之無鹼玻璃「AN100」）。 160610.doc •45- 201226888 表1之f(D)之欄分別記載相對於例1〜2〇之試料之自表面 至氣泡為止之距離D之值的£(〇)=0.0^〇16+15之值。 表1之氣泡之球換算直挫e之棚記載使用上述方法算出之 例1〜20之試料之氣泡之球換算直徑e。 表1之表面之隆起量之欄記載使用奥林巴斯股份有限公 司（Olympus Corporation)製造之3D雷射顯微鏡（機種名： LEXT OLS 3100 MODEL: OLS31-SU)來測定例卜汕之試 料之與表面垂直之方向之試料表面之隆起量的結果。表中 標記為N.D之試料係表示隆起量為測定界限以下（〇.丨以 下）。 表1之評估之欄記載進行自玻璃基板之表面至氣泡為止 之距離D與氣泡之球換算直徑e之間是否滿足f(D)之判 定之結果。於滿足egf(D)之情形是記載〇，不滿足e$f(D) 之情形時記載X。 由表1可明白’滿足eg 〇·〇1χβΐ·6+15之實施例之例卜⑺ 之試料中，與試料之表面垂直之方向之試料表面之隆起量 為測定界限以下。對此，不滿足e$〇〇lxDi.6+15之實施例 之例11〜例20之試料中，與試料之表面垂直之方向之試料 表面之隆起量為1.5〜6.2 μιη。 因此，可明白，藉由使自玻璃基板之表面至氣泡為止之 距離D與氣泡之球換算直徑e滿足eg0.01xDi.6+15而可將玻 璃基板之表面之隆起量設為測定界限以下。 已詳細地且參照確定之實施態樣對本申請案進行了說 月但業者當明白可在不脫離本發明之精神與範圍之情況 160610.doc -46 - 201226888 下添加各種變更或修正。 本申請案係基於2010年12月9曰申請之曰本專利申請案 (日本專利特願2G1G-275G49)者’其内容以參照之形式併入 本文中。 [產業上之可利用性] 本發明可較佳地應用於例如作為液晶顯示面板之透明基 板而利用之玻璃基板。 【圖式簡單說明】 圖1係表示本發明之玻璃基板之侧視圖之例之說明圖。 圖2係表示氣泡之形狀之說明圖。 圖3係表示自上方觀察氣泡之狀態之說明圖。 圖4係表示檢查於玻璃基板中之自表面至氣泡為止之距 離D與氣泡之球換算直徑e之間式（4)是否成立之檢查系統 之構成例的模式圖。 圖5A係表示中央線之說明圖。 圖5B係表示圖像内之相當於中央線之線之說明圖。 圖6係表示玻璃基板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7係表示第1玻璃基板檢查方法之例之流程圖。 圖8係表示兩個重合之像之外切矩形之區域之說明圖。 圖9係表示第2玻璃基板檢查方法之例之流程圖。 圖1 〇係表示於圖像内顯現之玻璃基板之例之說明圖。 圖11係表示第3玻璃基板檢查方法之例之流程圖。 圖12係表示顯現於圖像内之玻璃基板之例之說明圖。 160610.doc -47· 201226888 法之說明圖。 之缺陷之圖像之例的說 圖13A係模式性地表示第1測定方 圖13B係表示以第1測定方法拍攝 明圖。 圖14A係模式性地表示第2測定方法之說明圖。 圖MB係表示以第2測定方法拍攝之缺陷之圖像之例之說 明圖。 圖15A係模式性地表示第3測定方法之說明圖。 圖15B係表不以第3測定方法拍攝之缺陷之圖像之例之說 明圖。 圖16係表示所搬送之玻璃基板内之缺陷被線型相機拍攝 時之位置之說明圖。 圖17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 線型相機 4 運算裝置 5 玻璃基板 51 玻璃基板 57 氣泡 81 線型相機 81a 線型相機 81b 線型相機 82 玻璃基板 160610.doc • 48-[Table l] Ο(μηι) f(D)(—) ε(μιη) Surface bulge (μιη) Evaluation Example 1 15.35 15.79 9.32 ND Example 2 25.69 16.80 16.64 ND 0 Example 3 43.02 19.11 11.92 ND Example 4 47.88 19.88 17.24 ND Example 5 70.07 23.97 15.50 ND Example 6 75.39 25.09 16.94 ND Example 7 100.47 30.97 24.85 ND Example 8 134.98 40.61 33.07 ND Example 9 157.93 47.93 35.85 ND 0 Example 10 175.10 53.84 45.01 ND Example 11 3.34 15.07 20.35 3.1 X Example 12 6.69 15.21 23.91 3.0 X Example 13 24.32 16.65 31.03 1.5 X Example 14 28.12 17.08 51.30 2.3 X Example 15 44.23 19.30 53.48 1.6 X Example 16 52.74 20.69 85.92 3.5 X Example 17 59.74 21.95 145.14 6.2 X Example 18 74.94 24.99 171.89 3.7 X Example 19 78.43 25.74 141.41 2.1 X Example 20 100.78 31.05 166.68 1.9 X The sample of Examples 1 to 20 is a glass substrate which has been cut by a floating method and formed into an alkali-free glass "AN100" manufactured by Asahi Glass Co., Ltd. ). 160610.doc •45- 201226888 The column f(D) of Table 1 respectively describes the value of the distance D from the surface to the bubble of the sample of Examples 1 to 2〇(〇)=0.0^〇16+15 The value. The ball of the bubble of Table 1 is converted into a shed of e. e. The ball diameter of the bubble of the sample of Examples 1 to 20 calculated by the above method is converted. The column of the amount of swelling of the surface of Table 1 describes the use of a 3D laser microscope (model name: LEXT OLS 3100 MODEL: OLS31-SU) manufactured by Olympus Corporation to determine the sample of the sample. The result of the amount of bulging of the surface of the sample perpendicular to the surface. The sample labeled N.D in the table indicates that the amount of swelling is below the measurement limit (〇.丨). The evaluation column of Table 1 describes the result of determining whether or not f(D) is satisfied between the distance D from the surface of the glass substrate to the bubble and the ball-converted diameter e of the bubble. In the case where egf (D) is satisfied, it is described, and when it is not satisfied, e is described as X. In the sample of the example (7) which satisfies the example of eg 〇·〇1χβΐ·6+15, the amount of swelling of the surface of the sample perpendicular to the surface of the sample is below the measurement limit. On the other hand, in the samples of Examples 11 to 20 which did not satisfy the example of e$〇〇lxDi.6+15, the amount of the surface of the sample perpendicular to the surface of the sample was 1.5 to 6.2 μm. Therefore, it can be understood that the amount of swelling of the surface of the glass substrate can be made equal to or less than the measurement limit by making the distance D from the surface of the glass substrate to the bubble and the ball-equivalent diameter e of the bubble satisfying eg0.01xDi.6+15. The present application has been described in detail with reference to the embodiments of the present invention, and it is understood that various changes or modifications may be added without departing from the spirit and scope of the invention. The present application is based on the Japanese Patent Application No. 2G1G-275G49 filed on Dec. 29, 2010, the content of which is hereby incorporated by reference. [Industrial Applicability] The present invention can be preferably applied to, for example, a glass substrate which is used as a transparent substrate of a liquid crystal display panel. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is an explanatory view showing an example of a side view of a glass substrate of the present invention. Fig. 2 is an explanatory view showing the shape of a bubble. Fig. 3 is an explanatory view showing a state in which bubbles are observed from above. Fig. 4 is a schematic view showing a configuration example of an inspection system in which the equation (4) is established between the distance D from the surface to the bubble in the glass substrate and the sphere-converted diameter e of the bubble. Fig. 5A is an explanatory view showing a center line. Fig. 5B is an explanatory view showing a line corresponding to a center line in an image. Fig. 6 is an explanatory view showing the relationship between the direction of the long axis of the bubble in the glass substrate and the conveyance direction of the conveyance roller 1. Fig. 7 is a flow chart showing an example of a first glass substrate inspection method. Fig. 8 is an explanatory view showing a region in which two overlapping images are cut out of a rectangular shape. Fig. 9 is a flow chart showing an example of a second glass substrate inspection method. Fig. 1 is an explanatory view showing an example of a glass substrate which is visualized in an image. Fig. 11 is a flow chart showing an example of a third glass substrate inspection method. Fig. 12 is an explanatory view showing an example of a glass substrate which is displayed in an image. 160610.doc -47· 201226888 Illustration of the law. Example of the image of the defect Fig. 13A schematically shows the first measurement side. Fig. 13B shows the first measurement method. Fig. 14A is an explanatory view schematically showing a second measurement method. Fig. MB is an explanatory diagram showing an example of an image of a defect photographed by the second measurement method. Fig. 15A is an explanatory view schematically showing a third measuring method. Fig. 15B is an explanatory view showing an example of an image of a defect which is not captured by the third measuring method. Fig. 16 is an explanatory view showing a position at which a defect in the conveyed glass substrate is taken by a line camera. Fig. 17 is an explanatory diagram of the shooting distance yc. [Main component symbol description] 1 Transport roller 2 Light source 3 Linear camera 4 Arithmetic device 5 Glass substrate 51 Glass substrate 57 Bubble 81 Linear camera 81a Linear camera 81b Linear camera 82 Glass substrate 160610.doc • 48-