201035541 六、發明說明： 【發明所屬之技術領域】 本發明係有關於對形成於基板上之圖型的形狀進行攝 像檢查的裝置及方法。 【先前技術】 在電漿顯示裝置或液晶顯示裝置或太陽能電池等的製 造中，利用檢查光學系統203拍攝形成於是被檢查物之基 0 板基板109上之圖型的形狀，再檢査此圖型的形狀。在此 攝像和檢査中，使用第31圖所示的攝像檢查裝置。 有一種情形，係利用和檢查光學系統203爲不同而具 備的覆查光學系統204再拍攝藉由此檢查而被判斷爲第一 缺陷群之缺陷的部分。藉由如此再拍攝，而可得到更詳細 的資訊。 在此，利用覆査光學系統204拍攝第一缺陷群的周邊， 再根據其攝像結果，作爲第二缺陷群，更高精度地檢測出 〇 缺陷，以上稱爲覆查。 此覆查光學系統204有和檢查光學系統203光學解析 度相異的情況、或構成光學系統之透鏡的數値孔徑相異的 情況 '或是可進行彩色攝像之相機的情況。在此檢查裝置 之基板109之兩側的工作台，設置有X軸軌道2 07。檢查光 學系統203和覆查光學系統204安裝於Y軸可動件201。成 爲Y軸可動件201之移動路徑的Y軸軌道206設置於X軸 可動件208。此X軸可動件208在跨X軸軌道207間之基板 201035541 109的上方位置沿著X軸軌道207移動。 利用藉由該覆查所得之影像，可得到與利 系統203而被判斷爲第一缺陷群之缺陷的部分 細的資訊。然後，使用所得之資訊，作爲第二 高精度地進行該部分實際上是否是缺陷、或屬 的種類的判斷。 可是’在此檢查裝置中，因爲在藉檢查光 之檢查動作結束後進行該覆查動作，所以具有 〇 之總時間長的問題。例如，在專利文獻1中， 置，其具有一種功能，即將半導體LSI作爲對 . 手段檢測缺陷，並利用同一裝置內之別的電子 察該檢測結果。 爲了解決此問題，如第3 2圖所示，在專利 査裝置210中，係利用檢査光學系統203拍攝 查物的基板109,再轉移至和具備有覆查光學系 〇 查裝置210爲不同的覆查裝置211。藉由如此轉 式虛擬地平行進行2種動作，來使藉檢查光學 檢查和藉覆查光學系統204的覆查作動作。在 揭示一種裝置，其具有一種功能，即將半導體 象，以光學手段檢測缺陷，並將結傳送至同一 測資訊傳送至別的裝置，再利用別的電子光學 察。 [專利文獻1 ]特開昭60 - 2 1 8845號公報 用檢查光學 有關之更詳 缺陷群，更 於何種缺陷 學系統203 檢查和覆査 揭示一種裝 象，以光學 光學系統觀 文獻1的檢 並檢查被檢 統204之檢 移，以管線 系統203的 專利文獻2 LSI作爲對 裝置或將檢 系統來作觀 201035541 [專利文獻2]特開昭5 8 - 3 3 1 54號公報 【發明內容】 發明所欲解決之課題 在專利文獻2，雖然可縮短檢查時間，但是因爲檢查裝 置的面積需要約2倍，所以具有整體上裝置大型化的問題。 隨著被檢查物大型化，藉由這種方法之解決將變得困難。 本發明之目的在於提供一種攝像檢查裝置及攝像檢查 方法，可縮短檢查後之覆查動作的時間，而且亦可應付被 Ό 檢查物之大型化。 解決課題之手段 本發明的攝像檢查裝置，其特徵爲具備：載置被檢查 ' 物的載置部；檢查光學系統，係拍攝該被檢查物並將缺陷 檢測爲第一缺陷群；檢査光學系統移動機構，係使該被檢 查物和該檢查光學系統在X方向相對移動；覆查光學系 統’係拍攝以該檢査光學系統所檢測出的該第一缺陷群， 〇 並將缺陷檢測爲第二缺陷群；覆查光學系統移動機構，係 一面使該被檢查物和該覆查光學系統在該X方向相對移 動，一面使該檢查光學系統和該覆查光學系統在和該又方 向正父的Y方向相對移動；及控制機構，係在該檢查光學 系統的攝像中使該覆查光學系統拍攝該第一缺陷群，並將 該第二缺陷群和覆查光學系統未拍攝的該第一缺陷群檢測 爲缺陷。 具體而言’特徵爲：該覆査光學系統之Y方向的相對 201035541 移動量和該檢査光學系統之γ方向的相對移動量相等。 又’特徵爲：具有複數個該檢查光學系統和複數個該覆查 光學系統；在該Υ方向，該覆查光學系統和該檢查光學系 統獨立’並可在該被檢查物的整個寬度移動。又，特徵爲： 具有複數個該檢査光學系統和一個該覆查光學系統；在該 Υ方向’該覆査光學系統和該檢査光學系統獨立，並可在 該被檢查物的整個寬度移動。又，特徵爲：具有複數個該 檢查光學系統和複數個該覆査光學系統；在該γ方向，該 〇 複數個檢查光學系統可對該檢查光學系統相對地移動。 本發明的攝像檢查方法，其使檢査光學系統和覆查光 學系統在該檢査對象物上在X方向相對移動，同時使在Υ 方向移動’並利用該覆査光學系統拍攝根據該檢查光學系 統的攝像結果而檢測出之第一缺陷群的缺陷，來檢測第二 缺陷群的缺陷，該攝像檢查方法的特徵爲：每當該檢査光 學系統結束Υ方向的一掃描，對至那時爲止所檢測出之第 〇 一缺陷群中之未被該覆查光學系統拍攝的缺陷，以成爲可 利用該覆查光學系統拍攝之缺陷的個數變成最大之程序表 的方式計算藉該覆查光學系統之攝像的覆查程序表；在該 檢查光學系統之Υ方向的下一掃描，根據該覆查程序表使 該覆查光學系統拍攝。 具體而言，特徵爲：該覆查程序表的計算，係根據所 檢測出之該缺陷的特徵量決定優先順位，再根據該優先順 位及缺陷的個數算出評估値，並以該評估値成爲最大的方 201035541 式來被計算。又，特徵爲：該缺陷的特徵量係至少包含缺 陷之反射率、面積、形狀及座標之任一個。又，特徵爲： 該覆查程序表的計算，係比較第一程序表和第二程序表， 並被計算爲將評估値高的程序表當作覆查程序表，而該第 一程序表係根據所檢測出之第一缺陷群之缺陷的特徵量決 定優先順位，再根據該優先順位及缺陷之個數算出程序表 的評估値，並以該評估値成爲最大的方式來被計算，該第 二程序表係從該該第一程序表將優先順位高的缺陷除外 Ο 後，再根據該優先順位及缺陷的個數算出評估値，並以該 評估値成爲最大的方式來被計算。又，特徵爲：在將所檢 測出之複數個該缺陷之特徵量的總和設爲P，並將所檢測 出之該缺陷的個數設爲N的情況，根據201035541 VI. Description of the Invention: [Technical Field] The present invention relates to an apparatus and method for performing an image inspection of a shape of a pattern formed on a substrate. [Prior Art] In the manufacture of a plasma display device, a liquid crystal display device, a solar cell, or the like, the shape of the pattern formed on the base plate substrate 109 of the object to be inspected is imaged by the inspection optical system 203, and the pattern is checked. shape. In this imaging and inspection, the imaging inspection device shown in Fig. 31 is used. There is a case where the inspection optical system 204, which is different from the inspection optical system 203, is used to take a portion which is judged to be a defect of the first defect group by the inspection. More detailed information can be obtained by shooting again. Here, the periphery of the first defect group is imaged by the inspection optical system 204, and based on the imaging result, the defect defect is detected more accurately as the second defect group, which is referred to as a review. The review optical system 204 has a case where the optical resolution of the optical system 203 is different, or a case where the number of apertures of the lenses constituting the optical system is different, or a camera capable of color imaging. The table on both sides of the substrate 109 of the inspection apparatus is provided with an X-axis rail 206. The inspection optical system 203 and the inspection optical system 204 are mounted to the Y-axis movable member 201. A Y-axis rail 206 that becomes a moving path of the Y-axis movable member 201 is provided to the X-axis movable member 208. The X-axis movable member 208 moves along the X-axis rail 207 at a position above the substrate 201035541 109 between the X-axis rails 207. By using the image obtained by the review, it is possible to obtain a portion of the information which is determined to be a defect of the first defect group by the profit system 203. Then, using the obtained information, as a second high-precision determination as to whether or not the portion is actually a defect or a genus. However, in this inspection apparatus, since the review operation is performed after the inspection operation by the inspection light is completed, there is a problem that the total time of 〇 is long. For example, Patent Document 1 has a function of detecting a defect by using a semiconductor LSI as a means, and using the other electrons in the same device to detect the detection result. In order to solve this problem, as shown in FIG. 3, in the patent check device 210, the substrate 109 on which the object is inspected by the inspection optical system 203 is transferred to and different from the inspection optical system inspection device 210. The device 211 is reviewed. By performing two kinds of operations in parallel in this way, the inspection optical inspection and the review of the optical system 204 are performed. A device is disclosed which has a function of optically detecting a defect by a semiconductor image and transmitting the junction to the same measurement information to another device for use by another electron optical inspection. [Patent Document 1] Japanese Laid-Open Patent Publication No. SHO 60-201-8845 discloses a more detailed defect group related to inspection optics, and more than what kind of defect system 203 inspection and review reveals a type of image, and the optical optical system The inspection and inspection of the inspection system 204 is performed, and the patent document 2 LSI of the pipeline system 203 is used as a pair of devices or a system to be inspected. 201035541 [Patent Document 2] Japanese Patent Laid-Open No. Hei. No. 5 8 - 3 3 1 54 [Invention [Problem to be Solved by the Invention] In Patent Document 2, the inspection time can be shortened. However, since the area of the inspection apparatus needs to be doubled, there is a problem that the size of the apparatus is increased as a whole. As the object to be inspected becomes larger, it will become difficult to solve by this method. SUMMARY OF THE INVENTION An object of the present invention is to provide an image pickup inspection apparatus and an image inspection method which can shorten the time for the inspection operation after inspection, and can also cope with an increase in the size of the inspection object. Means for Solving the Problem An imaging inspection apparatus according to the present invention includes: a mounting portion on which an object to be inspected is placed; an inspection optical system that captures the object to be inspected and detects a defect as a first defect group; and an inspection optical system The moving mechanism moves the inspection object and the inspection optical system relative to each other in the X direction; the inspection optical system captures the first defect group detected by the inspection optical system, and detects the defect as a second a defect group; the optical system moving mechanism is configured to move the inspection object and the inspection optical system relative to each other in the X direction while the inspection optical system and the inspection optical system are in the direction of the parent The Y direction is relatively moved; and the control mechanism causes the review optical system to capture the first defect group in the imaging of the inspection optical system, and the first defect group and the first defect that is not photographed by the inspection optical system The group is detected as a defect. Specifically, the feature is that the relative amount of movement of the Y-direction of the inspection optical system in the Y direction is equal to the relative movement amount of the inspection optical system in the γ direction. Further, the feature is that there are a plurality of the inspection optical systems and a plurality of the inspection optical systems; in the meandering direction, the inspection optical system and the inspection optical system are independent and movable over the entire width of the inspection object. Further, it is characterized in that: the plurality of inspection optical systems and one of the inspection optical systems; the inspection optical system and the inspection optical system are independent in the Υ direction, and are movable over the entire width of the inspection object. Further, it is characterized in that a plurality of the inspection optical systems and a plurality of the inspection optical systems are provided; in the γ direction, the plurality of inspection optical systems can relatively move the inspection optical system. In the imaging inspection method of the present invention, the inspection optical system and the inspection optical system are relatively moved in the X direction on the inspection object while moving in the Υ direction, and the inspection optical system is used to image the optical system according to the inspection optical system. The defect of the first defect group is detected by the imaging result, and the defect of the second defect group is detected. The image inspection method is characterized in that each time the inspection optical system ends a scan in the x direction, it is detected until then. The defect in the first defect group that is not photographed by the review optical system is calculated by the method of the inspection optical system in such a manner that the number of defects photographed by the inspection optical system becomes the largest A review program table for imaging; the next scan in the direction of the inspection optical system is used to capture the review optical system based on the review schedule. Specifically, the calculation is that the calculation of the review program table determines the priority order based on the detected feature quantity of the defect, and then calculates the evaluation 根据 based on the priority order and the number of defects, and becomes the evaluation 値The largest party 201035541 is to be calculated. Further, it is characterized in that the feature quantity of the defect includes at least one of a reflectance, an area, a shape, and a coordinate of the defect. Further, the method is characterized in that: the calculation of the review program table compares the first program table and the second program table, and is calculated as the program table having the high evaluation value as the review program table, and the first program table system is Determining a priority order based on the detected feature quantity of the defect of the first defect group, and then calculating an evaluation 程序 of the program table based on the priority order and the number of defects, and calculating the maximum amount of the evaluation ,, the first The second program table excludes the defect having the highest priority from the first program table, and then calculates the evaluation 根据 based on the priority order and the number of defects, and calculates the evaluation 値 to be the largest. Further, the feature is that the total of the detected feature amounts of the plurality of defects is P, and the number of detected defects is N,
E=k P+j . N 此外，k是P的加權係數，j是N的加權係數。 計算程序表的評估値E，並將該評估値E大的程序表 Q 計算爲覆查程序表。 本發明之攝像檢査方法，其特徵爲：具有第一缺陷群 檢測步驟，係一面使被檢查物和檢查光學系統在X方向相 對移動，一面以該檢査光學系統拍攝該被檢查物，並將缺 陷檢測爲第一缺陷群；覆査步驟，係一面使被檢査物和覆 查光學系統在該X方向相對移動，一面以該覆查光學系統 拍攝在該第一缺陷群檢測步驟所檢測出之的缺陷，並將缺 陷檢測爲第二缺陷群；及缺陷檢測步驟，係將該第二缺陷 201035541 群和在該覆查步驟未拍攝到之該第一缺陷群檢測爲缺陷； 該檢查光學系統和該覆査光學系統係在該X方向獨立地相 對移動，在和該X方向正交的Y方向協同地相對移動，而 同時進行該第一缺陷群檢測步驟和該覆查步驟。 發明之效果 若依據該構成，因爲一面使被檢查物和覆查光學系統 在X方向相對移動，一面利用覆查光學系統移動機構使檢 查光學系統和覆查光學系統在γ方向相對移動，一面在檢 0 查光學系統之攝像掃描中在覆査光學系統拍攝缺陷的周 邊，所以和以往相比，可縮短檢査後之覆查動作的時間。 【實施方式】 以下，參照圖面詳細說明本發明之實施形態。此外， 在以下的說明，對相同之構成附加相同的符號來作說明。 (第1實施形態） 第1圖〜第17圖表示本發明之第1實施形態。 Q 在被設置有被檢查物之基板109的攝像檢查裝置200 之工作台上，兩側設有X軸軌道107。X軸可動件108被設 置成跨在此兩側的X軸軌道107。 具體而言，X軸軌道107由進給螺桿所構成，和此進給 螺桿螺合的螺帽設置於X軸可動件108。因應於X軸可動件 108的移動量而利用馬達旋轉驅動作爲X軸軌道107的進給 螺桿。 在X軸可動件108上，在和X軸軌道107交叉的方向 201035541 設置第一 Y軸軌道106和第二Y軸軌道105。第一 Y軸可動 件101設置於第一γ軸軌道106。在第一 Υ軸可動件101 上，將檢查光學系統103安裝成拍攝基板109。第二Υ軸可 動件102設置於第二Υ軸軌道105。在第二Υ軸可動件102 上，將覆査光學系統1Ό4安裝成拍攝基板109。 具體而言，第一、第二Υ軸軌道106、105由進給螺桿 所構成，和此進給螺桿螺合的螺帽設置於第一 Υ軸可動件 101、第二Υ軸可動件102。因應於第一 Υ軸可動件101的 〇 移動量而利用馬達旋轉驅動作爲第一 Υ軸軌道1 0 6的進給 螺桿。因應於第二Υ軸可動件102的移動量而利用馬達旋 轉驅動作爲第二Υ軸軌道105的進給螺桿。 在本發明中，以覆查光學系統1 04檢測出在以檢查光 學系統103所檢測出之是缺陷的第一缺陷群中更重要者來 作爲第二缺陷群。然後，作爲計算裝置301的檢查結果， 將第二缺陷群和未被覆查的第一缺陷群當作在該檢查中的 ◎ 缺陷。 控制機構300控制使被載置於載置部的基板109和檢 查光學系統103在X方向相對移動的檢查光學系統移動機 構、及使檢查光學系統103和覆查光學系統104在Υ方向 相對移動之覆查光學系統移動機構的動作。藉由此控制機 構300的控制，一面使基板109和覆查光學系統104在X 方向相對移動，一面在檢查光學系統103的攝像掃描中執 行覆查光學系統104的位置控制並進行拍攝。此控制機構 -10- 201035541 300由以微電腦爲主要部分的計算裝置301所構成。 按動作具體說明此計算裝置之構成。 首先，說明此攝像檢查裝置200在X方向的動作。 檢查動作如第2(a)、（b)圖所示，將基板109固定於攝 像檢查裝置200的工作台，藉由使X軸可動件1〇8沿著107 移動，而使檢查光學系統103和覆查光學系統1〇4在X方 向移動。 0 此外’在此，雖然列舉如第1圖所示構成的情況說明 攝像檢查裝置200，但是在如第3圖及第4(a)、（b)圖所示 以固定檢査光學系統103和覆查光學系統104並驅動基板 109之方式構成攝像檢查裝置2 00的情況亦相同。在第3 圖的攝像檢査裝置200，在工作台之Y軸方向的中央附近 隔著間隔設置X軸軌道107，基板109沿著X軸軌道107 在X軸方向移動。裝載檢查光學系統103和覆査光學系統 104的X軸可動件108被固定於工作台之X軸方向的中央 Q 附近。而，在基板109通過X軸可動件108的內側時進行 拍攝。 又，在以下情況亦相同，即攝像檢查裝置200如第 5(a)、（b)圖所示以藉由使基板109在X軸方向移動，同時 使X軸可動件108在X軸方向移動，而在X軸方向驅動檢 查光學系統103和覆查光學系統104的方式構成。 如第1圖〜第5圖所示，使基板109和檢査光學系統103 相對地移動，而在X軸方向掃描基板109。 -11 - 201035541 接著’說明攝像檢查裝置200在Y方向的動作。 因爲檢查光學系統1〇3的視野一般比基板1〇9之寬度 更小’所以無法在一次的動作中掃描基板1〇9的整個面。 因而’在本發明的攝像檢査裝置200，一面往γ方向進行 節距進給動作’一面使X軸可動件1〇8在X軸方向移動， 進行複數次的掃描，藉此掃描基板丨09的整個面。爲此， 本發明的攝像檢查裝置200具有檢查光學系統1〇3和覆查 光學系統1〇4進行往γ方向之移動所需的機構。 Ό 說明攝像檢査裝置20 0的覆査光學系統1〇4。 覆査光學系統104在X方向中係設置成和檢査光學系 統103 —體。此覆查光學系統1〇4如第1圖〜第5圖所示， 和檢查光學系統103獨立而被設置成具有可在γ方向移動 的機構。覆査光學系統104之可移動的範圍比基板109在Y 方向的寬度更大。 接著，說明覆査的動作。 〇 在覆査光學系統104的攝像，可和在檢查光學系統103 中的檢查動作同時進行。在檢查光學系統103中之檢查的 掃描進行中時，如上述所示，在X方向，在基板109和檢 査光學系統103具有相對速度（以下作爲X方向的相對速度） 之狀態，移動掃描檢査範圍。在此，因爲覆查光學系統104 在X方向是和檢查光學系統103 —體，所以覆查光學系統 1 04亦在對基板1 09具有相對速度之狀態，一樣地移動於檢 査範圍中。即，在X方向，覆査光學系統104的視野通過 -12- 201035541 基板109的整個寬度。 此外，在檢查光學系統103和覆查光學系統1〇4被設 置於在X方向偏移之位置的情況，需要使覆查光學系統1〇4 在X方向之移動範圍增加僅該偏移量。又，在Y方向，覆 查光學系統104本身具有移動機構，因爲其可移動之範圍 比基板109更大，所以覆查光學系統104可向基板1〇9的 任意位置移動。 依此方式，覆查光學系統104可將基板109的任意位 〇 置抓入視野並進行拍攝。在本實施形態，覆查光學系統1 04 雖然可在Y方向靜止而進行拍攝，但是在X方向在和基板 1 09之間無法相對上靜止地通過，所以藉由測量時序而進行 拍攝，而得到第一缺陷群之缺陷的影像。 在第6圖表示此時之動作例的流程圖。 首先，在步驟S601，使用檢查光學系統103進行基板 109之第1行的檢查掃描。根據裝置之構成，亦可在檢查光 Q 學系統103通過基板109之缺陷之上的瞬間掌握該缺陷的 位置。但是，在此爲了簡化說明，當作可在第1行之檢查 結束的時間點，掌握位於該掃描線上之缺陷的位置。在步 驟S601，掌握缺陷的位置，並檢測爲第一缺陷群。 在步驟S602，和檢查光學系統103對基板109之第2 行的掃描同時，對在步驟S601所檢測出之第一缺陷群之缺 陷的位置，利用計算求得要按照何種順序使用覆査光學系 統104進行第1行的覆查。 -13- 201035541 此外，未必需要覆査全部之第一缺陷群的缺陷，亦未 必可覆查全部之第一缺陷群的缺陷。因而，在此在需要覆 查之第一缺陷群的缺陷中，設置優先順位抽出可覆查之第 一缺陷群的缺陷，並製作用以移動覆查光學系統104的覆 查程序表。製作覆查程序表的具體例將後述。 若在步驟S602之覆查程序表製作結束，在步驟S603， 藉檢查光學系統103進行基板109之第2行的檢查掃描（步 驟S 603 — a)。和此檢查掃描（步驟S603 — a)同時平行地根據 C1 在步驟S 602所計算並製作的程序表覆查在第1行所檢測出 之第一缺陷群（步驟S603 — b)。步驟S604以後，一面逐次 挪移1行，一面進行此重複動作。 即，在步驟S604，對由在步驟S603所檢測出之第2 行的第一缺陷群中需要覆查之第一缺陷群的缺陷、和雖然 在步驟S601檢測出但是在步驟S603— a未覆查之第1行之 第一缺陷群的缺陷所合倂之第一缺陷群的集合，計算覆查 〇 程序表。在此’雖然在步驟S601檢測出，但是在步驟S603 - a未覆査之第1行之第一缺陷群的缺陷，是儘管需要覆查 卻由於程序表的關係而未覆查的第一缺陷群。 在步驟S605 ’同時平行地執行第3行的檢查掃描（步驟 S605 — a)、和根據在步驟S 604所製作的程序表之第1行及 第2行的覆査（步驟S605 - b)。步驟S606、步驟S607亦一 樣。 即，在步驟S606，對由在步驟S605所檢測出之第3 -14- 201035541 行的第一缺陷群中需要覆查者、和在步驟S601、步驟S603 一 a、步驟S 605 - a所檢測出之儘管需要覆査卻由於程序表 的關係而未覆查之第1〜第3行的第一缺陷群所合倂之第一 缺陷群的集合，g十算覆査程序表。 在步騾S607’同時平行地執行第4行的檢查掃描（步驟 S607-a)、和根據在步驟S606所製作的程序表之第丨~第3 行的覆查（步驟S607 — b)。 步驟S 608表示在檢查動作中之最後行（第N行）的動 C) 作。 在本實施形態，因爲定義爲至檢査掃描結束爲止無法 確定最後行之第一缺陷群之缺陷的位置，所以在此掃描線 所檢測出之第一缺陷群的缺陷無法和檢查同時覆查。因 此，在最後行之掃描檢查結束後需要覆查之第一缺陷群中 有覆查未結束之缺陷的情況，在步驟S 609進行不是和檢查 掃描同時平行進行之一般的覆查。 Q 在第8(a)圖表示檢查光學系統1〇3之動作的軌跡。 701表示檢查光學系統103之掃描第1行的軌跡，706 表示在掃描第1行所檢查的區域，以黑圓所示之705表示 所檢測出之第一缺陷群。一樣地，702〜704是檢查光學系統 103之掃描第2〜第4行的軌跡，707 ~7 09是在掃描第2〜第4 行所檢查的區域。 在第8(b)圖表示覆査光學系統1〇4之動作的軌跡。 710是覆查光學系統104之掃描第2行的軌跡，711是 -15- 201035541 覆査光學系統104之掃描第3行的軌跡，712是覆查光學系 統104之掃描第4行的軌跡。 此外’在此第8(a)、（b)圖中，檢查是進行往返2次共 4次的掃描。和第6圖所示的例子一樣，在各掃描結束的 時間點’當作可旱握則掃描之f了之第一缺陷群之缺陷的位 置。在此情況’可和檢査第N行之掃描同時覆查的第一缺 陷群是在第N行之前一行（第（N— 1)行）所檢測出者。實際 0 上’可掌握第一缺陷群之缺陷之位置的時序依構成而爲各 式各樣。 說明製作覆查程序表之具體例。 從和檢查之掃描同時實施覆查之目的可知，必然地由 第一缺陷群的座標唯一地決定覆查的順序。關於各個第一 缺陷群’因爲可取得覆查或不覆查之2種，所以在覆查N 個第一缺陷群時之覆査程序表有2N種組合。即，使用Landau 的記號0，可說用以得到覆查程序表的計算量是〇 ( 2n )的等 〇 級。—般，考慮成爲覆查程序表製作的對象之第一缺陷群 之缺陷的個數是約1 000個的等級時，對全部求覆查的程序 表並不切實際’必須講求某種高效率手段。作爲該高效率 手段’首先，表示可利用N次的計算，即以〇(n)之等級的 計算量決定對N個第一缺陷群之缺陷的覆查程序表的手 法。 所檢測出的缺陷各自具有特徵量。特徵量是以數値表 示缺陷之各種性質者。此特徵量例如是缺陷的反射率或面 -16- 201035541 積、形狀、缺陷的座標等。根據此特徵量，預先決定在第 一缺陷群中優先覆查何種缺陷。例如，使面積大的缺陷優 先、使反射率小的缺陷優先等。然後，求出如可使在第一 缺陷群中成爲覆査對象之缺陷的個數變成最大的路徑。 在不想對覆査之缺陷設定優先順位的情況，即在和特 徵量無關而想隨機地決定覆查之缺陷的情況，預先取亂數 作爲特徵量。因而，可認爲使根據亂數之特徵量大的缺陷 _ 優先。又，亦可使用2種以上的特徵量來決定優先順位。 〇 此外，以下爲了易於說明，表示僅以某一種特徵量爲對象 來製作優先覆查此特徵量之値大的缺陷之覆査程序表的例 子。 已檢測出之第一缺陷群中覆查尙未結束之第一缺陷群 的列表（以下稱爲覆査程序表列表）如第9圖所示，表示第 —缺陷群由缺陷801、802、803、804之4個所構成之情況 的例子。拍攝缺陷之優先順位亦按照此順序。 Q 在第7圖表示第6圖所示之程序表製作之處理流程之 前半部分的例子。 在此，從優先順位最高的缺陷進行處理。最初，當作 在覆查程序表列表中什麼都未包含（步驟S1)。首先，在第 一缺陷群中’將是優先順位最高之缺陷的缺陷801加入覆 查程序表列表。因而’覆查程序表列表由僅一個缺陷8 0 1 所構成。在第10圖表示此時的狀況。若覆查對象是僅一 個’該對象當然是可覆查。因爲預先將覆查光學系統104 -17- 201035541 移往缺陷的位置即可。此時，以箭號901表示覆査光 統104所採取的軌跡。 接著，將優先順位第2位的缺陷802加入覆查程 列表（步驟S2)。即，覆查程序表列表由2個缺陷801、 所構成。在第11圖表示此列表。在新的缺陷加入覆査 表列表的情況，作成該缺陷之檢査掃描軸向的座標成 近開始覆查時之覆査光學系統1 04之位置的順序。對 _ 入之缺陷802的前後，判定覆查光學系統104是否可 Ο (步驟S3)。覆查光學系統104是否可移動是根據覆查 系統104所安裝之軸的運動性能，即最高速度或加速 檢查掃描軸的速度及缺陷之間的距離決定。即，若相 從2個缺陷之檢查掃描方向的距離和檢查掃描軸的速 得之到達時間（設爲11 )，從2個缺陷之覆查光學系統 之軸向的距離和覆查光學系統1 04的運動性能所得之 時間（設爲12)的關係是如下之第（1)式，則覆查光學系轫 Ο 可移動。此外，d是餘裕時間。 ι1>ί2 + ά 第（1)式 在第11圖所示的例子，覆查光學系統104可從缺陧 往缺陷802移動。在加入覆査程序表列表之缺陷之不 後（在僅前或後存在的情況爲其中一個）的情況覆查光 統104亦可移動時，所加入的缺陷依然殘留於覆查程 列表（步驟S6)。此時，以箭號1〇〇1表示覆查光學系舒 所採取的軌跡。 學系 序表 802 程序 爲接 所加 移動 光學 度、 對於 度所 104 到達 :104 ^ 801 論前 學系 序表 :104 -18- 201035541 一樣地’將優先順位第3位的缺陷803加入覆查程序 表列表。即，覆査程序表列表由3個缺陷801、803、802 所構成。在第12圖表示此列表。 和上次一樣’對加入覆査程序表列表之缺陷803的前 後’判定覆查光學系統104是否可移動。在本例，因爲缺 陷801和缺陷8 03在檢查掃描方向的距離短，所以當作覆 查光學系統104無法從缺陷801往缺陷803移動。反之， 0 當作覆查光學系統104可從缺陷803往缺陷802移動。以 箭號1101表示覆查光學系統104所採取的軌跡。以虛線表 示無法移動的軌跡。 在覆查光學系統104無法對加入覆查程序表列表之缺 陷之前後其中一個或雙方，移動的情況，從覆查程序表列 表刪除所加入的缺陷（在此爲803 )(步驟S4)。即，覆查程序 表列表成爲回到第1 1圖之狀態。 藉由重複此處理（步驟S5)，可利用N次的循環處理得 〇 到對第一缺陷群中之N個缺陷的覆査程序表列表。在第13 圖表示處理的結果。此時，以箭號1201表示覆査光學系統 104所採取的軌跡。 可是，利用此方法所得之程序表受到特徵量之優先順 位的束縛。例如，如第1 4圖所示，藉由覆查優先順位爲第 5的缺陷1301，假設優先順位爲第6的缺陷1 3 02、優先順 位爲第7的缺陷1 3 0 3、優先順位爲第8的缺陷1 3 0 4、優先 順位爲第9的缺陷1 3 0 5、優先順位爲第1 〇的缺陷1 3 0 8之 -19- 201035541 5個缺陷全部無法覆查的狀況。在上述的手法，得到以覆 查優先順位爲第5之缺陷1301的軌跡1 307所示的程序表。 可是，因爲覆查缺陷的機會在檢查光學系統每次往返有複 數次，所以如軌跡1 308所示作成實施優先順位從第6至第 10 之 5 個缺陷 1302、1303、1304、1305、1308 的覆査，而 對優先順位第5的缺陷1 30 1，在下一以後之檢查掃描的機 會或在檢查掃描結束後再覆查，比較有整體上可覆査很多 缺陷的情況（在第一缺陷群中可覆查之缺陷的個數變成最 〇 大的情況）。 爲了覆査更多的缺陷，表示可藉由對暫時得到的覆查 程序表再計算程序表而作改善的例子。 首先，表示一例，即定量地評估所得之覆查程序表的 優劣，並爲了可比較幾個覆查程序表，算出覆查程序表之 評估値。作爲覆查程序表之評估値，例如有覆查對象之缺 陷之特徵量的總和。在假如優先覆查特徵量大之缺陷的情 Q 況，可說此評估値愈大之覆査程序表是愈優異的程序表。 又，若是覆查對象之缺陷的個數愈多愈優異的程序表，亦 可將該個數作爲評估値。當然，亦可組合兩者。若設特徵 量的總和爲P、覆査對象之缺陷的個數爲N，將覆查程序表 之評估値E定義爲如下第（2)式，藉由適當地決定係數k和 j，可組合兩者。 E = kxP + jxN 第（2)式 在優先覆查特徵量大之缺陷的情況，將k設爲正數， -20- 201035541 而在優先覆查特徵量小之缺陷的情況，將k設爲負數。j 設爲總是0或正數。依此方式，可說評估値E取愈大値的 覆查程序表是愈優異的程序表。 在第15圖表示使用定量評估覆查程序表的手法’高效 率求得評估ii之覆查程序表的具體例。以第一缺陷群之缺 陷的個數爲20個的情況，說明此流程。此外，此流程係在 第6圖所示之步驟S604、S606、S609中執行。 在步驟S1401，對這些20個缺陷，依前面所示的第7 〇 圖算出程序表，並將所得之程序表設爲S1。在第16圖表 示此狀況。在此，成爲覆查對象之缺陷全部有6個，各自 按照覆查的順序設爲1501~1506。將不是覆査對象之剩下的 14個缺陷總稱爲第一缺陷群1 507。 在步驟S1402，根據第（2)式算出程序表S1的評估値。 將所算出的結果設爲E1。 經由步驟S1402 — 1，在步驟S1403，如上述所示製作 Q 出最初之覆查對象之缺陷1501除外之情況（缺陷1501不存 在的情況）的覆查程序表。將所製作的程序表設爲S2。 在步驟S 1404，根據第（2)式算出程序表S2的評估値。 將所算出的結果設爲E2。 在步驟S 1 405，比較E1和E2。藉此，停止缺陷1501 之覆查的情況，可覆查第一缺陷群15 07的幾個，結果，調 査是否無法得到具有更高之評估値的覆查程序表。在第17 圖表示此時的狀況。在本例中，藉由停止缺陷1501之覆查， -21- 201035541 而可重新覆查缺陷丨6〇1、1602、1603。 在步驟S14 05的比較結果中，在程序表S1是比程序表 S2優異之程序表的情況’不應停止缺陷1501的覆查。在此 情況，經由步驟S 1 40 5 — 1、S 1 405 - 2，回到步驟S 1 403， 對下一缺陷1502 —樣地進行處理。 此外，在步驟S1405-1將N加1，在步驟S1405-2, 判定那時的N是否是成爲覆查程序表之對象的缺陷數，在 N不是成爲覆査程序表之對象之缺陷數的情況，回到步驟 〇 S 1 403 ° 在步驟S1405的比較結果中，在程序表S1是比程序表 S2優異之程序表的情況，停止缺陷1501的覆查較佳。在此 情況，經由步驟S 1 40 6、S 1 407，回到步驟S 1 402 — 1。 在步驟S1 406，從作爲覆查程序表之對象的第一缺陷 群刪除缺陷1 5 0 1。 在步驟S1407，將程序表S1置換成程序表S2，同時將 Q 評估値E1置換成評估値E2,再度從最初的缺陷進行處理。 藉由依序進行處理至最後的缺陷1506，可得到優異的 覆查程序表。設成爲程序表對象之缺陷的個數爲η、在最 初所得之覆查程序表可覆查之缺陷的個數爲m，在此所示 之手段的計算量是O(nxm)的等級。自定義可知，n2m。因 爲覆查光學系統104之軸移動速度是規定速率（determined rate) ’ 一般是n>>m。但，在η爲小値的情況，未限定如此。 即，雖然計算量最差的情況是〇(η2)的等級，但是因爲大部 -22- 201035541 分的情況接近〇(n)的等級，所以在此所示的演算法能以少 的計算量求得優異的程序表。 雖然覆査光學系統104可拍攝缺陷的機會和在基板 109上之缺陷的位置有關，但是一般有複數次。這是由於檢 查光學系統103和覆查光學系統104在X方向掃描基板1〇9 複數次的綠故。因爲以覆查光學系統104可拍攝在更早時 刻已進行之掃描所檢測出之缺陷的機會比在更晚時刻已進 行之掃描所檢測出之缺陷的機會更多，所以可和檢查之掃 〇 描同時覆查的可能性高。即，在最後的掃描所檢測出之缺 陷（最後行的缺陷）可和檢查之掃描同時覆查的可能性低。 又，覆查光學系統104很難在一次的掃描之間拍攝位 於同一行上之在X方向彼此接近的缺陷。這是由於覆查光 學系統104之Y方向移動機構的速度或加速度是規定速 率。即，在檢査掃描結束的階段覆查的完成程度主要和基 板109上之缺陷的分布狀態、X方向的檢查掃描速度及γ 〇 方向之覆查光學系統104的運動性能相依。在第一缺陷群 中，對無法和檢查之掃描同時覆査的缺陷，在檢査結束後 和以往一樣，逐一以覆査光學系統1 0 4拍攝即可。 如前面亦曾談到，在覆查光學系統104拍攝時，在X 方向成爲在移動中拍攝。若可暫時停止移動，雖然對覆查 光學系統1 04有利，但是困難的情況多。這是因爲，在檢 查光學系統 103使用線感測器或 TDI(Tirae Delay Integral ion)動作的圖像感測器的情況，將基板109和檢 -23- 201035541 查光學系統1 0 3的相對速度保持定値該感測器的性質較 佳。在不影響以檢查光學系統1 03所得之影像的方法中， 難在掃描的中途使在X方向的動作停止，或使速度變慢。 又，若使速度變慢’當然檢查所需的時間增長。 若在覆查光學系統104的攝像中以具有相對速度之狀 態進行拍攝（在移動中拍攝），則在所拍攝之影像中會發生 被拍攝物振動。因而，需要針對此被.拍攝物振動的對策， ^ 在此，被拍攝物振動的程度是根據X方向的相對速度、覆 查光學系統104的解析度及覆査光學系統104的曝光時間 來決定。X方向的相對速度愈大，又覆查光學系統104的解 析度愈小，又其曝光時間愈長，各自發生之被拍攝物振動 愈大。 反之，只要X方向的相對速度不是零，就無法消除被 拍攝物振動。被拍攝物振動的對策意指無法察覺被拍攝物 振動的影響，或者減輕至可看成在實用上無影響的程度。 〇 各種實驗的結果，得知若設相對速度爲V[m/s]、覆查 光學系統104的解析度爲R[m/pix]、覆查光學系統104之 相機的曝光時間（或在相機之曝光中照明所照射的時間，以 下相同）爲T[s]，無法察覺被拍攝物振動之覆查光學系統 104和基板109之相對速度的上限以如下第（3)式表示。 V = (0.5xR)/T 第（3)式E=k P+j . N Further, k is a weighting coefficient of P, and j is a weighting coefficient of N. The evaluation 値E of the program table is calculated, and the program table Q of the evaluation 値E is calculated as a review program table. In the imaging inspection method of the present invention, the first defect group detecting step is characterized in that the inspection object and the inspection optical system are relatively moved in the X direction, and the inspection object is imaged by the inspection optical system, and the defect is detected. Detecting as a first defect group; and performing a step of detecting the first defect group detecting step by the inspection optical system while moving the object to be inspected and the inspection optical system relative to each other in the X direction Defect, and detecting the defect as a second defect group; and the defect detecting step of detecting the second defect 201035541 group and the first defect group not photographed in the review step as a defect; the inspection optical system and the The review optical system is relatively independently moved in the X direction, and relatively moves in the Y direction orthogonal to the X direction, and simultaneously performs the first defect group detecting step and the review step. According to this configuration, when the inspection object and the inspection optical system are relatively moved in the X direction, the inspection optical system and the inspection optical system are relatively moved in the γ direction by the inspection optical system moving mechanism. In the scanning scan of the optical system, the inspection of the optical system is performed on the periphery of the defect, so that the inspection operation time after the inspection can be shortened compared with the prior art. [Embodiment] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following description, the same components are denoted by the same reference numerals. (First Embodiment) Figs. 1 to 17 show a first embodiment of the present invention. Q On the table of the image-inspecting apparatus 200 on which the substrate 109 of the object to be inspected is placed, X-axis rails 107 are provided on both sides. The X-axis movable member 108 is disposed to straddle the X-axis rails 107 on both sides. Specifically, the X-axis rail 107 is constituted by a feed screw, and a nut screwed with the feed screw is provided to the X-axis movable member 108. The feed screw as the X-axis rail 107 is rotationally driven by the motor in response to the amount of movement of the X-axis movable member 108. On the X-axis movable member 108, a first Y-axis rail 106 and a second Y-axis rail 105 are disposed in a direction 201035541 crossing the X-axis rail 107. The first Y-axis movable member 101 is disposed on the first γ-axis rail 106. On the first x-axis movable member 101, the inspection optical system 103 is mounted to photograph the substrate 109. The second yoke movable member 102 is disposed on the second yaw axis rail 105. On the second x-axis movable member 102, the inspection optical system 1Ό4 is mounted as the imaging substrate 109. Specifically, the first and second yoke rails 106, 105 are constituted by feed screws, and the nut screwed with the feed screw is disposed on the first yoke movable member 101 and the second yoke movable member 102. The feed screw as the first boring rail 1 0 6 is rotationally driven by the motor in response to the amount of 〇 movement of the first cymbal movable member 101. The feed screw as the second yoke rail 105 is rotationally driven by the motor in response to the amount of movement of the second yaw movable member 102. In the present invention, the inspection optical system 104 detects a more important one of the first defect groups detected as defects by the inspection optical system 103 as the second defect group. Then, as a result of the inspection by the computing device 301, the second defect group and the undetected first defect group are regarded as ◎ defects in the inspection. The control unit 300 controls the inspection optical system moving mechanism that relatively moves the substrate 109 and the inspection optical system 103 placed on the placing portion in the X direction, and relatively moves the inspection optical system 103 and the inspection optical system 104 in the x direction. Review the movement of the optical system moving mechanism. By the control of the control mechanism 300, the substrate 109 and the inspection optical system 104 are relatively moved in the X direction, and the position control of the inspection optical system 104 is performed in the imaging scan of the inspection optical system 103, and imaging is performed. This control mechanism -10- 201035541 300 is composed of a computing device 301 having a microcomputer as a main part. The composition of the computing device will be specifically described in terms of actions. First, the operation of the imaging inspection apparatus 200 in the X direction will be described. As shown in FIGS. 2(a) and 2(b), the inspection operation is performed by fixing the substrate 109 to the table of the imaging inspection apparatus 200, and moving the X-axis movable member 1〇8 along the 107 to cause the inspection optical system 103. And the inspection optical system 1〇4 moves in the X direction. 0. Here, the imaging inspection apparatus 200 will be described with reference to the configuration shown in Fig. 1. However, the inspection optical system 103 and the cover are fixed as shown in Fig. 3 and Figs. 4(a) and 4(b). The same applies to the case where the optical imaging system 104 is driven and the substrate 109 is driven to constitute the imaging inspection device 200. In the imaging inspection apparatus 200 of Fig. 3, the X-axis rail 107 is provided at intervals in the vicinity of the center of the table in the Y-axis direction, and the substrate 109 is moved in the X-axis direction along the X-axis rail 107. The X-axis movable member 108 of the loading inspection optical system 103 and the inspection optical system 104 is fixed near the center Q of the table in the X-axis direction. On the other hand, when the substrate 109 passes through the inner side of the X-axis movable member 108, photographing is performed. Further, in the case where the imaging inspection apparatus 200 moves the substrate 109 in the X-axis direction as shown in the fifth (a) and (b), the X-axis movable member 108 is moved in the X-axis direction. The inspection optical system 103 and the inspection optical system 104 are driven in the X-axis direction. As shown in FIGS. 1 to 5, the substrate 109 and the inspection optical system 103 are relatively moved, and the substrate 109 is scanned in the X-axis direction. -11 - 201035541 Next, the operation of the imaging inspection device 200 in the Y direction will be described. Since the field of view of the inspection optical system 1〇3 is generally smaller than the width of the substrate 1〇9, it is impossible to scan the entire surface of the substrate 1〇9 in one operation. Therefore, in the imaging inspection apparatus 200 of the present invention, the X-axis mover 1〇8 is moved in the X-axis direction while performing the pitch feed operation in the γ direction, and scanning is performed a plurality of times to scan the substrate 丨09. The whole face. For this reason, the imaging inspection apparatus 200 of the present invention has a mechanism required for the inspection optical system 1〇3 and the inspection optical system 1〇4 to move in the γ direction.覆 The inspection optical system 1〇4 of the imaging inspection device 20 0 will be described. The review optical system 104 is disposed in the X direction and inspects the optical system 103 as a body. The review optical system 1 to 4, as shown in Figs. 1 to 5, is provided separately from the inspection optical system 103 to have a mechanism movable in the γ direction. The movable range of the review optical system 104 is larger than the width of the substrate 109 in the Y direction. Next, the operation of the review will be described.摄像 The imaging of the inspection optical system 104 can be performed simultaneously with the inspection operation in the inspection optical system 103. When the scanning of the inspection in the inspection optical system 103 is in progress, as described above, in the X direction, the scanning inspection range is moved in a state where the substrate 109 and the inspection optical system 103 have relative speeds (hereinafter, the relative speed in the X direction). . Here, since the inspection optical system 104 is in the X direction with the inspection optical system 103, the inspection optical system 104 also moves in the inspection range in the same state as the relative speed of the substrate 109. That is, in the X direction, the field of view of the optical system 104 is examined through the entire width of the substrate 109 of -12-201035541. Further, in the case where the inspection optical system 103 and the inspection optical system 1〇4 are disposed at positions shifted in the X direction, it is necessary to increase the range of movement of the inspection optical system 1〇4 in the X direction by only the offset amount. Further, in the Y direction, the inspection optical system 104 itself has a moving mechanism, and since the movable range is larger than the substrate 109, the inspection optical system 104 can be moved to any position of the substrate 1〇9. In this manner, the review optical system 104 can capture any position of the substrate 109 into the field of view and take a picture. In the present embodiment, the inspection optical system 104 can perform imaging while being stationary in the Y direction, but cannot pass relatively statically between the substrate and the substrate 109 in the X direction. Therefore, imaging is performed by measuring the timing. An image of a defect in the first defect group. Fig. 6 is a flow chart showing an example of the operation at this time. First, in step S601, inspection scanning of the first row of the substrate 109 is performed using the inspection optical system 103. Depending on the configuration of the device, the position of the defect can also be grasped at the instant when the inspection optical system 103 passes over the defect of the substrate 109. However, here, in order to simplify the description, it is considered that the position of the defect located on the scanning line can be grasped at the time point when the inspection of the first line is completed. In step S601, the position of the defect is grasped and detected as the first defect group. In step S602, and inspecting the scanning of the second row of the substrate 109 by the optical system 103, the position of the defect of the first defect group detected in step S601 is used to calculate the order in which the inspection optical is to be used. System 104 performs a review of the first row. -13- 201035541 In addition, it is not necessary to review the defects of all the first defect groups, and it is not necessary to review the defects of all the first defect groups. Therefore, in the defect of the first defect group to be inspected, the defect of the first defect group which can be checked is set with priority, and the inspection program table for moving the inspection optical system 104 is created. Specific examples of the production of the review program table will be described later. When the preparation of the review schedule in step S602 is completed, in step S603, the inspection optical system 103 performs the inspection scan of the second row of the substrate 109 (step S603-a). In parallel with this inspection scan (step S603-a), the first defect group detected in the first row is overwritten in accordance with the program table calculated and created in step S602 by C1 (step S603-b). After step S604, the repeating operation is performed while shifting one line one by one. That is, in step S604, the defect of the first defect group that needs to be checked in the first defect group of the second row detected in step S603, and the detection in step S601 but not in step S603-a are not covered. The set of the first defect group merged with the defects of the first defect group in the first row is checked, and the review procedure table is calculated. Here, although it is detected in step S601, the defect of the first defect group in the first row which is not checked in step S603-a is the first defect which is not checked due to the relationship of the program table despite the need for review. group. At the same time, in step S605', the inspection scan of the third row is performed in parallel (step S605-a), and the review of the first row and the second row of the schedule created in step S604 (step S605-b). Steps S606 and S607 are also the same. That is, in step S606, the reviewer is required in the first defect group of the 3-14-201035541 line detected in step S605, and is detected in step S601, step S603 a, step S 605 - a In spite of the need to review, the first defect group of the first to third rows that have not been checked due to the relationship of the program table is a set of the first defect group. At the same time, the inspection scan of the fourth row (step S607-a) and the review of the third to third rows of the schedule created in step S606 are performed in parallel (step S607-b). Step S608 shows the operation of the last line (Nth line) in the checking operation. In the present embodiment, since the position of the defect of the first defect group in the last row cannot be determined until the end of the inspection scan, the defect of the first defect group detected by the scanning line cannot be checked at the same time as the inspection. Therefore, in the first defective group to be checked after the end of the scan inspection at the end of the line, there is a case where the unfinished defect is overwritten, and in step S609, a general review which is not performed in parallel with the inspection scan is performed. Q shows the trajectory of the inspection optical system 1〇3 in Fig. 8(a). Reference numeral 701 denotes a locus of the first line of the scanning of the inspection optical system 103, and 706 denotes an area inspected by scanning the first line, and 705 indicated by a black circle indicates the detected first defect group. Similarly, 702 to 704 are tracks of the second to fourth lines of the inspection optical system 103, and 707 to 7 09 are areas examined at the second to fourth lines of scanning. The trajectory of the operation of reviewing the optical system 1〇4 is shown in Fig. 8(b). 710 is the trajectory of the second line of the scanning of the optical system 104, 711 is the trajectory of the third line of the scanning of the optical system 104, and 712 is the trajectory of the fourth line of the scanning of the optical system 104. Further, in the eighth (a) and (b) drawings, the inspection is performed four times in a round trip. As in the example shown in Fig. 6, at the time point when each scanning is completed, the position of the defect of the first defect group which is scanned by the dry grip is regarded as a defect. In this case, the first defect group that can be checked at the same time as the scan of the Nth line is the one detected before the Nth line (the (N-1)th line). The timing at which the position of the defect of the first defect group can be grasped by the actual 0 upper is different depending on the configuration. Explain the specific examples of the production review schedule. From the purpose of performing the review simultaneously with the scan of the inspection, it is understood that the order of the review is uniquely determined by the coordinates of the first defect group. Regarding each of the first defect groups, since there are two types of review or no review, there are 2N combinations of the review schedules when reviewing the N first defect groups. That is, using Landau's symbol 0, it can be said that the calculation amount for obtaining the review program table is equal to 〇 ( 2n ). In general, when it is considered that the number of defects of the first defect group to be the object of the review program table is about 1,000, it is not practical for all the program tables to be reviewed. means. As the high-efficiency means, first, the calculation of the N-times calculation, that is, the method of reviewing the defects of the N first defect groups by the calculation amount of the level of 〇(n) is indicated. The detected defects each have a feature amount. The feature quantity is the number of defects that indicate the various properties of the defect. This characteristic amount is, for example, the reflectance of the defect or the surface, the shape, the coordinates of the defect, and the like. Based on this feature amount, it is determined in advance which defects are preferentially checked in the first defect group. For example, a defect having a large area is prioritized, a defect having a small reflectance is prioritized, and the like. Then, a path is obtained in which the number of defects that can be the object of review in the first defect group is maximized. In the case where it is not desired to set a priority order for the defect to be inspected, that is, in the case where it is desired to randomly determine the defect of the review regardless of the feature amount, the random number is previously taken as the feature amount. Therefore, it is considered that the defect _ which is large in the feature amount according to the random number is prioritized. Further, two or more feature quantities may be used to determine the priority order. In addition, in the following, for the sake of convenience of explanation, an example of a review program table in which a large defect having a large feature value is preferentially checked is created for only one type of feature amount. The list of the first defect group that has not been detected in the first defect group that has been detected (hereinafter referred to as the list of the review program table) is as shown in FIG. 9, indicating that the first defect group is defective 801, 802, 803. An example of a situation in which four of 804 are formed. The priority of shooting defects is also in this order. Q Fig. 7 shows an example of the first half of the processing flow of the program creation shown in Fig. 6. Here, the processing is performed from the defect with the highest priority. Initially, nothing is included in the list of review schedules (step S1). First, the defect 801, which is the defect with the highest priority, is added to the list of review procedures in the first defect group. Thus the 'review list list consists of only one defect 80 1 . Fig. 10 shows the situation at this time. If the object to be reviewed is only one, the object is of course subject to review. This is because the inspection optical system 104 -17- 201035541 is moved to the position of the defect in advance. At this time, the trajectory taken by the inspection system 104 is indicated by an arrow 901. Next, the defect 802 of the second priority is added to the review schedule list (step S2). That is, the list of review programs is composed of two defects 801. This list is shown in Figure 11. In the case where a new defect is added to the list of the checklist, the coordinates of the inspection scan axis in which the defect is made become the order of the position of the inspection optical system 104 at the start of the review. Before and after the defect 802 is entered, it is determined whether or not the inspection optical system 104 is falsified (step S3). Reviewing whether the optical system 104 is movable is determined based on the motion performance of the axis to which the review system 104 is mounted, i.e., the maximum speed or acceleration check the speed of the scan axis and the distance between the defects. That is, if the distance from the inspection scan direction of the two defects and the arrival time of the inspection scan axis are set to 11 (the angle), the distance from the axial direction of the two optical systems and the inspection optical system 1 are examined. The relationship (time set to 12) obtained by the exercise performance of 04 is the following equation (1), and the optical system is reviewed to be movable. In addition, d is the marginal time. Ip1> ί2 + ά Equation (1) In the example shown in Fig. 11, the review optical system 104 can move from the defect to the defect 802. In the case where the defect of the list of review procedures is added (in the case where only one of the front or the back is present), when the inspection system 104 can also move, the added defects remain in the review list (steps). S6). At this time, the trajectory taken by the optical system is checked by the arrow 1〇〇1. The program sequence 802 program is connected to the mobile optical degree, for the degree of the arrival of 104: 104 ^ 801 Pre-schools list: 104 -18- 201035541 The same as the third priority defect 803 added to the review List of program tables. That is, the list of review schedules is composed of three defects 801, 803, and 802. This list is shown in Figure 12. As before, it is judged whether or not the optical system 104 is movable to the front and back of the defect 803 added to the list of the check list. In this example, since the distance between the defect 801 and the defect 803 is short in the inspection scanning direction, the inspection optical system 104 cannot move from the defect 801 to the defect 803. Conversely, 0 as the review optical system 104 can move from the defect 803 to the defect 802. The trajectory taken by the inspection optical system 104 is indicated by an arrow 1101. A broken line indicates a track that cannot be moved. In the case where the review optical system 104 cannot move one or both of the defects before joining the list of the review program table, the added defect (here, 803) is deleted from the review program table list (step S4). That is, the list of the review schedules is returned to the state of Fig. 11. By repeating this processing (step S5), the N-cycle processing can be used to obtain a list of review procedures for the N defects in the first defect group. The result of the processing is shown in Fig. 13. At this time, the trajectory taken by the inspection optical system 104 is indicated by an arrow 1201. However, the program table obtained by this method is bound by the priority order of the feature quantity. For example, as shown in FIG. 14, by reviewing the defect 1301 whose priority order is the fifth, it is assumed that the priority order is the sixth defect 1 3 02, the priority order is the seventh defect 1 3 0 3 , and the priority order is The defect of the eighth is 1 3 0 4 , the defect with the priority order is the ninth defect 1 3 0 5 , the priority order is the defect of the first defect 1 3 0 8 -19- 201035541 All the five defects cannot be checked. In the above method, a program table shown by the track 1 307 in which the priority order of the fifth defect 1301 is checked is obtained. However, since the opportunity to review the defect is repeated multiple times per round trip of the inspection optical system, as shown in trace 1 308, five defects 1302, 1303, 1304, 1305, 1308 are implemented which are preferentially ranked from the sixth to the tenth. Review, and the defect of the fifth priority 1 30 1, in the next inspection scan opportunity or after the inspection scan is over, compared with the overall review of many defects (in the first defect group) The number of defects that can be checked can become the most significant case). In order to review more defects, an example of improvement can be made by recalculating the program table for the temporarily obtained review schedule. First, an example is given, that is, the merits and demerits of the obtained review schedules are quantitatively evaluated, and the evaluation of the review schedules is calculated in order to compare several review schedules. As an evaluation of the review schedule, for example, there is a sum of the feature quantities of the defects of the review object. In the case of a priority review of a defect with a large feature quantity, it can be said that the larger the review procedure table is, the better the schedule is. In addition, if the number of defects of the object to be inspected is more and more excellent, the number can be used as an evaluation. Of course, you can combine them. If the sum of the feature quantities is P and the number of defects of the review object is N, the evaluation 値E of the review procedure table is defined as the following formula (2), and can be combined by appropriately determining the coefficients k and j. Both. E = kxP + jxN In the case of the case where the feature with large feature quantity is preferentially checked, k is set to a positive number, -20-201035541, and k is set to a negative number in the case of preferentially checking for a defect with a small feature amount. . j is set to always 0 or a positive number. In this way, it can be said that the review procedure table in which the evaluation 値E is larger is the more excellent the schedule. Fig. 15 shows a specific example of the review procedure table for evaluating ii using the method of quantitative evaluation of the schedule. This flow will be described in the case where the number of defects of the first defect group is 20. Further, this flow is executed in steps S604, S606, and S609 shown in Fig. 6. In step S1401, the program table is calculated based on the seventh map shown above for these 20 defects, and the obtained program table is set to S1. This is shown in the 16th chart. Here, there are six defects that are to be examined, and they are set to 1501 to 1506 in the order of review. The remaining 14 defects that are not the subject of the review are collectively referred to as the first defect group 1 507. In step S1402, the evaluation 程序 of the program table S1 is calculated based on the formula (2). The calculated result is set to E1. In step S1402, in step S1403, a review procedure table in which the defect 1501 of the first inspection target is excluded (when the defect 1501 does not exist) is created as described above. Set the created program table to S2. In step S1404, the evaluation 程序 of the program table S2 is calculated based on the formula (2). The calculated result is set to E2. At step S 1 405, E1 and E2 are compared. Thereby, the case where the defect 1501 is checked can be stopped, and several of the first defect group 15 07 can be checked. As a result, it is checked whether or not the review program table having a higher evaluation threshold cannot be obtained. The situation at this time is shown in Fig. 17. In this example, the defect 丨6〇1, 1602, 1603 can be re-examined by stopping the review of the defect 1501, -21-201035541. In the comparison result of step S14 05, in the case where the program table S1 is a program table superior to the program table S2, the review of the defect 1501 should not be stopped. In this case, via the steps S 1 40 5 - 1 and S 1 405 - 2, the process returns to the step S 1 403 to process the next defect 1502. Further, in step S1405-1, N is incremented by one, and in step S1405-2, it is determined whether or not N at that time is the number of defects to be the object of the review program table, and N is not the number of defects of the object of the review program table. In other cases, the process returns to step 1S 1 403 °. In the comparison result of step S1405, in the case where the program table S1 is a program table superior to the program table S2, it is preferable to stop the inspection of the defect 1501. In this case, via steps S 1 40 6 , S 1 407, the process returns to step S 1 402-1. At step S1 406, the defect 1 5 0 1 is deleted from the first defect group which is the object of the review program table. In step S1407, the program table S1 is replaced with the program table S2, and the Q evaluation 値E1 is replaced with the evaluation 値E2, and the processing is again performed from the first defect. By sequentially processing to the last defect 1506, an excellent review schedule can be obtained. The number of defects to be the object of the program table is η, and the number of defects that can be checked in the initial inspection program table is m, and the calculation amount of the means shown here is O(nxm). Customization knows, n2m. Since the axis moving speed of the optical system 104 is checked, the determined rate is generally n>>m. However, in the case where η is small, this is not limited. That is, although the case where the calculation amount is the worst is the level of 〇(η2), since the case of most -22-201035541 points is close to the level of 〇(n), the algorithm shown here can be calculated with a small amount. Get an excellent schedule. While the opportunity to review the optical system 104 for photographing defects is related to the location of defects on the substrate 109, it is generally a plurality of times. This is because the inspection optical system 103 and the inspection optical system 104 scan the substrate 1〇9 several times in the X direction. Since the review optical system 104 can take more chances of detecting a defect detected by scanning performed at an earlier time than a defect detected by scanning at a later time, the broom can be checked. The possibility of reviewing at the same time is high. That is, the defect detected in the last scan (the defect of the last line) is less likely to be checked at the same time as the scan of the inspection. Further, it is difficult for the review optical system 104 to capture defects in the X direction which are close to each other on the same line between scans. This is because the speed or acceleration of the Y-direction moving mechanism of the inspection optical system 104 is a prescribed rate. That is, the degree of completion of the review at the end of the inspection scan is mainly dependent on the distribution state of the defect on the substrate 109, the inspection scanning speed in the X direction, and the motion performance of the inspection optical system 104 in the γ 方向 direction. In the first defect group, the defects that cannot be checked at the same time as the inspection scan can be photographed one by one by reviewing the optical system 104 after the inspection is completed. As also mentioned above, when the review optical system 104 is photographed, it is photographed while moving in the X direction. If the movement can be temporarily stopped, although it is advantageous for the review optical system 104, it is often difficult. This is because, in the case where the optical system 103 is inspected using a line sensor or a TDI (Tirae Delay Integration) image sensor, the substrate 109 and the inspection -23-201035541 check the relative speed of the optical system 103. It is better to keep the characteristics of the sensor. In the method of not inspecting the image obtained by the inspection optical system 103, it is difficult to stop the operation in the X direction or slow down the speed in the middle of scanning. Also, if the speed is made slower, of course, the time required for the inspection increases. If the image is taken at a relative speed during the imaging of the review optical system 104 (photographed while moving), the subject vibrates in the captured image. Therefore, countermeasures against the vibration of the subject are required. ^ Here, the degree of vibration of the subject is determined by the relative speed in the X direction, the resolution of the inspection optical system 104, and the exposure time of the inspection optical system 104. . The greater the relative velocity in the X direction, the smaller the resolution of the optical system 104 is examined, and the longer the exposure time, the greater the vibration of the subject. Conversely, as long as the relative speed in the X direction is not zero, the subject vibration cannot be eliminated. The countermeasure against the vibration of the subject means that the influence of the vibration of the subject is not detected, or is reduced to the extent that it can be regarded as having no effect in practical use. As a result of various experiments, it is found that if the relative speed is V [m/s], the resolution of the inspection optical system 104 is R [m/pix], and the exposure time of the camera of the optical system 104 is checked (or at the camera) The time during which the illumination is illuminated during exposure is the same as T[s], and the upper limit of the relative speed of the inspection optical system 104 and the substrate 109, which is incapable of detecting the vibration of the subject, is expressed by the following formula (3). V = (0.5xR)/T (3)
例如，若覆查光學系統 104的解析度 R爲 0.5x10" 6[m/pix]( = 0.5[/z m/pix])，其曝光時間 T -24- 201035541 爲1 X 10 - 3[s](=l[ms])，若使相對速度大於相對速度 V = 〇.〇〇〇25[m/s]，即V = 0.25[mm/s]，則發生被拍攝物振動。 可是，因爲檢查光學系統103 —般以100[m/s]以上的 相對速度動作，所以會發生被拍攝物振動，因而需要某種 對策。作爲此對策，自該第（3)式可知，想到縮短覆查光學 系統104之相機的曝光時間。若縮短曝光時間，可一面滿 足檢查動作所需的相對速度V，一面拍攝無法察覺被拍攝 物振動之影像。例如，在作爲相對速度V想得到100 [m/s] 的情況’將曝光時間T設爲自該第（3)式所算出之2.5[//s] 以下即可。具體而言，藉由覆査光學系統104的相機使用 電子快門等，而可實現短的曝光時間。 可是，只是單純縮短曝光時間，所得之影像將會變暗。 關於此問題，可知：藉由使用高靈敏度的相機或使用高亮 度的照明可解決，但是目前可得到高靈敏度的相機昂貴， 又就算使用那種相機，亦難將曝光時間設爲1 0 [ μ S ]以下。 〇 又，高亮度照明仍相同，亦擔心提高照明度對被檢查物的 不良影響。 因此’在本實施形態，不僅只是縮短曝光時間，而且 使用由閃光燈所代表的閃光照明。若使用此閃光照明，因 爲可縮短照明的時間，所以在連續光照明可得到和縮短曝 光時間的情況一樣之效果。在閃光燈存在閃光時間約1 [ y s]之高速者，每單位時間的光量亦大。因而，藉由組合這 種閃光燈和一般靈敏度的相機，即使在相對速度 V=約 -25- 201035541 5 00[mm/s]亦可拍攝無法察覺被拍攝物振動之影像。又’爲 了一樣之目的，亦可在照明使用雷射。在此情況，可實現 更短的閃光時間，又每單位時間的光量亦大。 此外，亦可採用一種構成’即藉由使用可修正被檢查 物之缺陷的雷射光或裝置而可和檢查同時修正缺陷。 (第2實施形態） 第18圖和第19(a)、（b)圖表示本發明之第2實施形態。 0 在上述的第1實施形態，如第1圖〜第5圖所示，檢查 光學系統103和覆查光學系統104構成爲各自獨立地在Y 方向移動，覆査光學系統104構成爲移動基板109的整個 寬度，但是在本第2實施形態，在構成爲使覆查光學系統 104相對檢查光學系統103而在Y方向相對移動這一點上 相異。 即，在裝載檢查光學系統103的第一Y軸可動件101 上，將覆查光學系統104安裝成在第一 Y軸可動件101之 〇 範圍內在Y方向移動。此外，覆查光學系統104之Y方向 的移動機構比檢查光學系統103 —次可掃描之Y方向的範 圍更大。 若依據此構成，因爲即使覆查光學系統104所具備之Y 方向之移動機構的可移動範圍窄，亦能以覆査光學系統拍 攝基板109的整個寬度，所以即使在覆查光學系統1〇4所 具備之Y方向之移動機構的移動範圍小的情況，覆查光學 系統104可拍攝第一缺陷群之缺陷的機會亦不會減少。 -26- 201035541 (第3實施形態） 第20(a)、（b)圖表示本發明之第3實施形態。 雖然在第2實施形態，覆查光學系統104之γ方向的 移動機構比檢査光學系統103 —次可掃描之Y方向的範圍 更大，但是在本第3實施形態，覆查光學系統丨04最低限 度應具備之Y方向的移動機構和檢査光學系統103 一次可 掃描之Y方向的範圍相等。 0 若依據此構成，覆查光學系統104可拍攝第一缺陷群 之缺陷的機會是1次。 (第4實施形態） 第21圖和第22(a)、（b)圖、第23圖和第24(a)、（b)圖、 第25圖、第26圖分別表示本發明之第4實施形態。 雖然在上述的各實施形態，檢查光學系統103和覆查 光學系統104各自是單數，但是在本第4實施形態，在具 有複數個檢査光學系統103和複數個覆查光學系統104這 〇 一點上相異。 在第21圖和第22(a)、（b)圖所示之第4實施形態的第 1.實施例，4 個 Y 軸可動件 1〇1-1、ιοί— 2、101— 3、101 —4設置於X軸可動件1〇8之上，第一檢查光學系統103 -1設置於Y軸可動件101— 1。第二檢查光學系統103— 2設 置於Y軸可動件101- 1。第三檢查光學系統103-3設置於 Y軸可動件101— 1。第四檢查光學系統103-4設置於Y軸 可動件101-4。又，2個Y軸可動件102—1、102—2設置 -27- 201035541 於X軸可動件108之上。第一覆查光學系統104 - 1設置於 Y軸可動件102— 1。第二覆查光學系統104-2設置於Y軸 可動件1 0 2 - 2。 雖然在第4實施形態的第1實施例Y軸可動件1 〇2 _ 1、102— 2構成爲在共同之覆查光學系統Y軸105之上移 動，但是在第23圖和第24(a)、（b)圖所示之第4實施形態 的第2實施例，覆查光學系統Y軸1〇5-1、105 — 2隔著間 ^ 隔並列地設置。在Y軸可動件102- 1構成爲在覆查光學系 Ο 統Y軸105 — 1之上移動、和Y軸可動件102-2構成爲在 覆查光學系統Y軸105— 2之上移動這一點上和第1實施例 相異。在此情況，藉由Y軸可動件102— 1、102— 2移動， 第一覆査光學系統104-1和第二覆查光學系統104-2可 在基板109的整個寬度依序進行拍攝。 在第25(a)、（b)圖所示之第3實施例中，構成爲覆查光 學系統Y軸105-1、105— 2在Y軸方向錯開地設置，並以 〇 第一覆查光學系統104— 1和第二覆查光學系統104— 2含 蓋基板109的整個寬度。 如本第4實施形態的第1 ~第3實施例所示，在實際的 檢查裝置中，有具備複數個檢査光學系統103和複數個覆 査光學系統104的情況，構成爲檢查光學系統103和覆査 光學系統104各自獨立地在Y方向可動，並具備整體上可 移動基板109之整個寬度的機構即可。又，亦可檢查光學 系統103和覆查光學系統104的個數相異。 -28- 201035541 又，如第26圖（a)、（b)圖所示之第4實施形態的第4 實施例所示，亦可將第二覆査光學系統104- 2和第三覆查 光學系統104-3之複數個覆查光學系統1〇4設置於覆查光 學系統Y軸105— 2。在複數個覆查光學系統1〇4配置於同 一軸上的情況，需要顧慮覆査光學系統104彼此不會碰撞。 爲此，亦可複數個覆查光學系統104預先決定各自對基板 109的覆查範圍，亦可以電氣式或機械式方法控制成不會碰 ^ 撞。當然，在檢查光學系統103爲1組的情況，亦可藉由 〇 如此地具備複數個覆查光學系統104,而提高可和檢查之掃 描同時覆査的機率。 (第5實施形態） 第27圖表示本發明之第5實施形態。 雖然在第20圖和第21(a)、（b)圖所示之第4實施形態 的第1實施例，在覆査光學系統Y軸105具備有複數個覆 查光學系統104,但是在本第5實施形態的第5實施例，僅 〇 在構成爲將單數的覆查光學系統104設置於覆査光學系統 Y軸105，而此覆查光學系統104可在基板109的整個寬度 移動這一點上相異。 (第6實施形態） 第28(a)、（b)圖、第29(a)、（b)圖各自表示本發明之第 6實施形態。 在第28(a)、（b)圖所示第6實施形態的第6實施例，在 沿著設置於X軸可動件108之上面的第一 Y軸軌道106移 -29- 201035541 動之Y軸可動件101’第一檢查光學系統103-1和第二檢 查光學系統103 - 2隔著間隔設置。第一、第二覆查光學系 統Υ軸1 05 — 1、1 05 — 2隔著間隔並列地設置於γ軸可動件 101。第一覆查光學系統104— 1設置於第一覆查光學系統Υ 軸105 -卜第二覆查光學系統ι〇4_2設置於第二覆查光學 系統Υ軸1 0 5 — 2。 在第29(a)、（b)圖所示第6實施形態的第7實施例’在 構成爲第一、第二覆查光學系統104—1、104— 2沿著共同 (1 的第一覆查光學系統Y軸105 - 1移動這一點上和第6實施 形態的第6實施例相異。 如本第6實施形態的第6實施例、第7實施例所示， 在作爲複數個覆查光學系統，裝載第一、第二覆查光學系 統104 - 1、104 — 2的情況，亦和在第18圖及第19(a)、（b) 圖所示之第2實施形態的說明一樣，可使用移動範圍窄的 移動機構。 〇 (第7實施形態） 第30圖表示本發明之第7實施形態。 如在第1圖〜第1 7圖所示之第1實施形態的說明所示， 爲了覆查光學系統104在不使基板109靜止下拍攝，需要 針對被拍攝物振動的對策。雖然在第1實施形態所說明的 方法亦可作爲對策，但是在本第7實施形態，在爲了減輕 被拍攝物振動的影響而另外在覆查光學系統104設置往X 方向的移動機構這一點上和第1實施形態相異。 -30- 201035541 此在覆查光學系統104往X方向的移動機構 光學系統104拍攝的期間，朝向抵消基板109和 系統104之相對速度的方向驅動此移動機構。 在第30圖，111表示基板109和覆査光學系 相對速度及方向。因爲相對速度及方向111是已 在第7實施形態，爲了減輕被拍攝物振動的影響 光學系統104以是速度和相對速度及方向111相 相反之11 0所示的抵消速度驅動即可。 () 具體而言，此移動機構如第5圖所示能以驅動 和光學系統之雙方的機構實現。又，因爲此移動 覆查光學系統1 04的視野中心可往X方向移動即 要利用往X軸方向的平行移動機構，亦可利用例 查光學系統104傾斜等之手段使視野中心可動的卷 工業上的可應用性 本發明可有助於提高在液晶顯示裝置、電 〇 置、太陽能電池等所使用之大型基板的生產力。 【圖式簡單說明】 第1圖係本發明之第1實施形態的檢査光學 動之攝像檢査裝置的立體圖。 第2(a)、（b)圖係該第1實施形態之第1圖的 裝置之正面的模式圖和側面的模式圖。 第3圖係該第1實施形態中之被檢查物作移 檢查裝置的立體圖。 是在覆查 覆查光學 統104的 知，所以 ，使覆查 同而方向 丨基板109 機構只要 可，未必 如以使覆 瓷構實現。 發顯示裝 系統作移 攝像檢查 動之攝像 -31- 201035541 第4(a)、（b)圖係該第1實施形態之第3圖的攝像檢査 裝置之正面的模式圖和側面的模式圖。 第5(a)、（b)圖係該第1實施形態之攝像檢查裝置之正 面的模式圖和側面的模式圖。 第6圖係在該第1實施形態同時進行檢查和覆查時的 流程圖。 第7圖係說明程序表製作之處理流程之前半部分的流 程圖。 〇 第8(a)、（b)圖係表示該第1實施形態之檢査光學系統 和覆查光學系統的軌跡圖。 第9圖係表示覆查未結束之第一缺陷群之列表的圖。 第10圖係表示在程序表製作中途之第一缺陷群列表 (缺陷數1個）的狀態圖。 第11圖係表示在程序表製作中途之第一缺陷群列表 (缺陷數2個）的狀態圖。 Q 第12圖係表示在程序表製作中途之第一缺陷群列表 (缺陷數3個）的狀態圖。 第13圖係表示程序表製作已結束之第一缺陷群列表 的狀態圖。 第14圖係因覆查某第一缺陷群而無法覆査很多第一 缺陷群之例子的說明圖。 第1 5圖係算出更高效率之程序表的流程圖。 第16圖係成爲改善對象之程序表的說明圖。 -32- 201035541 第1 7圖係表示程序表之改善中之狀態的說明圖。 第1 8圖係本發明之第2實施形態之攝像檢查裝置的立 體圖。 第19(a)、（b)圖係該第2實施形態之攝像檢查裝置之正 面的模式圖和側面的模式圖。 第20(a)、（b)圖係本發明之第3實施形態之攝像檢査裝 置之正面的模式圖和側面的模式圖。 _ 第21圖係本發明之第4實施形態之第1實施例之攝像 〇 檢査裝置的立體圖。 第22(a)、（b)圖係該第4實施形態之第1實施例之攝像 檢查裝置之正面的模式圖和側面的模式圖。 第23圖係該第4實施形態之第2實施例之攝像檢查裝 置的立體圖。 第24(a)、（b)圖係該第4實施形態之第2實施例之攝像 檢査裝置之正面的模式圖和側面的模式圖。 Q 第25(a)、（b)圖係該第4實施形態之第3實施例之攝像 檢查裝置的立體圖。 第2 6(a)、（b)圖係該第4實施形態之第3實施例之攝像 檢查裝置之正面的模式圖和側面的模式圖。 第27(a)、（b)圖係本發明之第5實施形態之攝像檢查裝 置之正面的模式圖和側面的模式圖。 第2 8(a)、（b)圖係本發明之第6實施形態之第6實施例 之攝像檢査裝置之正面的模式圖和側面的模式圖。 -33- 201035541 第29(a)、（b)圖係該第6實施形態之第7實施例之攝像 檢查裝置之正面的模式圖和側面的模式圖。 第30圖係說明本發明之第7實施形態之移動機構之控 制的圖。 第31圖係習知例之攝像檢査裝置的立體圖。 第3 2圖係別的習知例之攝像檢査裝置的立體圖。 【主要元件符號說明】 200 攝 像 檢 查 裝 置 300 控 制 機 構 30 1 計 算 裝 置 101 第 一 Y 軸 可 動 件 102 第 二 Y 軸 可 動 件 103 檢 查 光 學 系 統 104 覆 查 光 學 系 統 105 第 二 Y 軸 軌 道 106 第 一 Y 軸 軌 道 107 X 軸 軌 道 108 X 軸 可 動' 件 109 基 板 110 抵 消 速 度 和 方 向 111 基 板 109 和 覆 査 度 及 方 向 701 檢 查 光 學 系 統 之 〇 ◎ -34- 201035541 702-704 705 檢査光學系統之掃描第2~第4行的軌跡 缺陷 Ο 706 在 掃 描 第 1 行所檢查的 區域 7 0 7- 709 在 掃 描 第 2- -第4行所檢 查的區域 710 覆 查 光 學 系 統之掃描第 2行的軌跡 71 1 覆 查 光 學 系 統之掃描第 3行的軌跡 712 覆 查 光 學 系 統之掃描第 4行的軌跡 801、 802 ' 803、 804 缺陷 901、 1001、 1101、 1201 覆查光學系 統的軌跡 Ρ 特 徵 量 的 總 和 Ν 覆 查 對 象 之 第一缺陷群 的個數 Ε 程 序 表 的 評 估値 -35-For example, if the resolution R of the review optical system 104 is 0.5x10 " 6[m/pix] (= 0.5[/zm/pix]), the exposure time T -24 - 201035541 is 1 X 10 - 3 [s] (=l[ms]), if the relative speed is greater than the relative speed V = 〇.〇〇〇25 [m/s], that is, V = 0.25 [mm/s], the subject vibrates. However, since the inspection optical system 103 generally operates at a relative speed of 100 [m/s] or more, the subject is vibrated, and some countermeasure is required. As a countermeasure against this, it is understood from the above formula (3) that the exposure time of the camera that reviews the optical system 104 is shortened. If the exposure time is shortened, it is possible to capture an image that is incapable of detecting the vibration of the subject while satisfying the relative speed V required for the inspection operation. For example, when 100 [m/s] is obtained as the relative speed V, the exposure time T may be 2.5 [//s] or less calculated from the equation (3). Specifically, a short exposure time can be achieved by using an electronic shutter or the like for the camera that reviews the optical system 104. However, simply shortening the exposure time will result in a darker image. Regarding this problem, it can be seen that it can be solved by using a high-sensitivity camera or using high-intensity illumination, but it is expensive to obtain a camera with high sensitivity, and even if the camera is used, it is difficult to set the exposure time to 1 0 [ μ S] below. 〇 Also, high-brightness illumination remains the same, and there is also concern about improving the adverse effects of illumination on the object being inspected. Therefore, in the present embodiment, not only the exposure time is shortened but also the flash illumination represented by the flash is used. If this flash illumination is used, since the illumination time can be shortened, the same effect as in the case of shortening the exposure time can be obtained in continuous light illumination. In the case where the flash has a flash speed of about 1 [ y s], the amount of light per unit time is also large. Therefore, by combining such a flash lamp and a camera of general sensitivity, an image in which the subject is not vibrated can be photographed even at a relative speed V = about -25 - 201035541 5 00 [mm/s]. For the same purpose, lasers can also be used in lighting. In this case, a shorter flash time can be achieved, and the amount of light per unit time is also large. Further, it is also possible to use a laser light or a device which corrects the defect of the object to be inspected and correct the defect at the same time as the inspection. (Second Embodiment) Fig. 18 and Figs. 19(a) and (b) show a second embodiment of the present invention. In the above-described first embodiment, as shown in Figs. 1 to 5, the inspection optical system 103 and the inspection optical system 104 are configured to move independently in the Y direction, and the inspection optical system 104 is configured to move the substrate 109. The entire width is different in the second embodiment in that the inspection optical system 104 is relatively moved in the Y direction with respect to the inspection optical system 103. That is, on the first Y-axis mover 101 of the inspection optical system 103, the inspection optical system 104 is mounted to move in the Y direction within the range of the first Y-axis movable member 101. Further, the moving mechanism in the Y direction of the inspection optical system 104 is larger than the range in the Y direction in which the inspection optical system 103 can be scanned. According to this configuration, even if the movable range of the moving mechanism in the Y direction of the optical system 104 is narrow, the entire width of the substrate 109 can be photographed by the inspection optical system, so even in the inspection optical system 1〇4 When the moving range of the moving mechanism in the Y direction is small, the opportunity for the inspection optical system 104 to capture the defect of the first defect group is not reduced. -26- 201035541 (Third Embodiment) Figs. 20(a) and (b) show a third embodiment of the present invention. In the second embodiment, the moving mechanism in the gamma direction of the inspection optical system 104 is larger than the range in the Y direction in which the inspection optical system 103 can be scanned. However, in the third embodiment, the inspection optical system 最低04 is the lowest. The Y-direction moving mechanism and the inspection optical system 103, which are required to have a limit, have the same range in the Y direction that can be scanned at one time. 0 According to this configuration, the opportunity for the inspection optical system 104 to take a defect of the first defect group is once. (Fourth Embodiment) Fig. 21 and Figs. 22(a), (b), 23, and 24(a), (b), 25, and 26 show the fourth aspect of the present invention, respectively. Implementation form. In each of the above-described embodiments, the inspection optical system 103 and the inspection optical system 104 are each singular. However, in the fourth embodiment, the inspection optical system 103 and the plurality of inspection optical systems 104 are provided. The difference is different. In the first embodiment of the fourth embodiment shown in Fig. 21 and Figs. 22(a) and (b), four Y-axis movable members 1〇1-1, ιοί-2, 101-3, 101 4 is disposed above the X-axis movable member 1〇8, and the first inspection optical system 103-1 is disposed on the Y-axis movable member 101-1. The second inspection optical system 103-2 is disposed on the Y-axis movable member 101-1. The third inspection optical system 103-3 is disposed on the Y-axis movable member 101-1. The fourth inspection optical system 103-4 is disposed on the Y-axis movable member 101-4. Further, the two Y-axis movable members 102-1 and 102-2 are disposed above the X-axis movable member 108 at -27-201035541. The first review optical system 104-1 is disposed on the Y-axis movable member 102-1. The second review optical system 104-2 is disposed on the Y-axis movable member 1 0 2 - 2. In the first embodiment of the fourth embodiment, the Y-axis movers 1 〇 2 _ 1 and 102-2 are configured to move over the common inspection optical system Y-axis 105, but in the 23rd and 24th (a) (b) In the second embodiment of the fourth embodiment shown in the drawing, the inspection optical systems Y-axis 1〇5-1, 105-2 are arranged side by side with each other interposed therebetween. The Y-axis movable member 102-1 is configured to move over the inspection optical system Y-axis 105-1, and the Y-axis movable member 102-2 is configured to move over the inspection optical system Y-axis 105-2. It is different from the first embodiment at one point. In this case, by moving the Y-axis movable members 102-1, 102-2, the first review optical system 104-1 and the second review optical system 104-2 can sequentially photograph the entire width of the substrate 109. In the third embodiment shown in Figs. 25(a) and (b), the inspection optical system Y-axis 105-1, 105-2 is arranged to be shifted in the Y-axis direction, and is firstly reviewed by 〇 The optical system 104-1 and the second review optical system 104-2 cover the entire width of the cover substrate 109. As shown in the first to third embodiments of the fourth embodiment, the actual inspection apparatus includes a plurality of inspection optical systems 103 and a plurality of inspection optical systems 104, and is configured as an inspection optical system 103 and The review optical system 104 may be independently movable in the Y direction, and may have a mechanism that can move the entire width of the substrate 109 as a whole. Further, it is also possible to check that the number of the optical system 103 and the inspection optical system 104 are different. -28- 201035541 Further, as shown in the fourth embodiment of the fourth embodiment shown in Figs. 26(a) and (b), the second review optical system 104-2 and the third review may be used. A plurality of review optical systems 1 to 4 of the optical system 104-3 are disposed on the Y-axis 105-2 of the inspection optical system. In the case where a plurality of review optical systems 1 to 4 are disposed on the same axis, it is necessary to consider that the inspection optical systems 104 do not collide with each other. To this end, a plurality of review optical systems 104 may also determine in advance the respective coverage ranges of the substrate 109, and may be controlled electrically or mechanically so as not to collide. Of course, in the case where the inspection optical system 103 is one set, it is also possible to increase the probability of being able to check with the inspection while checking the plurality of inspection optical systems 104 as described above. (Fifth Embodiment) Fig. 27 shows a fifth embodiment of the present invention. In the first embodiment of the fourth embodiment shown in FIG. 20 and FIGS. 21(a) and (b), the review optical system Y-axis 105 is provided with a plurality of inspection optical systems 104, but In the fifth embodiment of the fifth embodiment, only the singular inspection optical system 104 is provided on the inspection optical system Y-axis 105, and the inspection optical system 104 can move over the entire width of the substrate 109. The difference is different. (Sixth embodiment) Figs. 28(a), (b), and 29(a) and (b) show a sixth embodiment of the present invention. In the sixth embodiment of the sixth embodiment shown in Figs. 28(a) and (b), the first Y-axis rail 106 disposed on the upper surface of the X-axis movable member 108 is moved -29-201035541. The shaft movable member 101' first inspection optical system 103-1 and second inspection optical system 103-2 are disposed at intervals. The first and second review optical system shafts 10 - 1 , 1 05 - 2 are arranged side by side in the γ-axis movable member 101 at intervals. The first review optical system 104-1 is disposed on the first review optical system 105 shaft 105 - the second review optical system ι 〇 4_2 is disposed on the second review optical system 1 axis 1 0 5-2. The seventh embodiment of the sixth embodiment shown in Figs. 29(a) and (b) is configured as the first and second review optical systems 104-1, 104-2 along the common (first of 1) The movement of the Y-axis 105-1 of the optical system is different from that of the sixth embodiment of the sixth embodiment. As shown in the sixth embodiment and the seventh embodiment of the sixth embodiment, it is plural. The inspection optical system, the first and second inspection optical systems 104-1, 104-2, and the second embodiment shown in Fig. 18 and 19(a) and (b) Similarly, a moving mechanism having a narrow moving range can be used. (Embodiment 7) Fig. 30 shows a seventh embodiment of the present invention. The description of the first embodiment shown in Figs. 1 to 17 is shown. In order to check that the optical system 104 is photographing without stopping the substrate 109, it is necessary to take measures against the vibration of the subject. Although the method described in the first embodiment can also be used as a countermeasure, in the seventh embodiment, In order to reduce the influence of the vibration of the subject, the moving mechanism 104 is additionally provided with a moving mechanism in the X direction. The point is different from that of the first embodiment. -30- 201035541 When the optical system 104 is inspected to the moving mechanism optical system 104 in the X direction, the moving mechanism is driven in a direction to cancel the relative speed of the substrate 109 and the system 104. In Fig. 30, reference numeral 111 denotes the relative speed and direction of the substrate 109 and the inspection optical system. Since the relative speed and direction 111 are in the seventh embodiment, the optical system 104 is required to reduce the speed of the object and the relative speed and relative. The speed and the direction 111 are opposite to the offset speed indicated by 110. () Specifically, the moving mechanism can be realized by the mechanism of both the drive and the optical system as shown in Fig. 5. Again, because of this movement The center of the field of view of the optical system 104 can be moved in the X direction, that is, the parallel movement mechanism to the X-axis direction is used, and the application of the optical system 104 can be used to make the center of the field of view movable. The present invention can contribute to improvement in productivity of a large substrate used in a liquid crystal display device, an electric device, a solar cell, etc. [Simplified Schematic] Fig. 1 Fig. 2(a) and (b) are a schematic view and a side view of the front side of the apparatus according to the first embodiment of the first embodiment. Fig. 3 is a perspective view of the inspection object moving inspection device in the first embodiment. It is a matter of reviewing the inspection optical system 104. Therefore, it is not necessary to have the same direction as the substrate 109 mechanism. The image of the front surface of the image-inspection apparatus of the third embodiment of the first embodiment is shown in Fig. 4 (a) and (b). Schematic diagram of the figure and side. Figs. 5(a) and 5(b) are a schematic view and a side view of the front surface of the image inspecting apparatus of the first embodiment. Fig. 6 is a flow chart at the time of inspection and review in the first embodiment. Figure 7 is a flow chart showing the first half of the processing flow for program creation. 〇 Figs. 8(a) and 8(b) are diagrams showing the trajectory of the inspection optical system and the inspection optical system according to the first embodiment. Figure 9 is a diagram showing a list of the first defect groups that have not been closed. Fig. 10 is a view showing a state of the first defect group list (one defect number) in the middle of the creation of the program table. Fig. 11 is a view showing a state of the first defect group list (two defects) in the middle of the creation of the program table. Q Fig. 12 is a state diagram showing the first defect group list (three defects) in the middle of the schedule creation. Figure 13 is a state diagram showing a list of the first defect groups in which the program table creation has ended. Figure 14 is an explanatory diagram of an example in which many first defect groups cannot be reviewed because a certain defect group is reviewed. Figure 15 is a flow chart for calculating a program of higher efficiency. Fig. 16 is an explanatory diagram of a program table to be improved. -32- 201035541 Fig. 17 is an explanatory diagram showing the state of improvement in the program table. Fig. 18 is a perspective view showing an imaging inspection apparatus according to a second embodiment of the present invention. 19(a) and 9(b) are a schematic view and a side view of the front surface of the image inspecting apparatus of the second embodiment. 20(a) and (b) are a schematic view and a side view of the front side of the imaging inspection apparatus according to the third embodiment of the present invention. Fig. 21 is a perspective view of the imaging 〇 inspection apparatus according to the first embodiment of the fourth embodiment of the present invention. 22(a) and (b) are a schematic view and a side view of the front side of the imaging inspection apparatus according to the first embodiment of the fourth embodiment. Fig. 23 is a perspective view showing an imaging inspection apparatus according to a second embodiment of the fourth embodiment. Figs. 24(a) and (b) are a schematic view and a side view of the front side of the imaging inspection apparatus according to the second embodiment of the fourth embodiment. Q (a) and (b) are perspective views of the imaging inspection apparatus according to the third embodiment of the fourth embodiment. Figs. 6(a) and 6(b) are a schematic view and a side view of the front side of the imaging inspection apparatus according to the third embodiment of the fourth embodiment. (a) and (b) are schematic diagrams and side views of the front surface of the imaging inspection apparatus according to the fifth embodiment of the present invention. (2) and (b) are schematic diagrams and side views of the front surface of the image pickup inspection apparatus according to the sixth embodiment of the sixth embodiment of the present invention. -33- 201035541 Fig. 29 (a) and (b) are a schematic view and a side view of the front side of the imaging inspection apparatus according to the seventh embodiment of the sixth embodiment. Fig. 30 is a view for explaining the control of the moving mechanism of the seventh embodiment of the present invention. Figure 31 is a perspective view of a conventional imaging inspection apparatus. Fig. 3 is a perspective view of a conventional imaging inspection apparatus. [Main component symbol description] 200 Camera inspection device 300 Control mechanism 30 1 Calculation device 101 First Y-axis movable member 102 Second Y-axis movable member 103 Inspection optical system 104 Review optical system 105 Second Y-axis track 106 First Y Axis track 107 X-axis track 108 X-axis movable 'Part 109 Substrate 110 Offset speed and direction 111 Substrate 109 and coverage and direction 701 Check optical system ◎ -34- 201035541 702-704 705 Check optical system scan 2 ~ Track defect 第 706 of line 4 706 In the area 7 0 7- 709 checked in the 1st line of scanning, in the area 710 inspected by scanning the 2nd - 4th line, the track of the scanning line 2 of the optical system is checked 71 1 Review the trajectory 712 of the scanning line of the optical system. Review the trajectory 801, 802 '803, 804 of the scanning line 4 of the optical system. The defects 901, 1001, 1101, 1201 check the trajectory of the optical system Ρ the sum of the characteristic quantitiesΝ Review the first defect group of the object The number of Ε program evaluations 値 -35-