201028681 六、發明說明： 【發明所屬之技摘^領域】 發明相關之技術領域 本發明係有關於用以檢驗形成在使用於大型液晶面板 等之大型基板上之電信號配線及電晶體等之圖案缺陷的缺 陷檢查裝置及缺陷檢查方法。 _ 先前技術 α彳主’在TFT液晶基板製造程序之微影程序中，會因製 造裝置或曝光裝置等之異常，於形成在基板上之圖案產生 缺陷。例如’顆粒附著於基板時，產生局部之圖案偏移， 此部份成為不均而顯現出來。 又，由於曝光裝置之異常，特別是在近年來作為曝光 手段之透鏡掃描方式的情形下，透鏡調整異常會發生透 鏡間之圖案偏移，以條紋狀之不均顯現。即使微小之圖案 ❿ 鮮也會發生這些不均，但由於在電氣上不會有異常，所 以在後續程序中很難發現。因此，進行藉以人的肉眼觀察 之巨觀檢查’進行不均之檢測。 自動定S化赶觀檢查之方法包括以比影像解析度小 之精度預先使用良品基板先取得重覆圖案之特徵點出現時 所預測之預測位置資m，藉求取冑述預測位置與測量位置 之誤差，檢驗圖案偏移之技術(例如，參照專利文獻 【專利文獻1】日本專利公開公報：特開2〇〇4_279244 號公報 201028681 【赛^明内溶1】 【發明欲解決之技術課題】 但是’先前技術中，必須預先使用真實基板產生預測 位置資Λ該基板必須是良品。然而，由於在製造程序開 始時製造1絲板，所以錢前技射有在此 之前無法進行符合檢查的問題。 又，亦有必須具有作為檢查對象之全機種及全程序中 之基板上檢查位置的全部位置資訊，因此其資訊量增 大的問題。 此外，亦有比較預測位置與測量位置時必須嚴密對 位，處理複雜化的問題。 本發明係有鑑於上述問題點而作成者，其目的在於提 供缺陷檢查裝置及缺鎌查方法，其可僅藉—部份良品之 圖案資訊，檢驗檢查基板之圖案偏移，且亦不需要檢查區 域全面之預測位置資訊，可抑制至所需最小限度之資訊量。 【用以解決課題採用之技術方法】 為解決上述課題之本發明缺陷檢查裝置包含對被檢查 基板攝影之攝㈣構；取得料職频構所攝得之基板 影像之影像取得機構；記‘_前《像取得機構所取得之 基板影像作為檢查影狀景彡像記憶_，·設定圖案匹配用 之參考影像與檢查條件之參考影像設定機構；使用前述檢 查影像與前述參考影像進行圖案匹配之_檢驗機構；及 藉前述圖案檢驗機構之難㈣之結果，算出前述檢查影 像與前述參考影像之偏移量之圖案偏移量算出機構。 201028681 【發明具有之優點】 藉本發明，可不必由良品基板事先作成資訊，且藉簡 易之機構檢驗圖案之偏移。 此外，藉本發明，可判定顯現之不均’因此可早期發 現製造程序之異常。 【實方式;3 發明之實施態樣 _ 本發明採用如下所述之構造。 即’依據本發明之-態樣，本發明之缺陷檢查裝置之 特徵在於包含對被檢查基板攝影之攝影機構；取得藉前述 攝影機構所攝得之基板影像之影像取得機構；記憶藉S述 影像取得機構所取得之基板影像作為檢查影像^像=憶 機構；設定圖案匹配用之參考影像與檢查條件之參考景 設定機構W吏用前述檢查影像與前述參考影像進行圖，案匹 配之圖案檢驗機構；及藉前述圖案檢驗 ” 結果，算出前述檢查影像與前述參考影 量= 偏移量算出機構。 偏移量之圖案 又，較佳地，本發明之缺陷檢查裝 量算出機構依據藉前述圖案偏移量算⑭_、圖案偏移 量，算出前述檢查影像與前述參考影像^算出之偏移 及偏移方向。 〈圖案的偏移距離 又，較佳地，本發明之缺陷檢查裝 量算出機構依據前述偏移量之分布，_^述圖案偏移 基板之不均的麵。 @ A於前述檢查 201028681 又’較佳地，本發明之缺陷檢查裝置具有偏移量顯示 機構，該偏移量顯示機構顯示藉前述圖案偏移量算出機構 所算出之偏移量的資訊。 又，較佳地’本發明之缺陷檢查裝置係前述偏移量顯 示機構視覺性地顯示藉前述圖案偏移量算出機構所算出之 前述檢查與前述參考影像之圖案的偏移距離及偏移方 向。 又，較佳地，藉本發明之一態樣，本發明之缺陷檢查 方法係檢查被檢查基板之缺陷的缺陷檢查方法其特徵在 於使用攝影前述被檢查基板之檢查影像與參考影像，進行 圖案匹配，藉前述圖案匹配之結果，算出前述檢查影像與 月’j述參考影像之偏移距離及偏移方向，判斷顯現於前述被 檢查基板之不均的種類。 以下，參照圖式說明本發明之實施形態。 首先，說明適用本發明之第丨實施形態。 第1圖是適用本發明之第1實施形態中之缺陷檢查裝置 的機能方塊圖。 第1圖中，缺陷檢查裝置10包含作為攝影機構之照像機 邛2、影像取得部3、影像記憶部4、參考影像設定部^、圖 案檢驗部7、圖案偏移量算出部8及偏移量顯示部9。 前述照像機部2可對被載置在平台上之被檢查基板1上 之預先設定多數位置的檢查對象攝影’作成〗維影像或2維 影像。前述影像取得部3取得藉前述照像機部2攝得之基板 影像，前述影像記憶部4可記憶藉前述影像取得部3取得之 201028681 月述參考影像攻定部5取出包含記憶於前述影像記憶 部4之基板影像之無缺陷特徵性圖案的—部份，蚊圖案匹 用之參考f/像’同時設定檢查區域。參考影像亦可由以 照像機部2攝影正確形成W案之良品基板騎之基板影像 選出無圖案偏移之像素，切出包含該錢案偏移之像素内[Technical Field] The present invention relates to a pattern for testing electrical signal wiring and a transistor formed on a large substrate used for a large liquid crystal panel or the like. Defect inspection device and defect inspection method. In the lithography process of the TFT liquid crystal substrate manufacturing process, defects occur in the pattern formed on the substrate due to abnormalities in the manufacturing apparatus or the exposure apparatus. For example, when a particle adheres to a substrate, a partial pattern shift occurs, and this portion becomes uneven. Further, in the case of the abnormality of the exposure apparatus, particularly in the case of the lens scanning method which is an exposure means in recent years, the lens adjustment abnormality causes a pattern shift between the lenses, and the unevenness of the stripes appears. Even if the tiny pattern is fresh, these unevenness will occur, but since there is no electrical abnormality, it is difficult to find in subsequent procedures. Therefore, the inspection of the unevenness by the naked eye of the human is performed. The method for automatically determining the S-view inspection includes pre-predicting the position of the predicted feature position when the feature point of the repeated pattern is obtained by using the good substrate with less precision than the image resolution, and taking the predicted position and the measurement position. In the case of the error, the technique for inspecting the pattern shift (for example, refer to the patent document [Patent Document 1] Japanese Patent Laid-Open Publication No. Hei. No. Hei. No. 2-279244 No. 201028681. However, in the prior art, it is necessary to use a real substrate in advance to generate a predicted position. The substrate must be a good product. However, since the first board is manufactured at the beginning of the manufacturing process, the pre-money technique has problems that cannot be checked before. In addition, it is necessary to have all the position information of the inspection position on the substrate in the entire model and the entire program, and therefore the information amount is increased. In addition, it is necessary to compare the predicted position and the measurement position. The problem of processing is complicated. The present invention has been made in view of the above problems, and its purpose is to provide a defect. Inspection device and defect inspection method, which can only check the pattern offset of the substrate by using the pattern information of some good products, and also does not need to check the comprehensive predicted position information of the area, and can suppress the minimum amount of information required. [Technical method for solving the problem] The defect inspection device of the present invention for solving the above-mentioned problems includes a photographing of a substrate to be inspected (four); an image obtaining mechanism for obtaining a substrate image obtained by a material frequency structure; _Before the "substrate image obtained by the acquisition mechanism as a reference image memory _, a reference image setting mechanism for setting a reference image for matching the image and the inspection condition; and pattern matching using the inspection image and the reference image And a pattern shift amount calculation means for calculating an offset between the inspection image and the reference image by the result of the difficulty of the pattern inspection mechanism (4). 201028681 [Advantages of the Invention] By the present invention, it is not necessary to The substrate is made into information in advance, and the offset of the pattern is checked by a simple mechanism. Further, by the present invention, It is judged that the appearance is uneven. Therefore, the abnormality of the manufacturing process can be found early. [Embodiment; 3 Embodiment of the invention] The present invention adopts the configuration described below. That is, the defect inspection of the present invention is based on the aspect of the present invention. The device is characterized by comprising: an image capturing mechanism for photographing the substrate to be inspected; an image obtaining mechanism for obtaining a substrate image captured by the photographing mechanism; and storing the substrate image obtained by the image capturing device as an inspection image; a reference scene setting mechanism for setting a reference image for matching the image and a reference image for the inspection condition, and a pattern inspection mechanism for matching the image with the reference image; and the result of the pattern inspection to calculate the inspection image and the foregoing Reference amount = offset calculation means. Offset pattern. Further, the defect inspection load calculation means of the present invention calculates the inspection image based on the pattern shift amount 14_ and the pattern shift amount. The reference image ^ calculates the offset and offset direction. <Offset Distance of Pattern Further, preferably, the defect inspection amount calculation means of the present invention shifts the uneven surface of the substrate in accordance with the distribution of the offset amount. Further, the defect inspection device of the present invention has an offset display means for displaying information on the offset amount calculated by the pattern shift amount calculation means. Further, in the defect inspection apparatus of the present invention, the offset amount display means visually displays an offset distance and an offset direction of the inspection and the pattern of the reference image calculated by the pattern shift amount calculation means. . Moreover, preferably, in one aspect of the present invention, the defect inspection method of the present invention is a defect inspection method for inspecting defects of an inspected substrate, which is characterized in that pattern inspection is performed by photographing the inspection image and the reference image of the substrate to be inspected. By the result of the pattern matching, the offset distance and the offset direction of the inspection image and the reference image are calculated, and the type of unevenness appearing on the substrate to be inspected is determined. Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, a third embodiment to which the present invention is applied will be described. Fig. 1 is a functional block diagram of a defect inspection device according to a first embodiment of the present invention. In the first drawing, the defect inspection device 10 includes a camera 2 as an imaging unit, a video acquisition unit 3, a video storage unit 4, a reference image setting unit, a pattern check unit 7, a pattern shift amount calculation unit 8, and a bias. The shift amount display unit 9. The camera unit 2 can create a video or a two-dimensional image of the inspection target image ‘by a predetermined number of positions on the substrate 1 to be inspected placed on the platform. The image acquisition unit 3 obtains the substrate image captured by the camera unit 2, and the image storage unit 4 can store the reference image acquisition unit 5 acquired by the image acquisition unit 3, which is stored in the image memory. The portion of the non-defective characteristic pattern of the substrate image of the portion 4, the reference f/image of the mosquito pattern is used to simultaneously set the inspection area. The reference image may also be selected from the substrate image of the substrate on which the camera substrate 2 is correctly formed by the camera unit 2, and the pixel having no pattern offset is selected, and the pixel including the offset of the money is cut out.
之特徵性形狀的—部份，作為圖案匹配用之參考影像來登 錄。 前述參考影像記憶部6記憶藉參考影像設定部5設定登 錄之參考影像及檢查條件。圖錄驗部7使用在前述被檢查 基板1之檢查時取得且記胁《料記‘_4之檢查影像 及記憶於前述參考影像記憶部6之參考影像，進行圖案匹 配，檢驗檢查影像與參考影像一致之位置。圖案偏移量算 :部8藉前述圖案檢驗部7之圖案匹配的結果，依據前述檢 —影像與參考影像之差，算出圖案之偏移距離與偏移方向。 此外，前述偏移量齡部9作成_藉㈣圖案偏移量 异出部8算出之圖案偏移的結果，且例如，於顯示前述偏移 =與偏移方向等之符號賦予顏色，或使前㈣號亮滅以 祝覽性地顯示偏移距離與偏移方向。 i如前所述，前述參考影像設定部5取出記憶於前述影像 記憶部4之檢查影像的一部份，設定圖案匹配用之參考二像 等，接著，說明其具體例。 第2圖是用以說明參考影像之設定的圖。 201028681 如第2圖所示，前述參考影像設定部5僅取出液晶1像素 量之記憶於檢查影像之一部份且未產生圖案偏移的影像。 接著，以此影像作為參考影像100,由該參考影像100中切 出適於圖案匹配之特徵性形狀的圖案，設定圖案匹配用之 模型101。又，亦可設定液晶顯示裝置之“象素量的全部作 為圖案匹配用之模型101。 又，前述參考影像設定部5設定檢測位置偏移之檢查區 域。在第2圖所示之例中，設定模型1〇1之特徵性形狀之中 心於基準座標點，同時將第1檢查區域1〇2設定在距離所設 定之基準座標點XI、Y1之位置處’且將第2檢查區域1〇3設 疋在距離基準座標點X2、Y2之位置處。又，前述參考影像 設定部5在前述被檢查基板1中設定成為檢查對象之多數點 的檢查位置。並且，如前所述，這些資訊被記憶於參考影 像記憶部6。 接著，對用以檢查前述被檢查基板1之檢查處理流程進 行說明。 第3圖是顯示用以檢查被檢查基板之檢查處理之流程 的流程圖，第4圖是用以說明檢查對象與模型之圖案匹配的 圖，第5圖是顯示差影像之例的圖，第6圖是用以說明偏移 距離之圖。 首先，在第3圖之步驟S31中，照像機部2對被檢查基板 1進行攝影。此外，省略缺陷檢查裝置1〇之機械動作的說明。 接著，在步驟S32中，在步驟S31攝得之基板影像透過 影像取得部3記憶於影像記憶部4中。 201028681 然後，在步驟S33中，圖案檢驗部7由在步驟S32記憶於 影像記憶部4之基板影像中，僅依序讀取檢查對象區域之影 像的預先設定區域部份。接著，在步驟834中，前述圖案檢 驗部使用記憶於前述參考影像記憶部6之圖案匹配用之模 型101實行圖案匹配。具體而言，如第4圖所示，全部地檢 驗基板影像中之符合圖案，以次像素之精度求出與模型 一致之檢驗位置M(x(ij)，。 然後，在第3圖之步驟S35中，由在步驟S34求出之檢驗 位置Μ，切出與參考影像1〇〇相同尺寸之檢查影像丨〗〇，算 出與參考影像100之差。在此，於檢查影像11〇與參考影像 100之差處理中，首先，於檢查影像110與參考影像100實施 平均化等之影像過濾。接著，如以下條件式1所示，就分別 對應2個影像之影像的差值dP(ij)為預先設定之對比閾值cs 以下的像素而言，同樣地設定像素值=128，作為背景影像， 且將差值大於對比閾值Cs成為負值者同樣地設定像素值 =0，而將成為正值者設定像素值=255。 ldP(ij)|$Cs時，dP(ij)=128 ldP(ij)|>CsndP(ij)>0時，dP(ij)=255 dP(ij)>CsndP(ij)<0時，dP(ij)=0 …條件式1 因此，如第5圖所示，成為僅強調圖案偏移之一部份的 差影像120。 接著，使用求取位置偏移之檢查區域102及檢查區域 103，如第6圖所示，例如’在檢查區域102中檢驗X方向之 9 201028681 偏移時’於圖案之兩側發生偏移，因此，以X方向之像素值 0之像素數Dx(ij)與像素值255之像素數Bx(ij)的平均值作為 X方向之偏移距離Px(ij)(條件式2)。又，在檢查區域103中檢 驗Y方向之偏移時，以Y方向之像素值〇之像素數〇7(出與像 素值255之像素數By(ij)的平均值作為γ方向之偏移距離 Py(ij)(條件式 2)。The part of the characteristic shape is registered as a reference image for pattern matching. The reference image storage unit 6 stores the registered reference image and the inspection condition by the reference image setting unit 5. The image recording unit 7 performs pattern matching and checks the inspection image and the reference image by using the inspection image of the material record '_4 and the reference image stored in the reference image memory unit 6 when the inspection of the substrate 1 to be inspected is performed. Consistent location. The pattern shift amount calculation unit 8 calculates the offset distance and the shift direction of the pattern based on the result of pattern matching by the pattern check unit 7 based on the difference between the detected image and the reference image. Further, the offset-aged portion 9 is a result of pattern shift calculated by the (four) pattern shift amount difference portion 8, and for example, a color is given by displaying a symbol such as the offset = offset direction or the like, or The front (4) is turned on to show the offset distance and the offset direction in a conspicuous manner. As described above, the reference image setting unit 5 takes out a part of the inspection image stored in the image storage unit 4, sets a reference image for pattern matching, and the like. Next, a specific example will be described. Fig. 2 is a diagram for explaining the setting of a reference image. 201028681 As shown in Fig. 2, the reference image setting unit 5 extracts only the image of the liquid crystal 1 pixel which is stored in one of the inspection images and which does not cause a pattern shift. Next, this image is used as the reference image 100, and a pattern suitable for pattern matching is cut out from the reference image 100, and the pattern matching pattern 101 is set. Further, all of the "amount of pixels of the liquid crystal display device" may be set as the pattern 101 for pattern matching. Further, the reference image setting unit 5 sets an inspection region for detecting a positional shift. In the example shown in Fig. 2, Setting the center of the characteristic shape of the model 1〇1 to the reference coordinate point, and setting the first inspection area 1〇2 at a position away from the set reference coordinate point XI, Y1' and the second inspection area 1〇3 The reference image setting unit 5 sets an inspection position of a plurality of points to be inspected in the inspection target substrate 1 as described above. Further, as described above, the information is The memory is stored in the reference image storage unit 6. Next, the flow of the inspection process for inspecting the substrate 1 to be inspected will be described. Fig. 3 is a flow chart showing the flow of the inspection process for inspecting the substrate to be inspected, and Fig. 4 is a flow chart. A diagram for explaining the matching between the inspection object and the model, FIG. 5 is a diagram showing an example of a difference image, and FIG. 6 is a diagram for explaining an offset distance. First, in step S31 of FIG. 3, image The machine unit 2 photographs the substrate 1 to be inspected. The description of the mechanical operation of the defect inspection device 1 is omitted. Next, in step S32, the substrate image captured by the image acquisition unit 3 in step S31 is stored in the image memory unit. 4: 201028681 Then, in step S33, the pattern check unit 7 reads only the predetermined area portion of the image of the inspection target region in the substrate image stored in the image storage unit 4 in step S32. In step 834, the pattern check unit performs pattern matching using the pattern 101 for pattern matching stored in the reference image storage unit 6. Specifically, as shown in Fig. 4, all of the matching patterns in the substrate image are inspected to The accuracy of the sub-pixel is obtained by finding the inspection position M (x(ij) in accordance with the model. Then, in step S35 of Fig. 3, the inspection position Μ obtained in step S34 is cut out and the reference image is cut. The same size inspection image 丨 〇 〇, calculate the difference from the reference image 100. Here, in the difference between the inspection image 11 〇 and the reference image 100, first, the inspection image 110 and the reference image 100 Image filtering is performed by averaging, etc. Next, as shown in the following conditional expression 1, the pixels d2 (ij) corresponding to the images of the two images are set to a pixel having a contrast threshold cs or less set in advance, and the pixels are set in the same manner. The value = 128 is used as the background image, and the pixel value = 0 is set in the same way that the difference value is larger than the comparison threshold Cs, and the pixel value is set to 255 in the positive value. ldP(ij)|$Cs, dP (ij)=128 ldP(ij)|>CsndP(ij)>0, dP(ij)=255 dP(ij)>CsndP(ij)<0, dP(ij)=0 ...condition Equation 1 Therefore, as shown in Fig. 5, the difference image 120 which emphasizes only one part of the pattern shift is obtained. Next, using the inspection area 102 and the inspection area 103 for ascertaining the positional deviation, as shown in FIG. 6, for example, when the inspection of the X direction is performed in the inspection area 102, the offset is made on both sides of the pattern. Therefore, the average value of the pixel number Dx (ij) of the pixel value 0 in the X direction and the pixel number Bx (ij) of the pixel value 255 is taken as the offset distance Px (ij) in the X direction (Condition 2). Further, when the offset in the Y direction is checked in the inspection region 103, the pixel value 〇7 of the pixel value Y in the Y direction (the average value of the number of pixels By(ij) of the pixel value 255 is taken as the offset distance in the γ direction. Py(ij) (Condition 2).
Px(ij)=(Dx(ij)+Bx(ij))/2 Py(ij )=(Dy (ij )+By(ij ))/2 …條件式2 此外，在差影像120中，在由像素值225變成像素值〇之 方向上圖案偏移’偏移方向亦可算出，若對偏移方向設定 符號，則可藉將距離R(Px(ij)，Py(ij))乘上符號而數值化。 接著，在第3圖之步驟S36中，使在步驟S34及步驟S35 求出之檢驗位置M(X(ij)，Y(ij))與偏移距離(px(ij)，py(ij))的 結果顯示於偏移量顯示部9之螢幕上，使操作者了解狀況。 又，使用向量箭號一下就可看出作為結果顯示之偏移距離 與偏移方向，可輕易掌握狀況。 以下’說明適用本發明之第2實施形態。 第2實加形態係依據所檢驗之偏移的檢驗位置μ與偏 移距離R的關係，將判斷所檢驗之偏移成為何種不均之程序 加入第1實施形態者。又，由於在算出檢驗位置河與偏移距 離R之前的過程與第一實施形態相同，故省略說明。 第7圖是顯示起因於附著於基板之顆粒之不均之例的 圖，第8圖是顯示起因於曝光裝置之掃描透鏡間之偏移之不 201028681 均之例的圖，第9圖是顯示曝光裝置之基板固定機構之位置 偏移之例的圖。 圖案偏移量算出部8在算出檢驗位置Μ與偏移距離R 後，所算出之偏移距離R(Px(ij)，Py(ij))的絕對值求出成為預 先設定之規格值ST以上之檢驗位置MM(X(ij)，Y(ij))，並且 求出檢驗位置MM之分散程度，求出對應不均之種類。 例如，如第7圖所示，以向量表示偏移距離與偏移方向 時’於某範圍内向量聚在一起成圓形時，判斷為起因於附 著於被檢查基板1之顆粒的不均。又，如第8圖所示，當看 見向量呈於某方向上連續之條紋狀時，判斷為起因於曝光 裝置之掃描透鏡間之偏移的不均。此外，如第9圖所示，同 樣地分散於基板影像之全體且偏移方向一定時，判斷為曝 光裝置之基板固定機構發生位置偏移。 如此’顯示作為結果顯示之偏移距離與偏移方向時， 藉所檢驗之偏移特徵判斷顯現何種不均，藉此可早期發現 被檢查基板1之異常。 本發明不限於以上所述之各實施形態等，在不偏離本 發明之要旨之範圍内，可以採用種種構造或形狀。 【圖式*簡曰月】 第1圖是適用本發明之第i實施形態中之缺陷檢查裝置 的機能方塊圖。 第2圖是用以說明參考影像之設定的圖。 第3圖是顯不用以檢查被檢查基板之檢查處理之流程 的流程圖。 201028681 第4圖是用以說明檢查對象與模型之圖案匹配的圖。 第5圖是顯示差影像之例的圖。 第6圖是用以說明偏移距離之圖。 第7圖是顯示起因於附著於基板之顆粒之不均之例的 圖。 第8圖是顯示起因於曝光裝置之掃描透鏡間之偏移之 不均之例的圖。 第9圖是顯示曝光裝置之基板固定機構之位置偏移之 例的圖。 【主要元件符號說明】Px(ij)=(Dx(ij)+Bx(ij))/2 Py(ij )=(Dy (ij )+By(ij ))/2 (Conditional expression 2) Further, in the difference image 120, The pixel value 225 becomes the pixel value 〇 in the direction of the pattern offset 'offset direction can also be calculated. If the symbol is set to the offset direction, the distance R (Px(ij), Py(ij)) can be multiplied by the symbol. Numerical. Next, in step S36 of FIG. 3, the test positions M (X(ij), Y(ij)) and the offset distance (px(ij), py(ij)) obtained in steps S34 and S35 are obtained. The result is displayed on the screen of the offset display portion 9, so that the operator can understand the situation. Also, using the vector arrow, you can see the offset distance and offset direction displayed as results, and you can easily grasp the situation. The following describes the second embodiment to which the present invention is applied. The second actual form is a program for determining the unevenness of the detected offset based on the relationship between the inspection position μ and the offset distance R of the detected offset, and is added to the first embodiment. Further, since the process before the calculation of the check position river and the offset distance R is the same as that of the first embodiment, the description thereof is omitted. Fig. 7 is a view showing an example of unevenness of particles due to adhesion to a substrate, and Fig. 8 is a view showing an example of not all of the offsets between the scanning lenses of the exposure device, 201028681, and Fig. 9 is a view showing A diagram showing an example of the positional shift of the substrate fixing mechanism of the exposure apparatus. After calculating the inspection position Μ and the offset distance R, the pattern shift amount calculation unit 8 obtains the absolute value of the calculated offset distance R (Px(ij), Py(ij)) to be a predetermined specification value ST or more. The position MM (X(ij), Y(ij)) is checked, and the degree of dispersion of the inspection position MM is obtained, and the type of the corresponding unevenness is obtained. For example, as shown in Fig. 7, when the offset distance and the offset direction are represented by a vector, when the vectors are gathered into a circle within a certain range, it is determined that the unevenness of the particles attached to the substrate 1 to be inspected is caused. Further, as shown in Fig. 8, when it is seen that the vector is continuous in a stripe shape in a certain direction, it is determined that the shift due to the shift between the scanning lenses of the exposure device is uneven. Further, as shown in Fig. 9, when the entire image of the substrate is dispersed in the same manner and the offset direction is constant, it is determined that the substrate fixing mechanism of the exposure device is displaced. When the offset distance and the offset direction displayed as the result are displayed, it is judged by the offset characteristic of the check which unevenness is exhibited, whereby the abnormality of the substrate 1 to be inspected can be found early. The present invention is not limited to the above-described embodiments and the like, and various configurations or shapes can be employed without departing from the gist of the present invention. [Picture * Simple Moon] Fig. 1 is a functional block diagram of a defect inspection device to which the i-th embodiment of the present invention is applied. Fig. 2 is a diagram for explaining the setting of a reference image. Fig. 3 is a flow chart showing the flow of the inspection process which is not used to inspect the substrate to be inspected. 201028681 Figure 4 is a diagram for explaining the matching of the inspection object with the pattern of the model. Fig. 5 is a view showing an example of a difference image. Figure 6 is a diagram for explaining the offset distance. Fig. 7 is a view showing an example of unevenness of particles due to adhesion to a substrate. Fig. 8 is a view showing an example of the variation in the offset between the scanning lenses caused by the exposure device. Fig. 9 is a view showing an example of the positional deviation of the substrate fixing mechanism of the exposure apparatus. [Main component symbol description]
1.. .被檢查基板 2.. .照像機部 3.. .影像取得部 4.. .影像記憶部 5.. .參考影像設定部 6.. .參考影像記憶部 7.. .圖案檢驗部 8.. .圖案偏移量算出部 9.. .偏移量顯示部 10.. .缺陷檢查裝置 100.. .參考影像 101.. .圖案匹配用之模型 102.. .第1檢查區域 103…第2檢查區域 110.. .檢查影像 120.. .差影像 S31-S36...步驟1.. Inspection substrate 2: Camera unit 3. Image acquisition unit 4. Image memory unit 5.. Reference image setting unit 6. Reference video storage unit 7.. Inspection unit 8. Pattern offset calculation unit 9. Offset display unit 10. Defect inspection device 100.. Reference image 101.. Pattern matching model 102.. . Area 103...Second inspection area 110.. .Check image 120..Difference image S31-S36...step
1212