200916763 九、發明說明： c發明所屬之技術領域3 發明領域 本申請案主張2007年3月3日申請之第60/941,672號美 5 國繼續申請案之優先權。 本發明關於自動光學檢測系統。 L先前技術3 發明背景 在生產線終端，在封裝與運送之前，拋光HDI(高密度互連) 10 基板會以視覺檢查表面缺陷及/或違反整體性。這類視覺檢 查將會依據品管要件把它們分類為有效或失效。 表面缺陷係令不滿意的產品品質及/或令不滿意的處理結 果。在此，所有的品質係關於全部的表面類型，即。金屬 化表面諸如在層壓基材上之鍍金互連塾以及焊接遮罩。這 15 類表面之每一類型係被特定類型的缺陷所影響，譬如其原 始材料之產物及製造過程。 從光學的角度來看，各材料引用不同的光學特性，因此需 要在被檢測時定址（手動或經由自動機器）。 提供有效率的缺陷檢測方法和系統是與日俱增之需求。 20 【發明内容】 發明概要 一種系統，其包括一處理器、一成像路徑、以及多重 照明路徑；該處理器從多重照明路徑中選擇一照明路徑， 使得選定的照明路徑照明一導體表面以及一導體表面缺 200916763 陷，使得不論導體導體表面之粗链程度，表面缺陷之-影 :實質:為黑色的。，而導體表面之-影像實質上為白： ' 、中該處㈣選擇—照明路徑，其照明-半透明介 電表面以及-半透明介電表面缺陷，使得不論半透明表面 之粗链私度為何，半透明表面缺陷之影像實質上為白色 的’而半透明表面之影像實質上為黑色的。 、一種藉由-多重照明路徑系統來檢測缺陷之方法，該 2法包含下列步驟：在照明—導體表面時，在—缺陷檢測 間選擇一選定的照明路徑來被致用，使得該選定照 1〇明路控照明導體表面與導體表面缺陷時，不論導體導體表 面之粗糖程度，表面缺陷之一影像實質上為黑色的，而導 體表面之#像實質上為白色的；當照明一半透明表面 寺在心檢測掃描期間選擇一選定的照明路徑來被致 用使仟不，半透明表面之粗縫程度為何，半透明表面缺 5陷之衫像實質上為白色的，而半透明表面之影像實質上 黑色的。 、’ 。。-種用於缺陷檢測之系統，該系統包括包括一處理 益、-成像路徑、以及多重照明路徑；其中該系統適應於 僅藉由與該節段相關聯之一選定照明路徑來照明一物件之 〇每節段，其中該選定照明路徑係響應於自該節段之—主 要材料與該節段資訊之一表面粗糙程度中之至少-個參數 所選出’其中該成像路徑適應於響應該照明產生表示缺陷 之檢測H 其中該處理器處理該檢測信號來檢測缺陷。 種夕重照明路徑系統，該系統包括多重照明路徑、 200916763 處理l§、以及-成像路徑資訊，其中該處理器適應於： 控制-物件資訊之-初始掃描，響應於在該初始掃描資訊 期間獲得的檢測信號來將該物件分節成節段；判定每一節 5 10 15 段資訊之-表面粗糖度；以及響應於自該節段之一主要材 料與該節段資訊之-表面粗#程度巾選定之至少一個參 數，每-節段選擇-選定照明路徑來在該節段之一缺陷檢 測處理期間被致動；其中該等多重照明路徑適應於在該初 始掃描資訊期間同步照明該物件，且其中一選定照明路徑 找物件資訊之-缺陷檢測掃描期間照明與該選定照明路 U目關如之又，且其中該成像路徑適應於響應於該照 明產生表示缺陷之檢測信號。 〜於缺陷檢測之方法，該方法包含：僅藉由與一 1段資訊相關聯之-選定照明路徑來照明一物件之每一節 段’其中該單一選定照明路徑係響應於從該節段之一主要 ㈣及該節段資訊之-粗糙程度中選出之至少-個參數被 選疋，響應於該昭明姦斗主-U ^ …、產生表不缺陷—貝訊之檢測信號；以及 處理該檢測信號來檢測缺陷。 種藉由多重照明路徑系統來檢測缺陷之方法該 方法U列步驟·執行—物件資訊之一初始掃描，其中 在該初始掃描期間，多重照明路徑同步照明該物件；變库 =該初始掃描期間獲得的檢測信號來將該物件分節成節 貝& a應於自4節段之—主要材料與該節段資訊之一 表面粗糙程度中選出至少一個參數，每一節段選擇一選定 L來在°亥即段之一缺陷檢測掃描期間被致動；僅藉 20 200916763 由該物件資訊之一選定照明路徑來照明該物件之每一節 段；以及響應於該照明產生表示缺陷之檢測信號。 圖式簡單說明 本發明將由下列說明伴隨圖式而令人更全盤地瞭解： 5 第1圖繪示依據本發明之一實施例之缺陷以及表現這 類缺陷之理想灰階信號之範例； 第2A-2D圖繪示各種類型之表面拋光與反射模式； 第3 A - 3 C圖繪示依據本發明之各種實施例之各種類型 之照明與成像方式； 10 第3D圖繪示依據本發明之一實施例一缺陷之一影像及 其背景； 第4 A圖繪示依據本發明之一實施例之一照明與成像方 式； 第4B-4D圖繪示一缺陷之成像及其響應於苐4A圖不同 15 的照明所獲得之背景； 第5 A圖繪示依據本發明之一實施例之一照明及成像方 式； 第5B-5D圖繪示一缺陷之成像及其響應於第4A圖不同 的照明所獲得之背景； 20 第6A圖繪示依據本發明之一實施例之一照明及成像方 式； 第6B圖繪示一缺陷之成像及其響應於第4A圖不同的 照明所獲得之背景； 第7A圖依據本發明之一實施例之一照明及成像方式， 8 200916763 以及各種以不同角度獲得之成像； 第7B-7C圖繪示施以不同的照明所獲得之檢測信號； 第8A圖繪示依據本發明之一實施例之一多重照明路徑 系統； 5 第8B圖繪示依據本發明之一實施例之一多重照明路徑 糸統， 第8C圖繪示依據本發明之一實施例之一多重照明路徑 系統； 第9 A圖繪示依據本發明之一實施例之一多重照明路徑 10 系統； 第9 B - 9 D圖繪示依據本發明之一實施例之一可組配之 照明路徑； 第10圖繪示依據本發明之一實施例之一方法； 第11圖繪示依據本發明之一實施例之一方法； 15 第12圖繪示依據本發明之一實施例之一方法；以及 第13圖繪示依據本發明實施例之一種方法。 t實施方式3 較佳實施例之詳細說明 本發明提供一種多重照明路徑系統。其包括多重照明 20 路徑、一處理器、以及一成像路徑。處理器可適應於：⑴ 控制一物件之一初始掃描；（ii)將物件分成節段，響應於初 始掃描期間所獲之檢測信號；（i i i)判定每一節段之一表面粗 糙度；以及(iv)響應於從一節段之主要材料及節段之一表面 粗糙程度之至少一參數中之選擇，每一個別節段選擇一選 200916763 定照明路徑來在一節段之缺陷檢測程序期間被致動。多重 照明路徑係適應於在物件初始掃描期間連續照明，一選定 照明路徑在物件之一缺陷檢測掃描期間照明與選定照明路 徑相關聯之一節段。成像路徑係適應於響應照明產生指示 5 缺陷之檢測信號。 較佳地，處理器係適應於處理檢測信號以檢測缺陷。 較佳地，處理器係適應於響應於指示物件節段之物件 資訊來分段評估物件。 較佳地，系統係適應於：⑴以多重照明路徑一次一個 10 照明路徑地照明節段；其中不同的照明路徑適用不同的表 面粗糙程度；（ii)以成像路徑產生照明造成之檢測信號；以 及（iii)以處理器處理檢測信號並選擇一個用於節段之照明 路徑。 較佳地，處理器係適應於選擇一照明路徑，其照明一 15 導體(例如一金屬導體）表面以及一導體表面缺陷，使得不論 導體導體表面之粗糙程度，表面缺陷之一影像實質上為黑 色的，而導體表面之一影像實質上為白色的。 較佳地，處理器係適應於選擇一照明路徑，其照明一 半透明介電表面以及一半透明介電表面缺陷，使得不論半 20 透明表面之粗糙程度，半透明（介電）表面缺陷之影像實質上 為白色的，而半透明表面之影像實質上為黑色的。 較佳地，多重照明路徑包含一鏡射照明路徑，一散射 性傾斜通道，以及一近鏡射照明路徑。 較佳地，成像路徑包含一相機，一物件透鏡，以及一 Γ '1 L ^ j 10 200916763 遮罩；其中相機包含一感測表面，其平行於受評估物件之 一表面；物件透鏡係將物件成像於相機之感測區域，且其 中遮罩引用成像路徑之一非對稱角度測量範圍。 較佳地，多重照明路徑包括一鏡射照明路徑，一前傾 5 斜照明路徑，以及一後分散傾斜照明路徑。 較佳地，系統包括一可組配的照明路徑，其可具有兩 種組配中之一種；其中在第一中組配中，可組配的照明路 徑以以連續角度範圍照明物件；其中在第二種組配中，可 組配的照明路徑以連續角度範圍中之兩個彼此間隔開的次 10 範圍來照明物件。 較佳地，多重照明路徑包含一前傾斜照明路徑，一後 分散傾斜照明路徑，一偏軸鏡射照明路徑；其中成像路徑 包含一相機，其以實值上以異於至物件表面之正常角度成 像物件。 15 較佳地，處理器係適應於響應於一節段分類為一節段 選擇一選定照明路徑，譬如一光滑金屬表面，一半霧面金 屬表面，一霧面金屬表面或一半透明介電表面。 較佳地，照明路徑係適應於以一選定照明路徑來照明 另一物件之一節段，其與物件之一相應路徑相關聯，其中 20 其他物件較理想地相同於該物件，而其中該物件與其他物 件實值上受制於相同的製造條件。 較佳地，照明路徑係適應於以與物件之一相應節段相 關聯的一選定照明路徑來照明另一物件之一節段；其中物 件與其他物件係屬於相同批次之電路。 11 200916763 本發明提供一種用於缺陷檢測之一系統。此系統包括 一處理器，一成像路徑及多重照明路徑。系統係適應於僅 藉由與節段相關聯之一選定照明路徑來照明一物件之每一 節段；其中選定照明路徑係響應於從節段之一主要材料及 5 節段之表面粗糙程度中至少一參數來選定。成像路徑係適 應於響應於照明產生指示缺陷之檢測信號。處理器處理檢 測信號來檢測缺陷。 較佳地，多重照明路徑包含一鏡射照明路徑，一散射 性傾斜通道，以及一近鏡射照明路徑。 10 較佳地，成像路徑包含一相機及一遮罩；其中相機包 含平行所評估物件之一表面之一感測表面；以及其中遮罩 引用角度上非對稱之成像路徑集成範圍。 較佳地，多重照明路徑包含一鏡射照明路徑，一前傾 斜照明路徑，以及一後分散傾斜照明路徑。 15 較佳地，系統包括一可組配的照明路徑，其具有兩種 組配；其中第一組配中可組配的照明路徑以一連續角度範 圍照明物件；其中第二種組配中，可組配的照明路徑以連 續角度範圍中相間隔開之兩個次範圍來照明物件。 較佳地，多重照明路徑包含一前傾斜照明路徑，後分 20 散傾斜照明路徑，偏軸鏡射照明路徑；以及其中成像路徑 包含藉由一相機以實值上以異於至物件表面之正常角度的 一角度來成像物件。 較佳地，處理器響應於節段分類來選擇選定照明路 徑，該分類包含一光滑金屬表面、一半霧面金屬表面、一 12 200916763 霧面金屬表面、或一半透明介電表面。 應用不變影像外觀及一體式缺陷檢測演算法 為使得缺陷檢測處理應用穩定及標準化，並因而快速 且可靠，目標灰階影像需為下列極性：⑴一均勻金屬表面 5 呈「白色」，由高強度畫素提供；（ii)一半透明介電表面呈 「黑色」，由低強度畫素提供；（iii)金屬表面缺陷呈在「白 色」背景上的「黑色」一區域質地異常，由被高強度均勻 金屬質地環繞之低強度晝素提供；以及(iv)介電表面和次表 面缺陷呈在「黑色」背景上之「白色」。介電表面與次表面 10 異常由低強度均勻介電質地所環繞之高強度畫素提供。 在物件（亦稱作應用）中，包括兩種金屬與介電部（譬 如-BGA)，其影像會是一高動態範圍灰階影像，其具有高 強度、「白色」而具有「黑色」缺陷（14)的金晝素（13)、低 強度、「黑色」而具有「白色」缺陷（12)的之之焊接遮罩畫 15 素（11)，如第1圖所繪示。 第1圖亦繪示這些缺陷-焊接遮罩括痕17、導體三維 (3D)缺陷18、焊接遮罩空隙19、在焊接遮罩異質材料16之 下。 前述標準化方式支配預設的影像極性，其中金屬三維 20 (3D)缺陷呈「白色」上的「黑色」，反之亦然，表面與次表 面介電缺陷呈「黑色」上的「白色」。利用特殊設計的照明 配置以在3D缺陷顯示上產生被白色包圍之黑色畫素，以及 導致吸收效應。應用分散技術使介電缺陷呈黑色包圍的白 色晝素。 13 200916763 表面加工不變性 多重照明路徑系統具有一光學組配，其不論各式各樣 的表面加工均提供順應標準化程式的灰階影像。所有表 面’被疋義為將被系統檢測的選擇性物件，可以其反射屬 5性分類為下列幾組：⑴非常平坦光滑的不透明表面212(光 滑金屬）’幾乎以鏡射之方式反射入射光，如第2A圖所繪 示，且特別是以光源210和照明圖樣214所繪示者；（i〇梢粗 糙的半霧面不透明表面216(半霧面金屬），以方向性散射之 方式反射入射光，如第2B圖所繪示，特別是光源21〇及照明 圖樣218所繪不者；（〗„)粗糙的霧面不透明表面幻％霧面金 屬8)，以一完全散射（朗伯表面甚至是等向）方式反射入射 光，如第2C圖所繪示，特別是光源21〇及照明圖樣22〇所示 者，（iv)平滑介電透明與半透明層24〇(光滑絕緣體)組合表面 鏡射與容積漫反射，如第2D圖所繪示，特別光源21〇和 15圖樣230是所示者。 灰階影像係包括：⑴成像之角度位置及相對於所檢測 表面之㈣光學（_作成料妙錢賴）；⑼表面反 射屬性：㈣成像光學的數值角度與解析度；㈣照明光風 之角度覆蓋範圍；以及(v)照明光譜之函數。 予 20 其包括 較佳地，多重照明路#系統包括_成像路押， 相對於物件置於一固定角度位置之一單 限為一 CCD)。 ^ 1—不 在這樣的-系統中，-表面無變化的_ 照明角度彼此相異且適於各種表面加 由M、 J夕重照明路徑所。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The present invention relates to an automated optical inspection system. L Prior Art 3 Background of the Invention At the end of the production line, the HDI (High Density Interconnect) 10 substrate is polished to visually inspect surface defects and/or violate integrity before packaging and shipping. Such visual inspections will classify them as valid or invalid based on quality requirements. Surface defects are unsatisfactory product quality and/or unsatisfactory results. Here, all qualities are related to all surface types, ie. Metallized surfaces such as gold-plated interconnects on laminated substrates and solder masks. Each of these 15 types of surfaces is affected by specific types of defects, such as the products of their original materials and the manufacturing process. From an optical point of view, each material references a different optical characteristic and therefore needs to be addressed when it is detected (either manually or via an automated machine). Providing efficient defect detection methods and systems is an increasing demand. 20 SUMMARY OF THE INVENTION A system includes a processor, an imaging path, and a plurality of illumination paths; the processor selecting an illumination path from the plurality of illumination paths such that the selected illumination path illuminates a conductor surface and a conductor The surface lacks 200916763, so that regardless of the degree of the thick chain of the conductor conductor surface, the surface defect - shadow: the essence: black. And the surface of the conductor - the image is essentially white: ', where (four) selection - illumination path, its illumination - translucent dielectric surface and - translucent dielectric surface defects, making the thick chain of the translucent surface Why, the image of a translucent surface defect is substantially white' and the image of the translucent surface is substantially black. A method for detecting defects by a multiple illumination path system, the method comprising the steps of: selecting a selected illumination path between the defects detection during illumination-conductor surface to enable the selected illumination 1 When the surface of the illuminated lighting conductor and the surface of the conductor are defective, regardless of the degree of coarse sugar on the surface of the conductor conductor, one of the surface defects is substantially black, and the surface of the conductor surface is substantially white; when the illumination is half transparent surface temple Selecting a selected illumination path during the heart-detection scan is used to cause sputum, the degree of sag of the translucent surface, and the semi-transparent surface is substantially white, while the image of the translucent surface is substantially Black. , '. . a system for defect detection, the system comprising a processing benefit, an imaging path, and a plurality of illumination paths; wherein the system is adapted to illuminate an object by only selecting one of the illumination paths associated with the segment 〇 each segment, wherein the selected illumination path is selected in response to at least one of a surface roughness of the primary material and the segment information from the segment, wherein the imaging path is adapted to respond to the illumination generation A detection indicating a defect H in which the processor processes the detection signal to detect a defect. An illumination path system comprising a plurality of illumination paths, 200916763 processing, and imaging path information, wherein the processor is adapted to: control - object information - initial scan, responsive to obtaining during the initial scan information The detection signal is used to segment the object into segments; to determine the surface roughness of each section of the information, and to select the surface material from the primary material and the segment information At least one parameter, each-segment selection-selected illumination path to be actuated during one of the segment defect detection processes; wherein the multiple illumination paths are adapted to simultaneously illuminate the object during the initial scan information, and wherein A selected illumination path finds object information - the illumination during the defect detection scan is again related to the selected illumination path U, and wherein the imaging path is adapted to generate a detection signal indicative of the defect in response to the illumination. a method for defect detection, the method comprising: illuminating each segment of an object only by selecting a lighting path associated with a piece of information - wherein the single selected illumination path is responsive to one of the segments The main (4) and at least one of the parameters selected from the roughness information of the section are selected, in response to the detection of the main character of the smuggler-U ^ ..., the generation of the defect - the detection signal of Beixun; and processing the detection signal To detect defects. Method for detecting defects by multiple illumination path systems. The method U-steps-execution-one of the object information initial scans, wherein during the initial scan, multiple illumination paths simultaneously illuminate the object; variable library = obtained during the initial scan The detection signal is used to segment the object into scallops & a should select at least one parameter from the surface roughness of the main material and one of the segment information, and select a selected L for each segment. One of the segments is activated during the defect detection scan; only 20 200916763 selects an illumination path from one of the object information to illuminate each segment of the object; and generates a detection signal indicative of the defect in response to the illumination. BRIEF DESCRIPTION OF THE DRAWINGS The invention will be more fully understood from the following description: FIG. 1 is a diagram showing an example of an embodiment of the invention and an example of an ideal gray scale signal for representing such a defect; -2D diagram showing various types of surface polishing and reflection modes; Figures 3A - 3C illustrate various types of illumination and imaging modes in accordance with various embodiments of the present invention; 10 Figure 3D depicts one of the aspects of the present invention An image of one of the defects of the first embodiment and its background; FIG. 4A illustrates an illumination and imaging mode according to an embodiment of the present invention; and FIG. 4B-4D illustrates an image of a defect and its response to the image of the 苐4A Background obtained by illumination of 15; FIG. 5A illustrates an illumination and imaging mode according to an embodiment of the present invention; FIG. 5B-5D illustrates imaging of a defect and a different illumination device in response to FIG. 4A Background obtained; 20 FIG. 6A illustrates an illumination and imaging mode according to an embodiment of the present invention; FIG. 6B illustrates an image of a defect and a background obtained by different illuminations in accordance with FIG. 4A; Figure according to one of the inventions One illumination and imaging method of the embodiment, 8 200916763 and various images obtained at different angles; 7B-7C depict detection signals obtained by applying different illuminations; FIG. 8A illustrates an embodiment according to the present invention One multiple illumination path system; 5 FIG. 8B illustrates a multiple illumination path system in accordance with an embodiment of the present invention, and FIG. 8C illustrates a multiple illumination path system in accordance with an embodiment of the present invention; FIG. 9A illustrates a multiple illumination path 10 system according to an embodiment of the present invention; and FIG. 9B - 9D illustrates an illumination path that can be assembled according to an embodiment of the present invention; FIG. A method according to an embodiment of the invention is shown; FIG. 11 illustrates a method in accordance with an embodiment of the present invention; 15 FIG. 12 illustrates a method in accordance with an embodiment of the present invention; and FIG. A method in accordance with an embodiment of the invention is illustrated. t. Embodiment 3 DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention provides a multiple illumination path system. It includes multiple illuminations 20 paths, a processor, and an imaging path. The processor can be adapted to: (1) control an initial scan of an object; (ii) divide the object into segments, in response to a detection signal obtained during the initial scan; (iii) determine a surface roughness of each segment; and ( Iv) responsive to selection of at least one of the primary material of the segment and the degree of surface roughness of one of the segments, each individual segment is selected to select the 200916763 illumination path to be actuated during the defect detection procedure of the segment . The multiple illumination paths are adapted to continuously illuminate during an initial scan of the object, and a selected illumination path illuminates one of the segments associated with the selected illumination path during one of the defect detection scans of the object. The imaging path is adapted to produce a detection signal indicative of a defect in response to illumination. Preferably, the processor is adapted to process the detection signal to detect defects. Preferably, the processor is adapted to segment the evaluation of the item in response to the item information indicative of the item segment. Preferably, the system is adapted to: (1) illuminate the segments one at a time with multiple illumination paths; wherein different illumination paths are adapted to different surface roughness; (ii) detection signals resulting from illumination of the imaging path; (iii) processing the detection signal with the processor and selecting an illumination path for the segment. Preferably, the processor is adapted to select an illumination path that illuminates a surface of a 15 conductor (eg, a metal conductor) and a surface defect of the conductor such that one of the surface defects is substantially black regardless of the roughness of the surface of the conductor conductor. One of the conductor surfaces is substantially white. Preferably, the processor is adapted to select an illumination path that illuminates half of the transparent dielectric surface and half of the transparent dielectric surface defects such that the image of the semi-transparent (dielectric) surface defect is substantial regardless of the roughness of the semi-transparent surface. The upper is white, and the image of the translucent surface is substantially black. Preferably, the multiple illumination paths comprise a mirrored illumination path, a diffuse oblique channel, and a near-mirror illumination path. Preferably, the imaging path comprises a camera, an object lens, and a '1 L ^ j 10 200916763 mask; wherein the camera comprises a sensing surface parallel to one surface of the object being evaluated; the object lens is an object Imaging the sensing area of the camera, and wherein the mask references an asymmetric angle measurement range of one of the imaging paths. Preferably, the multiple illumination paths include a mirrored illumination path, a forward tilted 5 oblique illumination path, and a rear dispersed oblique illumination path. Preferably, the system comprises an configurable illumination path, which may have one of two combinations; wherein in the first medium assembly, the illumination path can be configured to illuminate the object in a continuous angular range; In the second assembly, the configurable illumination path illuminates the object in a sub-10 range that is spaced apart from each other by a continuous range of angles. Preferably, the multiple illumination paths comprise a front oblique illumination path, a post-dispersion oblique illumination path, and an off-axis mirror illumination path; wherein the imaging path comprises a camera that is different in value from the normal angle to the surface of the object Imaging objects. Preferably, the processor is adapted to select a selected illumination path for a segment in response to a segment classification, such as a smooth metal surface, a half matte metal surface, a matte metal surface or a semi-transparent dielectric surface. Preferably, the illumination path is adapted to illuminate a segment of another object with a selected illumination path associated with a respective path of one of the objects, wherein 20 other items are desirably identical to the object, wherein the object is Other objects are subject to the same manufacturing conditions in real value. Preferably, the illumination path is adapted to illuminate a segment of another object with a selected illumination path associated with a respective segment of the object; wherein the object and other objects belong to the same batch of circuitry. 11 200916763 The present invention provides a system for defect detection. The system includes a processor, an imaging path and multiple illumination paths. The system is adapted to illuminate each segment of an object by only selecting one of the illumination paths associated with the segment; wherein the selected illumination path is responsive to at least one of the primary material of the segment and the surface roughness of the five segments A parameter is selected. The imaging path is adapted to generate a detection signal indicative of a defect in response to illumination. The processor processes the detection signal to detect defects. Preferably, the multiple illumination paths comprise a mirrored illumination path, a diffuse oblique channel, and a near-mirror illumination path. Preferably, the imaging path comprises a camera and a mask; wherein the camera comprises a sensing surface parallel to one of the surfaces of the evaluated object; and wherein the mask is angularly asymmetric with respect to the imaging path integration range. Preferably, the multiple illumination paths include a mirrored illumination path, a front tilted illumination path, and a rear dispersed oblique illumination path. Preferably, the system comprises an configurable illumination path having two combinations; wherein the first group of configurable illumination paths illuminate the object in a continuous angular range; wherein the second combination, The configurable illumination path illuminates the object in two sub-ranges spaced apart in a continuous angular range. Preferably, the multiple illumination paths comprise a front oblique illumination path, a back-split oblique illumination path, and a off-axis mirror illumination path; and wherein the imaging path comprises a normal value on the surface of the object by a camera An angle of the angle to image the object. Preferably, the processor selects a selected illumination path in response to the segment classification, the classification comprising a smooth metal surface, a half matte metal surface, a 12 200916763 matte metal surface, or a semi-transparent dielectric surface. Applying the constant image appearance and integrated defect detection algorithm to make the defect detection processing application stable and standardized, and thus fast and reliable, the target gray scale image needs to have the following polarities: (1) a uniform metal surface 5 is "white", high The intensity pixel is provided; (ii) the semi-transparent dielectric surface is "black" and is provided by a low-intensity pixel; (iii) the metal surface defect is a "black" region on the "white" background. The low-strength elemental material surrounded by a uniform metallic texture; and (iv) the dielectric surface and subsurface defects are "white" on a "black" background. The dielectric surface and subsurface 10 anomalies are provided by high intensity pixels surrounded by a low intensity uniform dielectric texture. In objects (also known as applications), including two metals and dielectrics (such as -BGA), the image will be a high dynamic range grayscale image with high intensity, "white" and "black" defects. (14) The ruthenium (13), low-intensity, "black" and "white" defect (12) of the welding mask 15 (11), as shown in Figure 1. Figure 1 also depicts these defects - the solder mask footprint 17, the conductor three-dimensional (3D) defect 18, the solder mask void 19, under the solder mask heterogeneous material 16. The aforementioned normalization method governs the preset image polarity, wherein the metal three-dimensional 20 (3D) defect is "black" on the "white", and vice versa, and the surface and sub-surface dielectric defects are "white" on the "black". A specially designed lighting configuration is used to create a black pixel surrounded by white on the 3D defect display and cause an absorption effect. The use of dispersion technology to make the dielectric defects black surrounded by white halogen. 13 200916763 Surface Machining Invariance The Multiple Illumination Path System has an optical assembly that provides a grayscale image that conforms to a standardized program, regardless of the variety of surface finishes. All surfaces 'deprecated as selective objects to be detected by the system can be classified into the following groups by their reflection: 5 (1) Very flat and smooth opaque surface 212 (smooth metal) 'reflects incident light almost in a specular manner , as shown in FIG. 2A , and particularly as shown by the light source 210 and the illumination pattern 214 ; (i) the rough semi-matte opaque surface 216 (half-matte metal), reflected by directional scattering The incident light, as shown in Fig. 2B, especially the light source 21〇 and the illumination pattern 218 are not drawn; (〗 〖) rough matte opaque surface illusion fog metal 3), with a complete scattering (Lambert The surface reflects the incident light even in an isotropic manner, as shown in Fig. 2C, particularly as shown by the source 21〇 and the illumination pattern 22〇, (iv) smooth dielectric transparent and translucent layer 24〇 (smooth insulator) Combined surface mirroring and volumetric diffuse reflection, as shown in Fig. 2D, the special light source 21〇 and 15 pattern 230 are shown. The grayscale image system includes: (1) the angular position of the image and the (4) optics relative to the detected surface. (_作成妙妙钱赖); (9) Surface reflection properties: (iv) numerical angle and resolution of imaging optics; (iv) angular coverage of illumination light; and (v) function of illumination spectrum. 20 Preferably, the multiple illumination road system includes _ imaging road, One of the fixed angle positions relative to the object is limited to a CCD). ^ 1 - not in such a system, - the surface has no change _ illumination angles are different from each other and are suitable for various surfaces plus by M, J Heavy lighting path
14 200916763 達成。 利用一單一相機大幅地減少系統成本。一多重照明路 徑與單一相機系統比多重相機系統便宜許多。 利用一不對稱照明及/或成像路徑 5 為了增強光滑金屬表面上3D缺陷視覺化之陰影與吸收 效應，利用對稱破缺原理。精心設計的具有内建非對稱之 光學組配，其增強相對均勻表面之缺陷而不產生不合意的 傾斜效應。唯有具感生非對稱之鏡射光學組配會提供目標 灰階影像及增強光滑金屬表面之缺陷。此方式如下列光學 10 組配所繪示： 許多可提供位在一白色鄰近區中之一黑色缺陷之一影 像（第3D圖）的光學組配如第3A、3B、3C圖所示。 第3 A圖繪示在一鏡射角度位置之一傾斜的成像與照明 光學。其由光源210、照明圖樣302、和相機310所示。 15 第3B圖繪示具軸上鏡射照明之一正交成像光學，其中 照明光學產生一非對稱（較佳地利用一遮罩）。此由光源 210、不對稱照明圖樣312、光束分離器320、與相機310所 示。 第3C圖繪示具軸上鏡射照明之正交成像光學。遮罩330 20 之一不對稱成像角度產生非對稱。此由光源210、不對稱成 像圖樣332、光束分離器320、與相機310所示。 第3D圖繪示以第3A-3C圖之各組配獲得之一合意影 像。合意的影像包括被一白色背景（畫素340)包圍的一黑色 缺陷（畫素350)。 15 200916763 針對半霧面與霧面金屬表面之缺陷增強照明：較佳與 禁止角度區 藉由固定成像光學位置（傾斜或正交），缺陷極性（白色 上的黑色或黑色上的白色）係表面加工（粗糙程度）和照明角 5度的函數。缺陷在照明角度變化下改變他們的極性。 第4A圖繪示一半霧面表面4〇〇之一3D表面缺陷4〇2、_ 不規則光滑質地404、與一均勻半霧面質地406。不規則光 滑質地404係由包括一 CCD相機和一透鏡（統稱31〇)之—成 像路徑所成像。有許多照明路徑，在相機光學軸之兩側為 10 物件表面之法線。 這些照明路徑包括：傾斜散射照明路徑41〇和420、近 鏡射散射照明路徑414及424、以及不合意的照明路徑（亦稱 作禁止角度區）412及422。第4B，4C及4D圖繪示各種照明 半霧面表面400所和得之成像。 15 半霧面表面上之3D缺陷和質地誤差在近鏡射散射照明 (如路徑414及424繪示者）下呈白色背景上的黑色畫素(430 及432)。在低傾斜照明（路徑410及420所繪示）下相同缺陷改 變成其相反的極性：黑色背景（444)上的白色畫素(434及 435)。存在有中介角度區(412及422)，其中相同的缺陷與其 20周圍呈不可分辨的狀態，「灰色」在「灰色」上，如第4C 圖（灰畫素442)所示。 第5A圖繪示—霧面表面500之一 3D表面缺陷502、一不 規則光滑質地504、與一均勻半霧面質地5〇6。不規則光滑 質地504係以包括一CCD相機和一透鏡（統稱310)之一成像 16 L 3 200916763 路徑所成像。有許多照明路徑，相機光學軸兩側者為物件 表面之法線。 這些照明路徑包括：傾斜散射照明路徑510及520、近 鏡射散射照明路徑514及524、及不合意的照明路徑（亦稱作 5 禁止角度區）512及522。第5B、5C、5D圖繪示各種照明半霧 面表面400所獲得之成像。 低傾斜照明（520及510)提供「右」缺陷（第5D圖）；近鏡 射照明（514及524)導致缺陷顏色之模糊（第5B圖），而中介角 度照明（512及524)表示「全灰」，如第5C圖所示。 10 近鏡射角度區較佳係針對半霧面表面。霧面表面應以 低度傾斜照明。中介角度區被定義為對所有金屬加工禁 止。提供相對極性之照明組合亦被禁止，因其導致缺陷不 被看見。 光滑半透明介電層之缺陷增強照明。 15 此類應用之目標影像以在「黑色」背景上之「白色」 缺陷晝素呈現。最佳解析度符合前述需求為低傾斜照明。 其提供「黑色」一致的均勻光滑表面影像，無反射。表面 及次表面缺陷，表現強列的分散中心，呈「白色」，如第6 圖所示。 20 第6圖繪示在一黑色背景（第6B圖之晝素646)上之白色 缺陷（晝素642及644)。第6A圖繪示由照明一焊接遮罩表面 造成之光圖樣610該表面包括成像處理中增強的一次表面 缺陷。 角度覆蓋範圍及表面缺陷增強 17 200916763 影像呈現上角度覆蓋範圍的影響由第7A-7C圖示範。越 寬的角度覆蓋範圍產生越光滑的影像呈現，如第7B圖之圖 表740所繪示（特別相較於第7C圖之750)。平滑效應包括信 號之一光學平均，信號係從所繪示之角度範圍中獲得，包 5 括以方向在角度α 1之光所獲之影像73卜以角度α2之光獲 得之影像732、及從這些影像所獲得的一平均影像733。CCD 相機310獲得經平均之影像。 越強的表面對比誤差，需要越寬的角度覆蓋範圍來達 到一平滑效應。為了增強表面缺陷，必須抑制均勻質地雜 10 訊（背景平滑）及增加表面不均勻之區域對比。角度覆蓋範 圍，針對一特定應用填補較佳角度區，保存所需的影像極 性，以及抑制背景雜訊。這使得表面檢測更穩定，具有最 小的故障危機。 表面缺陷檢測之最佳光學組配 15 為提供應用無變化的灰階影像，發展出多角度照明光 學組配。 依據本發明諸多實施例之正交組配呈現於第8Α及8Β 圖。成像光學，其包括一CCD相機和物件透鏡(統稱310)， 正交檢視受測表面（CCD相機平行物件使得CCD相機之光 20 學軸正交於物件）。 動態成像遮罩角度810物件透鏡之集合角度811中插入 非對稱。照明光學由個獨立的照明通道表現： 包括光源310、遮罩820、及光束分離器830之軸上鏡射 照明路徑，可用於光滑金屬表面上之缺陷檢測。其亦包括 18 200916763 打破對稱遮罩角度來改變角度覆蓋範圍，從全[_8。…+8。] 至半[0 .“+8 ]或[-8。…0。]。前者如第8B圖之照明圖樣831 所繪示。 近鏡射對稱照明路徑850及860用於半霧面金屬表面上 5之缺陷檢測。相應角度覆蓋範圍落在[-32。…-8。]&[8。…32。] 之角度區内，如第8B圖之照明圖樣861及851所繪示 低傾斜對稱照明路徑84〇及870用於霧面金屬表面上及 半透明光滑介電層上、中、下之缺陷檢測。相應角度覆蓋 範圍落在[-63。…-47。]&[47。…63。]之角度區内，如第8C圖 10 之照明圖樣841及871所繪示。 依據特定規格（尤其是每一節段為何種主要材料及其 相關的表面粗糙度)選擇合適的照明通道。任何不同照明通 道之組合是被禁止的。 系統系由處理器801所控制，其可施用方法1 〇〇〇、 15 1丨〇〇、或1200以外之任何方法。 傾斜組配呈現於第8C圖。成像光學，其包括一CCD相 機310及物件透鏡890 ’其以小的傾斜角度_16。檢視受測表 面。亦可接受其他在[5。·..25。]之範圍内的檢視角度。照明 光學係由三個獨立的照明通道呈現： 20 一可組配的照明路徑可被組配來針對光滑與半霧面金 屬表面上之缺陷檢測提供偏軸寬鏡射和近鏡射照明路徑， 位於相對於表面法線16。之鏡射角度。此通道包括在兩工作 位置間進行切換的角度操縱光學元件： 用於光滑表面之第一部份（亦稱作〇 -位置）(繪示以連續14 200916763 Completed. Significantly reduce system cost with a single camera. A multiple illumination path is much less expensive than a single camera system than a multiple camera system. Utilizing an Asymmetric Illumination and/or Imaging Path 5 To enhance the shadowing and absorption effects of 3D defect visualization on smooth metal surfaces, the principle of symmetry breaking is utilized. A well-designed optical assembly with built-in asymmetry that enhances defects on relatively uniform surfaces without undesired tilting effects. Only the specular optics with a sense of asymmetry will provide the target grayscale image and enhance the defects of the smooth metal surface. This mode is illustrated by the following optical composition: Many optical combinations that provide an image of one of the black defects (Fig. 3D) in a white neighboring region are shown in Figures 3A, 3B, and 3C. Figure 3A shows the imaging and illumination optics tilted at one of the mirrored angular positions. It is illustrated by light source 210, illumination pattern 302, and camera 310. 15 Figure 3B illustrates one of the orthogonal imaging optics with on-axis mirror illumination, wherein the illumination optics produce an asymmetry (preferably using a mask). This is illustrated by light source 210, asymmetric illumination pattern 312, beam splitter 320, and camera 310. Figure 3C depicts orthogonal imaging optics with on-axis mirror illumination. One of the asymmetric imaging angles of the mask 330 20 produces an asymmetry. This is illustrated by light source 210, asymmetric imaging pattern 332, beam splitter 320, and camera 310. Fig. 3D shows a desirable image obtained by each of the combinations of the 3A-3C. A desirable image includes a black defect (pixel 350) surrounded by a white background (pixel 340). 15 200916763 Enhanced illumination for defects on semi-matte and matte metal surfaces: preferred and prohibited angular regions by fixed imaging optical position (tilted or orthogonal), defect polarity (black on black or white on black) surface Processing (roughness) and a function of the illumination angle of 5 degrees. Defects change their polarity as the illumination angle changes. Figure 4A depicts one of the half matte surface 4's 3D surface defects 4〇2, _ irregular smooth texture 404, and a uniform semi-matte texture 406. Irregular smooth texture 404 is imaged by an imaging path comprising a CCD camera and a lens (collectively 31 〇). There are many illumination paths on the sides of the camera's optical axis that are the normal to the surface of the 10 object. These illumination paths include obliquely scattered illumination paths 41A and 420, near-mirror scattered illumination paths 414 and 424, and undesired illumination paths (also referred to as forbidden angle regions) 412 and 422. Figures 4B, 4C and 4D illustrate the imaging of various illumination semi-matte surfaces 400. 15 3D defects and texture errors on the semi-matte surface are black pixels (430 and 432) on a white background under near-mirror scattering illumination (as depicted by paths 414 and 424). The same defect is changed to its opposite polarity in low tilt illumination (illustrated by paths 410 and 420): white pixels (434 and 435) on a black background (444). There are intermediate angle regions (412 and 422) in which the same defect is indistinguishable from the periphery of 20, and "gray" is on "gray" as shown in Fig. 4C (grey pixel 442). Figure 5A illustrates a 3D surface defect 502, an irregular smooth texture 504, and a uniform semi-matte texture 〇6. The irregular smooth texture 504 is imaged by imaging 16 L 3 200916763 path including one of a CCD camera and a lens (collectively referred to as 310). There are many illumination paths, and the sides of the camera's optical axis are the normals on the surface of the object. These illumination paths include obliquely scattered illumination paths 510 and 520, near-mirror scattered illumination paths 514 and 524, and undesired illumination paths (also referred to as 5 forbidden angle regions) 512 and 522. Figures 5B, 5C, and 5D illustrate images obtained with various illumination semi-matte surfaces 400. Low-tilt illumination (520 and 510) provides "right" defects (Fig. 5D); near-mirror illumination (514 and 524) causes blurring of defect colors (Fig. 5B), while intermediate angle illumination (512 and 524) means " All gray, as shown in Figure 5C. 10 The near-mirror angle region is preferably directed to a semi-matte surface. The matte surface should be tilted with low illumination. The intermediate angle zone is defined as a ban on all metal processing. Lighting combinations that provide relative polarity are also prohibited as they result in defects not being seen. Defects in the smooth translucent dielectric layer enhance illumination. 15 The target image for such applications is presented as a "white" defect in the "black" background. The best resolution meets the aforementioned requirements for low tilt illumination. It provides a "black" uniform uniform smooth surface image with no reflection. Surface and subsurface defects, showing the center of dispersion of the strong column, are "white", as shown in Figure 6. 20 Figure 6 shows the white defects (Alizarins 642 and 644) on a black background (Fig. 6B). Figure 6A illustrates a light pattern 610 caused by illuminating a solder mask surface that includes an enhanced primary surface defect in the imaging process. Angle Coverage and Surface Defect Enhancement 17 200916763 The effect of angular coverage on image presentation is demonstrated by Figures 7A-7C. The wider angular coverage results in a smoother image presentation, as depicted by graph 740 of Figure 7B (especially compared to 750 of Figure 7C). The smoothing effect includes one of the optical averaging of the signal, the signal is obtained from the range of angles depicted, and the image 732 obtained by the light obtained by the light at the angle α 1 is obtained by the light of the angle α2, and An average image 733 obtained from these images. The CCD camera 310 obtains an averaged image. The stronger the surface contrast error, the wider the angular coverage is needed to achieve a smoothing effect. In order to enhance the surface defects, it is necessary to suppress uniform texture (background smoothing) and increase the contrast of the surface unevenness. Angle coverage covers a better angle area for a specific application, preserves the desired image polarity, and suppresses background noise. This makes surface inspection more stable and has the smallest failure crisis. The best optical combination for surface defect detection 15 To provide a gray-scale image with no change in application, a multi-angle illumination optical combination has been developed. The orthogonal combinations according to various embodiments of the present invention are presented in Figures 8 and 8A. Imaging optics, which includes a CCD camera and object lens (collectively 310), and a quadrature view of the surface under test (the parallel object of the CCD camera is such that the CCD camera's light axis is orthogonal to the object). The dynamic imaging mask angle 810 is inserted into the set angle 811 of the object lens to be asymmetric. Illumination optics are represented by separate illumination channels: an on-axis mirror illumination path including source 310, mask 820, and beam splitter 830 for defect detection on smooth metal surfaces. It also includes 18 200916763 breaking the symmetrical mask angle to change the angular coverage from full [_8. ...+8. ] to half [0."+8] or [-8....0.]. The former is shown as illumination pattern 831 of Figure 8B. Near-mirror symmetric illumination paths 850 and 860 are used on semi-matte metal surfaces. 5 Defect detection: The corresponding angular coverage falls within the angle range of [-32....-8.]&[8....32.], as shown in Fig. 8B illumination patterns 861 and 851 show low tilt symmetry Illumination paths 84〇 and 870 are used for the detection of defects on the matte metal surface and on the translucent smooth dielectric layer. The corresponding angular coverage falls in [-63....-47.]&[47... 63.] The angled area is shown as illuminating patterns 841 and 871 of Figure 8C Figure 8. Select the appropriate lighting channel according to the specific specifications (especially the main material and its associated surface roughness for each segment). Any combination of different illumination channels is prohibited. The system is controlled by processor 801, which can apply any method other than Method 1, 丨〇〇, 15 1丨〇〇, or 1200. Tilting is presented in Figure 8C Imaging optics, which includes a CCD camera 310 and an object lens 890' with a small tilt angle _16. Other viewing angles in the range [5....25.] are also acceptable. The illumination optics are presented by three separate illumination channels: 20 A configurable illumination path can be combined to smooth Defect detection on a semi-matte metal surface provides an off-axis wide mirror and near-mirror illumination path at a mirror angle relative to the surface normal 16. This channel includes angled optics that switch between the two working positions Component: The first part of the smooth surface (also known as the 〇-position) (illustrated as continuous
19 200916763 角度範圍892)、或鏡射位置，其中操縱元件關閉，且不隱 藏在[0°···32°]間之角度範圍任何全角度覆蓋範圍。 5 10 用於半霧面表面之第二位置（亦稱作丨_位置）或近鏡射 位置，操縱元件打該，並排除[1(r_22。]之角度覆蓋範圍的 中央、純鏡射部份。主動角度覆蓋範圍填滿[〇。...1〇。]&[22 。…32。]之近鏡射角度區。角度區域892的這兩個次角度區 域標為893及894及並以區895間隔開。 前傾斜照明針對霧面金屬表面及半透明光滑介電層上 之缺陷檢測。照明路徑896之相應角度覆蓋範圍包括[8 32°]之角度區。角度範圍[_32。 光學路徑891中故因而未示。 -8 ]之另一（對稱）區包括在 背散傾斜照明針對半透明光滑介電層上、中、下之缺 陷檢測。相應角度«範圍填滿[u。]之角度區。其 如光圖樣891所繪示。 15 W田叼，尽乃逋道係依據特定規格 20 π 进疋：⑴偏釉鏡射 照明，0-位置-針對光滑金屬 芝屬’（11)偏軸鏡射照明，1_位置一 針對半務面金屬S，㈣讀斜照明—針對霧面金屬s ;㈣ 前傾斜與背散傾斜財透明光滑介電層。 較佳地’在缺陷檢測期Η ’間，任何(a)、（b)、（C)、或(d) 之組合是禁止的。 CSP及其他HDI應用 ，其已被設計來針對19 200916763 Angle range 892), or mirror position where the steering element is closed and does not hide any angular coverage of the angular range between [0°···32°]. 5 10 For the second position (also known as 丨_position) or near-mirror position of the semi-matte surface, the manipulating element hits this and excludes the center of the angle coverage of [1 (r_22.], pure mirror part The active angular coverage is filled with the near-mirror angle range of [〇....1〇.]&[22....32.] The two sub-angle regions of the angle region 892 are marked 893 and 894 and And spaced apart by zone 895. The front tilt illumination is for the detection of defects on the matte metal surface and the translucent smooth dielectric layer. The corresponding angular coverage of illumination path 896 includes an angular region of [8 32°]. Angle range [_32. The optical path 891 is thus not shown. The other (symmetric) region of -8] includes the detection of defects on the transmissive oblique illumination for the translucent smooth dielectric layer, the middle and the lower. The corresponding angle «range is filled [u. The angle of the area is as shown by the light pattern 891. 15 W Tian Hao, the 逋 逋 is based on a specific specification 20 π advance: (1) glaze mirror illumination, 0-position - for smooth metal genus' ( 11) off-axis mirror illumination, 1_position one for semi-surface metal S, (iv) read oblique illumination - for matte metal s (d) Front tilt and backscattered transparent transparent dielectric layers. Preferably, any combination of (a), (b), (C), or (d) is prohibited during the defect detection period Η. CSP and other HDI applications that have been designed to target
L 範例線掃描照明裝置針對pbga、 上之自動表面缺陷檢測。 第9A圖繪示傾斜線掃描照明裝置 各種HDI應狀自練面缺陷檢測。 20 200916763 所呈現光學系統包括：⑴傾斜16°之成像光學，其包含 線性CCD陣列908及物件透鏡907 ; (ii)三個線性光源9〇1、 902、及903(可為線性光纖或線狀LED陣列）。光源係垂直於 掃描方向。 5 每一線性光源具有以平行方式配置的集中圓柱光學， 譬如圓柱透鏡對。各集中圓柱對被設計來產生：⑴組配規 定的角度覆蓋範圍；（ii)適於光線寬路的應用。 三個集中的光線均同時發生在所檢查的表面上。 具圓柱對904之線性光源901形成自表面法線傾斜16。 1〇的一鏡射光通道914。圓柱對4提供之總角度覆蓋範圍為士16 。。第9b圖之旋轉線性角度912產生光角度操場效應。 具圓柱對905之線性光源902形成自表面法線傾斜55。 之一前傾斜光通道925。圓柱對5提供的總角度覆蓋範圍為土 8、 15 具圓柱對90 6之線性光源9 〇 3形成自表面法線傾斜_ 4 4。 之一後分散傾斜光通道930。圓柱對6提供的總角度覆蓋範 圍為±16 °。 在〇-位置上具旋轉角度角度912(第9b圖的）的鏡射光通 道914(第9a圖的）表示光滑金照明路徑。 2〇 在丨―位置上具旋轉角度角度912(第9b圖的）之鏡射光通 道914(第9a圖的）表示半霧面金照明路徑。 前傾斜光通道925表示霧面金照明路徑。 後分散傾斜光通道936及前傾斜光通道925表示焊接_ 遮罩照明路徑。 21 200916763 —選定光學路徑之選擇 為提供與本發明多角度裝置相關聯之自動表面識別之 應用^變影像形成方法。表面識別演算法如第1〇圖所示。 第10圖之方法1000始於階段1010，以全照明（所有照明 5路徑被同步致用）掃描一物件。 、階段1010接下來是階段1020,產生-物件節段,如定 義金節段和烊接遮罩節段。分節動作係響應於一節段中各 材料的相對比重。可依其尺寸與形狀不同來分節。 可找到許多節段。 10 針對每一焊接遮罩節段，階段1020接著是階段1025， 其選擇一照明路徑來照明一焊接遮罩。 針對每-金節段，階段刪接著是階段刪其儲存 形成節段之金畫素之金節段位置。 階段麵接著是階段聰，其選擇從_之—索^， 15如同有(在此範例中）4個可適於不同表面粗糖程度的不同照 明路控。 針對索引1的各值執行階段1040 ' 1〇45、及忉刈。 階段1 _包括掃描具細日轉徑之節段_。 階段1040接著是階段1〇45，其為節段判定一金加工參 2〇數。這是-組參數，譬如{金標準誤差(std);影像全域對 比卜其中金STD係金表面信號之標準誤差；影像全域對比 被定義為平均金信號和平均焊接信號之比） 階段麵接著是階段娜，其利用準則來分析所計算 的金加工參數。前述定義，運用的準則包括兩個同時執行L Example line scanning illuminator for automatic surface defect detection on pbga. Figure 9A shows the oblique line scanning illumination device. Various HDI-like self-leveling surface defect detection. 20 200916763 The presented optical system comprises: (1) imaging optics tilted at 16°, comprising a linear CCD array 908 and an object lens 907; (ii) three linear light sources 9〇1, 902, and 903 (which may be linear optical fibers or linear LED array). The light source is perpendicular to the scanning direction. 5 Each linear source has a concentrated cylindrical optics arranged in a parallel manner, such as a cylindrical lens pair. Each concentrated pair of cylinders is designed to produce: (1) a defined angular coverage; (ii) an application suitable for wide light paths. The three concentrated rays occur simultaneously on the surface being inspected. A linear light source 901 having a cylindrical pair 904 is formed to be inclined 16 from the surface normal. A mirrored light path 914. The total angular coverage provided by the cylinder pair 4 is ±16. . The rotational linear angle 912 of Figure 9b produces a light angle playground effect. A linear light source 902 having a cylindrical pair 905 is formed from a surface normal tilt 55. One of the front tilted light channels 925. The total angular coverage provided by the cylinder pair 5 is soil 8, 15 with a cylindrical pair 90 6 linear light source 9 〇 3 formed from the surface normal inclination _ 4 4 . One of the rear dispersed optical channels 930. The total angular coverage provided by the cylinder pair 6 is ±16 °. A mirrored light path 914 (Fig. 9a) having a rotational angle 912 (Fig. 9b) at the 〇-position represents a smooth gold illumination path. 2 Mirror light path 914 (Fig. 9a) having a rotation angle 912 (Fig. 9b) at the 丨 position indicates a semi-matte gold illumination path. The front oblique light channel 925 represents a matte gold illumination path. The rear dispersed oblique light channel 936 and the front oblique light channel 925 represent the soldering-mask illumination path. 21 200916763 - Selection of Selected Optical Paths To provide an adaptive image forming method for automatic surface recognition associated with the multi-angle apparatus of the present invention. The surface recognition algorithm is shown in Figure 1. The method 1000 of Figure 10 begins at stage 1010 by scanning an object with full illumination (all illumination 5 paths are used synchronously). Stage 1010 is followed by stage 1020, which produces an object segment, such as a defined gold segment and a spliced mask segment. The segmentation action is responsive to the relative specific gravity of each material in a section. It can be divided according to its size and shape. Many segments can be found. 10 For each weld mask segment, stage 1020 is followed by stage 1025, which selects an illumination path to illuminate a weld mask. For each-gold segment, the phase is deleted by the stage to delete the gold segment of the gold pixel that forms the segment. The stage is followed by Stage Cong, which chooses from _ _ _ ^, 15 as there are (in this example) 4 different lighting roads that can be adapted to the degree of coarse sugar on different surfaces. Stages 1040 '1〇45, and 忉刈 are executed for each value of index 1. Stage 1 _ includes scanning the section with a fine day turn. Stage 1040 is followed by stage 1〇45, which determines a gold processing parameter for the segment. This is a group parameter, such as {gold standard error (std); image global comparison, the standard error of the gold STD gold surface signal; image global comparison is defined as the ratio of the average gold signal to the average welding signal) Phase Na, which uses criteria to analyze the calculated gold processing parameters. The aforementioned definition, the criteria used include two simultaneous executions
22 200916763 的條件：最小金標準誤差和最大全域對比。 前述程序完成後’方法1000進行至找尋最適合照明路 徑（找最佳金加工準則照明通道_標為i0)之階段丨060。 階段10 6 0接著是在缺陷檢測程序期間選擇此照明路徑 5來照明此節段和相應節段(選擇i〇為最適合用於一特定應用 之照明通道）。 第11圖繪示1100用來以一多重照明路徑系統檢測缺陷 之方法，依據本發明之一實施例。 方法1100始於包括階段1110-1140之一測定程序，期間 10選定每一節段之照明路徑。 广理應瞭解的是，測定處理—完成，每—節段之照明路 征選擇可在一或多個理想的相同物件之缺陷檢測掃描期間 被使用。 譬如’ 一批電氣物件可在一特定電路上藉由執行測定 15 2〇 處理魏查，_在檢錄次之其他電路時彻照明路徑 (每一節段)選擇。 因此’多次巡迴方法㈣之各階段以㈣—⑽件中 之缺陷1些多次巡迴係、優先的，藉著接收 同物件(在理想條件下的物件被為相 於測定處理之結果）。物件和其他物件^_檢測得益 造條件。因此他們的表面㈣ 貝上”於相同製 徵。 貫貝上相同袓糙程度之特 此接收動作接下來是測定處理。；則定處理接著是應用 r t», I ^ 23 200916763 在其他全部（或多重）物件之一缺陷檢測處理。 方法mo始於階段111G，其執行-物件之_初始掃 描。初始掃描期間多重照明路徑會同步地照明物件。 階段1110接著是階段1120，其響應於初始掃描期間獲 5得的檢測信號將物件分節成節段。每一節段特徵在於 要材料。譬如，-節段大部份包括金屬或絕大部份可包括 半透明介電。 依據本發明之諸多實施例，階段1120接下來可為階段 1140、階段U25、階段113〇之組合。 1〇 階段1125包括接收節段資訊(譬如，物件的-設計擋， 先前掃描的結果，等等）。階段II25接著是階段1U0,其響 應於節段資訊判定各節段之一主要材料。 曰 階段1130判定各節段之一主要材料。 階段1130接著可為階段⑴5，其在知道各節段之主要 b材料之後，判定各節段之表面紐度。因此，主要材料— 、關疋m判疋何者將照明最適於節段粗韃程度之路捏。 又1135可包括，譬如：⑴以多重照明路徑—次—個 地照明節段，在判定節段之一主要材料後；其中不同的照 明路徑適於不同的節段可能粗糙程度；(ii)產生照明造成的 2〇檢測信號；以及㈣處理檢測信號來為節段選擇—照明路 徑。 又114 G包括在g卩段之—缺陷檢測掃描顧每—節^ 選擇一選定照明路徑來被致動。此選擇係響應於節段之—又 主要材料但也可響應於節段之表面粗糙程度。 24 200916763 階段1140接著是一缺陷檢測程序。缺陷檢測程序始於 階段1150，其僅以物件之一選定照明路徑照明每一物件節 段。 階段1150接著是階段1160，其響應於照明產生指示缺 5 陷之檢測信號。 階段1160接著可儲存檢測信號，並額外或另外處理檢 測信號來檢測缺陷。缺陷檢測處理為習知技藝。缺陷處理 程序一收到一標準物件影像變得更快更強健。 方法1100可應用在前述任何系統。 10 階段1140可包括選擇一節段之一選定照明路徑，若節 段主要材料為金屬，其照明一金屬表面和一金屬表面缺 陷，使得不論金屬表面之粗糙程度為何，金屬表面缺陷之 一影像實質上為黑色的而金屬表面之一影像實質上為白色 的。 15 階段1140可包括選擇一節段之一選定照明路徑，若節 段主要材料為一半透明介電材料，其照明節段之半透明表 面以及一半透明表面缺陷，使得不論半透明表面之粗糙程 度為何，半透明表面缺陷之一影像實質上為白色的，而半 透明表面之一影像實質上為黑色的。 20 階段1140可包括從一鏡射照明路徑、一散射性傾斜通 道、及一近鏡射照明路徑中選擇。此一選擇可與前述某些 系統有關，譬如第8A及8B圖之系統。 階段1160可包括藉由一成像路徑產生檢測信號，成像 路徑包括一相機和一遮罩。相機具有一感測表面，其平行22 200916763 conditions: minimum gold standard error and maximum global comparison. After the foregoing procedure is completed, the method 1000 proceeds to the stage 丨 060 which is most suitable for the illumination path (finding the best gold processing guide illumination channel _ labeled i0). Stage 10 6 0 then selects this illumination path 5 during the defect detection procedure to illuminate this segment and the corresponding segment (select i is the most suitable illumination channel for a particular application). Figure 11 illustrates a method 1100 for detecting defects using a multiple illumination path system, in accordance with an embodiment of the present invention. The method 1100 begins with a process that includes one of the stages 1110-1140, during which the illumination path for each segment is selected. It should be understood that the measurement process is completed, and the illumination path selection for each segment can be used during one or more defect detection scans of the same object. For example, a batch of electrical objects can be processed on a particular circuit by performing a measurement 15 2 〇 processing Wei Cha, _ in the inspection of other circuits, the illumination path (each segment) is selected. Therefore, each stage of the 'multiple patrol method (4) is a plurality of patrols in the (4)-(10) pieces, and is preferentially received by the same object (the object under ideal conditions is the result of the measurement process). Objects and other objects ^_ detection benefits Make conditions. Therefore, their surface (four) on the shell is "in the same pattern. The same degree of roughness on the top of the shell is the measurement action. The next step is the measurement process.; then the process is followed by the application rt», I ^ 23 200916763 in all other (or multiple) One of the object defect detection processes. The method mo begins at stage 111G, which performs an initial scan of the object. The multiple illumination paths illuminate the object synchronously during the initial scan. Stage 1110 is followed by stage 1120, which is obtained in response to the initial scan. The resulting detection signal segments the object into segments. Each segment is characterized by a material. For example, - most of the segments include a metal or a substantial portion may include a translucent dielectric. According to various embodiments of the invention, the stages 1120 may next be a combination of stage 1140, stage U25, stage 113. 1 stage 1125 includes receiving segment information (eg, object-design block, previous scan results, etc.). Stage II25 is followed by stage 1U0 And determining a main material of each segment in response to the segment information. 曰 Stage 1130 determines one of the main materials of each segment. Stage 1130 can then be stage (1) 5 After knowing the main b material of each segment, it determines the surface tens of each segment. Therefore, the main material - 疋 疋 疋 疋 疋 将 将 将 将 将 疋 疋 疋 疋 疋 照明 照明 照明 照明 照明 照明 照明 照明 照明 照明 照明 1 1 1 1 1 1 1 1 1 1 1 1 1 Including, for example: (1) multiple lighting paths - sub-ground lighting segments, after determining one of the main materials of the segment; wherein different illumination paths are suitable for different segments may be rough; (ii) caused by illumination 2〇 detection signal; and (4) processing the detection signal to select the segment for the segment—the illumination path. The 114 G is included in the g卩 segment—the defect detection scan is performed—the section selects a selected illumination path to be activated. In response to the segment - the primary material but also the surface roughness of the segment. 24 200916763 Stage 1140 is followed by a defect detection procedure. The defect detection procedure begins at stage 1150, which only illuminates the illumination path with one of the objects selected. Each object segment. Stage 1150 is followed by stage 1160, which generates a detection signal indicative of a defect in response to illumination. Stage 1160 can then store the detection signal and additionally or additionally process it. The signal is detected to detect defects. The defect detection process is a well-known technique. The defect handler becomes faster and more robust as soon as a standard object image is received. Method 1100 can be applied to any of the foregoing systems. 10 Stage 1140 can include selecting one of the segments The illumination path is selected. If the main material of the segment is metal, it illuminates a metal surface and a metal surface defect, so that regardless of the roughness of the metal surface, one of the metal surface defects is substantially black and one of the metal surfaces is substantially The upper portion is white. 15 Stage 1140 may include selecting one of the segments to select the illumination path, if the main material of the segment is a semi-transparent dielectric material, the semi-transparent surface of the illumination segment and the semi-transparent surface defect, such that the translucent surface The roughness is such that one of the translucent surface defects is substantially white, while one of the translucent surfaces is substantially black. The 20 stage 1140 can include selecting from a specular illumination path, a diffuse oblique channel, and a near-mirror illumination path. This option may be associated with some of the aforementioned systems, such as the systems of Figures 8A and 8B. Stage 1160 can include generating a detection signal by an imaging path that includes a camera and a mask. The camera has a sensing surface that is parallel
25 200916763 一物件表面。遮罩在成像路徑之一角度集合範圍中導入一 非對稱。此一檢測之一樣本可與前述某些系統有關，譬如 第8A及8B圖之系統。 階段1140可包括從一鏡射照明路徑、一前傾斜照明路 5 徑、及一後分散傾斜照明路徑中選擇。此一檢測之一樣本 可與前述某些系統有關，譬如第8C圖之系統。 較佳地階段1140可包括選擇一可組配照明路徑之兩種 組配中的一組配。在第一組配下，可組配的照明路徑以一 連續角度範圍照明物件。在第二組配下，可組配的照明路 10 徑以連續角度範圍之彼此相間隔開的兩個次範圍照明物 件。此一可組配的照明路徑之一範例如第8C、9、和9A圖 所繪示。 較佳地，階段1140包括從一前傾斜照明路徑、後分散 傾斜照明路徑、偏軸鏡射照明路徑中選擇；其中產生該檢 15 測信號之一相機以實質上不同於該物件之該表面上一法線 之一角度來成像該物件。此一檢測之一樣本可與前述某些 系統有關，譬如第8C圖之系統。 階段1140可包括為一節段選擇照明路徑，響應於包括 一光滑金屬表面，一半霧面金屬表面、一霧面金屬表面、 20 或一半透明介電表面之節段分類。 較佳地，階段1100之初始掃描可不同於階段1150之一 缺陷檢測掃描，基於下列至少一參數：速度、信號雜訊比、 及解析度。 較佳地，階段1350中判定一表面粗糙程度之掃描可不 26 200916763 同於P白段115G之-缺陷檢測掃描描，基於下列至少—表 數：速度，信號雜訊比、及解析度。 乂 -缺陷檢測掃描通倾其他高解析度和較佳信號 比之掃描低。 。 5 10 15 第12圖繪示依據本發明之一實施例之方法咖。 方法12 G G包括-缺陷檢測處理及彻—測定處理期間 所獲資訊(有關照明路徑每—節段之選擇）。 3 測定處理及缺陷檢測處理可應用在相同物件上，但非 必然如此。測定處理可應用在_物件而缺陷檢測處理可應 用在一或更多其他物件。譬如，物件和其他物件屬 抵次電路。 方法η姆™121G，其鋪由㈣抑關聯之— =照明路徑來照明一物件之各節段。選定照明 係響應於節段之-主要材料亦可響應於節段之—粗鍵程 度0 明產生指示缺 階段1210接著是階段1220，其響應於照 陷之檢測信號。 陷 階段1220接著是階段123〇， 其處理檢測信 號來檢測缺 20 又1210可包括以從一鏡射照明路徑、一 通道、及-近鏡射照明路徑中選出之—選定^路 明—物件節段。 L來照 ’成像路 件之一表 階段122 0可包括以-成像路經產生檢測信號 #包括-相機和-遮罩。相機包括平行所評估物 27 200916763 面的一感測表面。遮罩在成像路徑之一角度集合範圍中導 入一非對稱。 階段1210可包括以自一鏡射照明路徑、一前傾斜照明 路徑、及一後分散傾斜照明路徑中選出之一選定照明路徑 5 來照明一節段。 階段1210可包括選擇一可組配照明路徑之兩種組配中 的一組配。在一第一組配下，可組配的照明路徑以一連續 角度範圍照明物件。在第二組配下，可組配的照明路徑以 連續角度範圍中相間隔開的兩個次範圍來照明物件。 10 階段1210可包括以自一前傾斜照明路徑、後分散傾斜 照明路徑、偏軸鏡射照明路徑中選出之一選定照明路徑來 照明一節段選定。階段1220可包括產生檢測信號，其以實 質上不同於該物件之該表面上一法線之一角度來成像該物 件。 15 階段1210可包括以一選定照明路徑來照明一節段，其 選定係響應於包含一光滑金屬表面、一半霧面金屬表面、 一霧面金屬表面、或一半透明介電表面之一節段分類。 第13圖繪示依據本發明實施例之一種方法1300。 方法1300始於階段1310或1320其中一者。 20 階段1310包括在照明一導體表面時，在一缺陷檢測掃 描期間選擇一選定的照明路徑來被致用，使得該選定照明 路徑照明導體表面與導體表面缺陷時，不論導體導體表面 之粗糙程度，表面缺陷之一影像實質上為黑色的，而導體 表面之一影像實質上為白色的。 28 200916763 階段1320包括當照明一半透明表面時，在一缺陷檢測 掃描期間選擇一選定的照明路徑來被致用，使得不論半透 . 明表面之粗糙程度為何，半透明表面缺陷之影像實質上為 白色的，而半透明表面之影像實質上為黑色的。 5 階段1310接下來是階段1330，其照明該導體表面，特 : 別是照明一節段，該節段中之主材料是一導體。 階段1320接下來是階段1340，其照明該半透明表面， 特別是照明一節段，該節段中之主材料是一半透明表面。 1 階段1330和1340接下來是階段1350，其響應於照明產 10 生表示缺陷之檢測信號。 階段1350接下來是階段1360，其處理該檢測信號來檢 測缺陷。 理應瞭解的是，僅管本發明說明書中所提供之數值係 針對本發明之多重參數，多重不同的值可應用於這些參 15 數，依據本發明之不同實施例。 又進一步應理解的是，本發明之細節說明、本發明實 、. 施例之圖式細節、以及本發明諸不同實施例，有差異地實 施本發明。 I：圖式簡單說明3 20 第1圖繪示依據本發明之一實施例之缺陷以及表現這 類缺陷之理想灰階信號之範例； 第2A-2D圖繪示各種類型之表面拋光與反射模式； 第3A-3C圖繪示依據本發明之各種實施例之各種類型 之照明與成像方式； 29 200916763 第3D圖繪示依據本發明之一實施例一缺陷之一影像及 其背景； 第4 A圖繪示依據本發明之一實施例之一照明與成像方 式； 5 第4B-4D圖繪示一缺陷之成像及其響應於第4A圖不同 的照明所獲得之背景； 第5 A圖繪示依據本發明之一實施例之一照明及成像方 式； 第5B-5D圖繪示一缺陷之成像及其響應於第4A圖不同 10 的照明所獲得之背景； 第6 A圖繪示依據本發明之一實施例之一照明及成像方 式； 第6B圖繪示一缺陷之成像及其響應於第4A圖不同的 照明所獲得之背景； 15 第7A圖依據本發明之一實施例之一照明及成像方式， 以及各種以不同角度獲得之成像； 第7B-7C圖繪示施以不同的照明所獲得之檢測信號； 第8 A圖繪示依據本發明之一實施例之一多重照明路徑 糸統； 20 第8B圖繪示依據本發明之一實施例之一多重照明路徑 系統； 第8 C圖繪示依據本發明之一實施例之一多重照明路徑 系統； 第9A圖繪示依據本發明之一實施例之一多重照明路徑 30 200916763 系統； 第9B-9D圖繪示依據本發明之一實施例之一可組配之 • 照明路徑； ， 第10圖繪示依據本發明之一實施例之一方法； 5 第11圖繪示依據本發明之一實施例之一方法； ^ 第12圖繪示依據本發明之一實施例之一方法；以及 第13圖繪示依據本發明實施例之一種方法。 I：主要元件符號說明3 16 焊接遮罩下異質材料 404 不規則的光面質地 17 烊接遮罩括痕(表面缺陷） 406 規則的半霧面質地 19 焊接遮罩空隙(表面下缺陷） 514 近反射漫射發光 18 導體3D表面缺陷 512 禁止角度區域 11 影像灰階位準 510 傾斜漫射發光 212 光面拋光金屬表面 502 3D表面缺陷 216 半霧面拋光金屬表面 504 不規則的光面或半霧面質 218 霧面拋光金屬表面 地 240 半透明光滑表面 506 規則的半霧面質地 310 CCD透鏡 801 處理器 414 近反射漫射發光 810 成像遮罩角度 412 禁止角度區域 850 近-鏡像通道·半霧面金屬 410 傾斜漫射發光 表面 402 3D表面缺陷 820 發光遮罩角度 31 200916763 130 鏡射通道：光面金屬表面 1120 步驟 870 散射傾斜通道：霧面金屬及 1125 步驟 光滑介電表面處理器 1130 步驟 890 透鏡 1135 步驟 820 發光遮罩角度 1140 步驟 1010 步驟 1160 步驟 1020 步驟 1210 步驟 1025 步驟 1220 步驟 1030 步驟 1230 步驟 1035 步驟 1310 步驟 1040 步驟 1320 步驟 1045 步驟 1330 步驟 1050 步驟 1340 步驟 1060 步驟 1350 步驟 1070 步驟 1360 步驟 1080 步驟 1110 步驟 3225 200916763 The surface of an object. The mask introduces an asymmetry into the range of angle sets in one of the imaging paths. One of the samples of this test may be associated with some of the systems described above, such as the systems of Figures 8A and 8B. Stage 1140 can include selecting from a mirrored illumination path, a front tilted illumination path 5, and a post-dispersed oblique illumination path. One of the samples of this test may be associated with some of the aforementioned systems, such as the system of Figure 8C. Preferably stage 1140 can include selecting a set of two of the combinations of an configurable illumination path. In the first set, the configurable illumination path illuminates the object in a continuous angular range. In the second set, the illuminable path 10 can be assembled with two sub-range illumination objects spaced apart from each other in a continuous angular range. One of the configurable illumination paths is depicted, for example, in Figures 8C, 9, and 9A. Preferably, stage 1140 includes selecting from a front oblique illumination path, a rear dispersion oblique illumination path, and a off-axis mirror illumination path; wherein one of the detection signals is generated to be substantially different from the surface of the object The object is imaged at an angle from a normal. One of the samples of this test may be associated with some of the aforementioned systems, such as the system of Figure 8C. Stage 1140 can include selecting an illumination path for a segment, in response to segment classification including a smooth metal surface, a half matte metal surface, a matte metal surface, a 20 or a half transparent dielectric surface. Preferably, the initial scan of stage 1100 can be different from one of the stage 1150 defect detection scans based on at least one of the following parameters: speed, signal to noise ratio, and resolution. Preferably, the stage 1350 determines that a surface roughness scan may be the same as the P white segment 115G - defect detection scan based on at least the following: number of speeds, signal to noise ratio, and resolution.乂 - Defect detection scans other high resolution and better signals than scans. . 5 10 15 Figure 12 illustrates a method coffee in accordance with an embodiment of the present invention. Method 12 G G includes - defect detection processing and information obtained during the measurement process (selection of each segment of the illumination path). 3 The measurement process and the defect detection process can be applied to the same object, but this is not necessarily the case. The measurement process can be applied to the object and the defect detection process can be applied to one or more other objects. For example, objects and other objects are subordinate circuits. The method ηmTM121G, which is laid out by (iv) the associated -= illumination path to illuminate the segments of an object. The selected illumination system is responsive to the segment - the primary material may also be responsive to the segment - the coarse key degree 0 is indicative of the absence phase 1210 followed by phase 1220, which is responsive to the detection signal of the illumination. The trapping stage 1220 is followed by a phase 123, which processes the detection signal to detect the missing 20 and 1210, which may include selecting from a mirrored illumination path, a channel, and a near-mirror illumination path. segment. L illuminates one of the imaging paths. Stage 122 0 may include a - imaging path to generate a detection signal #include - camera and - mask. The camera includes a sensing surface that is parallel to the evaluation object 27 200916763. The mask introduces an asymmetry in the range of angular extents of the imaging path. Stage 1210 can include illuminating a segment by selecting one of the selected illumination paths 5 from one of a mirrored illumination path, a front tilted illumination path, and a rear dispersed oblique illumination path. Stage 1210 can include selecting a set of two of the combinations of an configurable illumination path. In a first set, the configurable illumination path illuminates the object in a continuous angular range. In the second set, the configurable illumination path illuminates the object in two sub-ranges spaced apart in a continuous angular range. The 10 stage 1210 can include illuminating the segment selection by selecting one of the selected ones from a front oblique illumination path, a rear dispersed oblique illumination path, and a off-axis mirror illumination path. Stage 1220 can include generating a detection signal that images the object at an angle that is substantially different from a normal on the surface of the object. 15 Stage 1210 can include illuminating a segment with a selected illumination path responsive to a segment classification comprising a smooth metal surface, a half matte metal surface, a matte metal surface, or a semi-transparent dielectric surface. FIG. 13 illustrates a method 1300 in accordance with an embodiment of the present invention. Method 1300 begins with one of stages 1310 or 1320. 20 stage 1310 includes, when illuminating a conductor surface, selecting a selected illumination path to be utilized during a defect detection scan such that the selected illumination path illuminates the conductor surface and the conductor surface defect, regardless of the roughness of the conductor conductor surface, One of the surface defects is substantially black, and one of the surface of the conductor is substantially white. 28 200916763 Stage 1320 includes selecting a selected illumination path to be utilized during a defect detection scan when illuminating a semi-transparent surface such that the image of the translucent surface defect is substantially the same regardless of the roughness of the semi-transparent surface The image of the white, translucent surface is substantially black. 5 Stage 1310 is followed by stage 1330, which illuminates the surface of the conductor, specifically: a section of illumination in which the primary material is a conductor. Stage 1320 is followed by stage 1340, which illuminates the translucent surface, particularly the illumination segment, where the primary material is a semi-transparent surface. The 1 stage 1330 and 1340 are followed by a stage 1350 which, in response to the illumination, produces a detection signal indicative of a defect. Stage 1350 is followed by stage 1360, which processes the detection signal to detect defects. It should be understood that although the numerical values provided in the present specification are directed to the multiple parameters of the present invention, multiple different values may be applied to these parameters, in accordance with various embodiments of the present invention. It is further understood that the invention may be variously described in detail, the details of the invention, the details of the invention, and the various embodiments of the invention. I: BRIEF DESCRIPTION OF THE DRAWINGS 3 20 FIG. 1 illustrates an example of a defect and an ideal gray scale signal for expressing such a defect according to an embodiment of the present invention; and FIGS. 2A-2D illustrate various types of surface polishing and reflection modes. 3A-3C illustrate various types of illumination and imaging modalities in accordance with various embodiments of the present invention; 29 200916763 3D depicts an image of a defect and its background in accordance with an embodiment of the present invention; The figure illustrates one illumination and imaging mode according to an embodiment of the present invention; 5 4B-4D illustrates the image of a defect and the background obtained by the different illumination in accordance with FIG. 4A; FIG. 5A An illumination and imaging method according to an embodiment of the present invention; FIG. 5B-5D illustrates an image of a defect and a background obtained by the illumination of different 10 in FIG. 4A; FIG. 6A illustrates the invention according to the present invention One embodiment of illumination and imaging mode; FIG. 6B illustrates a defect image and its background obtained by different illuminations in FIG. 4A; 15 FIG. 7A is an illumination according to an embodiment of the present invention Imaging method, and various Imaging obtained at the same angle; 7B-7C is a detection signal obtained by applying different illumination; FIG. 8A is a diagram showing a multiple illumination path system according to an embodiment of the present invention; A multiple illumination path system in accordance with an embodiment of the present invention; FIG. 8C illustrates a multiple illumination path system in accordance with an embodiment of the present invention; FIG. 9A illustrates an embodiment in accordance with the present invention One of the multiple illumination paths 30 200916763 system; the 9B-9D diagram illustrates an illumination path that can be assembled in accordance with one embodiment of the present invention; and FIG. 10 illustrates a method in accordance with an embodiment of the present invention FIG. 11 illustrates a method in accordance with an embodiment of the present invention; ^ FIG. 12 illustrates a method in accordance with an embodiment of the present invention; and FIG. 13 illustrates a method in accordance with an embodiment of the present invention. I: Main component symbol description 3 16 Welded material under heterogeneous material 404 Irregular smooth texture 17 遮 遮 括 ( (surface defect) 406 Regular semi-matte texture 19 Welded mask void (subsurface defect) 514 Near-reflective diffused light 18 Conductor 3D surface defect 512 Prohibited angle region 11 Image gray scale level 510 Inclined diffused light 212 Smooth polished metal surface 502 3D surface defect 216 Semi-matte polished metal surface 504 Irregular smooth or half Matte 218 Matte Polished Metal Surface 240 Translucent Smooth Surface 506 Regular Half Matte Texture 310 CCD Lens 801 Processor 414 Near Reflection Diffuse Light 810 Imaging Mask Angle 412 Prohibited Angle Area 850 Near-Mirror Channel Half Matte Metal 410 Inclined Diffuse Light Emitting Surface 402 3D Surface Defect 820 Illuminated Matte Angle 31 200916763 130 Mirror Channel: Glossy Metal Surface 1120 Step 870 Scattering Slope Channel: Matte Metal and 1125 Step Smooth Dielectric Surface Processor 1130 Step 890 Lens 1135 Step 820 Illumination Mask Angle 1140 Step 1010 Step 1160 Step 1020 Step 1210 step 1025 step 1220 step 1030 step 1230 step 1035 step 1310 step 1040 step 1320 step 1045 step 1330 step 1050 step 1340 step 1060 step 1350 step 1070 step 1360 step 1080 step 1110 step 32