201018895 九、發明說明： 【發明所屬之技術領域】 本發明是有關於一種檢測裝置及取得檢測影像的方法 ，特別是指一種具有半球形反射罩的檢測裝置及取得檢測 影像的方法。 【先前技術】 隨著光電產業的進步，作為光電產品中重要元件之一 的鏡片（lens)，其發展也一日千里，鏡片上的微小瑕疵會 造成光電產品之品質上的巨大影響，因此對於出廠或待組 裝的鏡片進行嚴格的檢測及把關，乃供應及生產線上極為 重要的一環。 現有鏡片檢測裝置，主要技術為照亮待測鏡片後取像 ，所取得的影像傳輸到電腦端可進行後續比對分析工作， 藉此找出有缺陷的鏡片。然而，照亮待測鏡片的方式繁多 ，取像的模式也各不同，因此取得的影像以及後續比對分 析的效果也就有所差異。例如圖1所示的一習知鏡片檢測 裝置，其利用兩組發光裝置72、73對一待測鏡片71發出 光線，每一組發光裝置72、73各包括等距交錯環設於待測 鏡片71周圍的二個發光件721、722、723及731、732、 733 ’再以一取像裝置74對該被照射的待測鏡片71擷取影 像，所擷取之影像傳送到比對裝置75進行比對分析。 然而’實際檢測狀況是，當對待測鏡片71直接發射過 多過亮的光線，常導致待測鏡片71發生反光現象，微小的 瑕疯反而被掩蓋而無法辨識；但若照射光線不足，也會發 201018895 生影像太暗而難以比對的狀況。 又如圖2所示另一種習知的鏡片檢測裝置，主要是將 光源81設置於一待測鏡片82下方位置處，利用背光的打 光方式，將鏡片82影像投攝於一屏幕（screen) 83上再 以影像擷取裝置84擷取屏幕83上受光源81照明的鏡片82 影像，當鏡片82有缺陷時，其影像會呈不同的光線明暗。 由於上述的打光方式通常利用長管狀的日光燈以背光 方式進行，光線照射上容易產生不均勻之情形，而導致檢 測的盲點。 綜合來說’現有的檢測 因素影響，容易出現誤檢， 測出鏡片缺陷的檢測方法。 【發明内容】 方式受到反光或光線不均勻等 因此，實有必要尋求能有效檢 因此’本發明之目的，即在 朝待測物取得影像的檢測裝置。〃彳兩角度斜向 待熟發月之3目的’在於提供—種以兩個角度斜向朝 】物取得影像的取得檢測影像的方法。 定義於ί吉Γ明之檢測裝置，適用於檢測-待測物，並 含-第一照明單元、—二直線’該檢測裝置包 -第二取像單元。該第單象=-第二照明單元及 該待測物上方的第一燈罩，、及複數;=口朝下地位於 頂部貫穿形成一第—取像口 ：D亥第-燈罩 偏離該第-直線的第—位置，該等 1 - 位於朝-側 x等第一光源發出的光束部 201018895 分經該第一燈罩内表面反射而由該開口朝向該待測物提供 複數檢測光束，部分穿出該第一取像口；該第一取像單元 與該第-直線夾1角度地通過該第一取像口朝該待測^ 取得一第-影像，該第-影像存在一偏心的第_暗區；該 第二照明單元包括一開口朝下地位於該待測物上方的第二 燈罩’及複數第二光源’該第二燈罩頂部貫穿形成—第二 取像口，且該第三取像口位於朝另一侧偏離該第一直線的 •帛二位置’該等第二光源發出的光束部分經該第二燈罩内 表面反射而由該開口朝向該待測物提供複數檢測光束部 分穿出該第二取像口；豸第二取像單元與該第一直線夹一 /5角度地通過該第二取像口朝該待測物取得一第二影像， 該第二影像存在一偏心的第二暗區，1在該第一影像與第 二影像疊置情況下，其第一暗區與第二暗區不重疊。上述 該待測物是先後在一第一站及一第二站分別進行取像而 該第一照明單元及第一取像單元是設於該第一站上方且 % 該第二照明單元及第二取像單元是設於該第二站上方。 較佳地，更包含一用以將該待測物由該第一站輸送至 該第二站的輸送單元，該輸送單元是輸送帶、轉盤或機械 手臂其中之一。 另外’本發明之檢測裝置亦可使待側物不移動，利用 第一照明單元及第一取像單元移動來達成檢測目的。即， 該檢測裝置包含一第一照明單元及一第一取像單元；該第 一照明單元，包括一開口朝下地位於該待測物上方的第一 燈罩，及複數第一光源，該第一燈罩頂部貫穿形成一第一 201018895 取像口，且該第一燈罩可伤 如“ 澄罩了位移，使該第-取像口在一朝一 側偏離該第一直線的第一位置與一 ^ AA ^ . 、朝另侧偏離該第一直 該第-^星内^間變換’該等第一光源發出的光束部分經 檢測光St反射而由該開口朝向該待測物提供複數 檢初先束’部分穿出該第一取像口；該第-取像單元，可 位移地㈣該第-燈罩上方，當該第—取像口位於該第一 =置’該第-取像單元與該第_直線夹1角度地通過該 第一取像口朝該待測物取得一第一影像，該第一影像存在 € -偏心的第-暗區；當該第一取像口位於該第二位置，該 第-取像單元與該第-直線夹_万角度地通過該第一取像 口朝該待測物取得-第二影像，該第二影像存在一偏心的 第二暗區，且在該第一影像與第二影像疊置情況下，其第 —暗區與第二暗區不重疊。 上述該第一暗區不超過一通過該第一影像的中心的第 基準線，該第二暗區不超過一通過該第二影像的中心的 第二基準線，在該第一影像與第二影像不旋轉而疊置情況 下’該第一基準線與第二基準線重合》 © 較佳地，α。 另外，該第一燈罩及第二燈罩皆概呈半球形，該等第 一光源是環狀排列地設於該第一燈罩内邊緣鄰近其開口處 ’而該等第二光源是環狀排列地設於該第二燈罩内邊緣鄰 近其開口處。 本發明取得檢測影像的方法，包含以下步驟： (A)由一側斜通過一燈罩的一取像口朝—待測物取得 201018895 一影像，該影像存在一偏心的第一暗區； (B) 由另一侧斜通過該取像口朝該待測物取得另—影 像，該另一影像存在一偏心的第二暗區，且在該二影像疊 置情況下，該第一暗區與第二暗區不重疊；及 (C) 組合該二影像中不含暗區的部份，獲得一可供判 讀是否有缺陷的檢測影像。 本發明之功效藉由第一照明單元及第一照明單元提供 均勻光源，並經由組合第一影像與第二影像不含暗區的部 分，以避免誤檢。 【實施方式】 有關本發明之前述及其他技術内容、特點與功效，在 以下配合參考圖式之較佳實施例的詳細說明中，將可清楚 地呈現。 本發明主要是利用照明單元對待測物進行反射式照明 ，再對該待測物取像。然而，取像裝置所在位置若介於照 明單元與待測物之間，將遮掩照射到待測物上的光線，若 由側邊朝待侧物取像，則因角度的關係將無法取得待側物 較遠一侧的影像。因此，取像裝置最好位於照明單元與待 侧物所連成的直線之上，且在照明單元上方，才能以較佳 角度取像而不遮掩光線。因此，本發明使照明單元的燈罩 開設取像口，以供取像裝置穿過該取像口對該待側物取像 。然而’取像口的設置會導致部分光線未被反射到待側物 上’在取得的影像中形成暗區，因此，發展出本發明之設 計。 201018895 圖3與圖4所示為本發明檢測裝置1〇〇的較佳實施例 ’適用於檢測-待測物1 ’並包含一輸送單元2、一第一照 明單元3、-第-取像單元4…第二照明單元$及一第二 取像單元6。 本實施例的輸送單元2是以一輸送帶為例，其概略呈 水平設置’可沿輸送方向⑶3中箭頭所示方向）將待測 物1先後輸送至兩預定檢測站，即一第一站21與一第二站 22。第一照明單元3及第一取像單元4是設置於第一站η ❿ 上方，而第二照明單元5及第二取像單元6是設置於第二 站22上方。另外’輸送單元2亦可以是轉盤或機械手臂等 ，只要是能達成先後輸送至兩預定檢測站的效果之裝置即 可。 配合參閱圖5,第一照明單元3為半球形擴散光源式照 明裝置，包括第-燈罩31及複數第_光;原&第一燈罩 31呈半球形，並具有-朝下的開σ 311及貫穿頂部之圓形 的第-取像口 312，而該等第一光源32是環狀排列地設於 ❿ 該第-燈罩Μ内邊緣鄰近其開σ 311冑。定義一垂直通過 待測物1中心的第一直綾η 面 直、银U以® 5的方位來說，當待測 物！位於第-站21時’第一取像σ3ΐ2是位於朝右側偏離 該第-直線U的第一位置，而該等第—光源32發出的光束 321部分經該第一燈罩31内表面反射而由開口 m朝向待 測物1提供複數檢測光束322,部分穿出該第一取像口 (如圖5的斜線區域所示）。 第一取像單元4設於第_燈罩31上方，並與第一直線 10 201018895 U夹一 α角度地通過第一取像口 312朝待測物1取得一第 一景像41’第一影像41存在一偏心的第一暗區42(如圖6 所示）。 有關偏心的第一暗區42之形成原因，乃是因為光束 321穿出該第一取像口 .312的部分，不會經第一燈罩31内 表面反射，因此會在待測物1上產生一相對照射亮度較低 的陰影區’該陰影區的形狀對應取像口 312的形狀概呈圓201018895 IX. Description of the Invention: [Technical Field] The present invention relates to a detecting device and a method for obtaining a detected image, and more particularly to a detecting device having a hemispherical reflecting cover and a method for obtaining a detected image. [Prior Art] With the advancement of the optoelectronic industry, the lens, which is one of the important components in optoelectronic products, has developed rapidly. The tiny flaws on the lens will have a huge impact on the quality of optoelectronic products, so for the factory or The rigorous testing and inspection of the lenses to be assembled is an extremely important part of the supply and production line. In the existing lens detecting device, the main technology is to illuminate the image to be tested, and the obtained image is transmitted to the computer for subsequent comparison analysis to find the defective lens. However, there are many ways to illuminate the lens to be tested, and the modes of image acquisition are also different, so the effects of the acquired images and subsequent comparison analysis are also different. For example, a conventional lens detecting device shown in FIG. 1 uses two sets of light-emitting devices 72, 73 to emit light to a lens 71 to be tested, and each group of light-emitting devices 72, 73 includes an equidistant staggered ring disposed on the lens to be tested. The two illuminating members 721, 722, 723 and 731, 732, 733 ′ around the 71 are further captured by the image capturing device 71 by the image capturing device 74, and the captured image is transmitted to the comparing device 75. Perform an alignment analysis. However, the actual detection condition is that when the lens 71 to be tested directly emits too much bright light, the lens 71 to be tested often causes a reflection phenomenon, and the tiny madness is concealed and cannot be recognized; but if the illumination light is insufficient, it will also be emitted. 201018895 The situation is too dark and difficult to match. Another conventional lens detecting device is shown in FIG. 2, which mainly places the light source 81 at a position below the lens 82 to be tested, and uses the backlighting method to project the image of the lens 82 on a screen. The image capturing device 84 captures the image of the lens 82 illuminated by the light source 81 on the screen 83. When the lens 82 is defective, the image will have different light and darkness. Since the above-mentioned lighting method is usually performed by using a long tubular fluorescent lamp in a backlight manner, unevenness in light irradiation is likely to occur, resulting in a blind spot of detection. In general, the influence of existing detection factors is prone to false detection and detection of lens defects. SUMMARY OF THE INVENTION The method is subject to reflection or uneven light. Therefore, it is necessary to seek a detection device capable of effectively detecting the object of obtaining the image toward the object to be tested. 〃彳 Two angles obliquely The purpose of the accommodating month is to provide a method for obtaining images by obtaining images at two angles. The detecting device defined in ί吉Γ明 is applicable to the detecting-test object, and includes a first lighting unit, a two-line unit, a detecting device package, and a second image capturing unit. The first single image = the second lighting unit and the first lamp cover above the object to be tested, and the plurality; the mouth is located at the top of the mouth to form a first image capturing port: the Dhai-the lampshade deviates from the first line a first position, the first light beam portion 201018895, which is emitted from the first light source, such as toward the side x, is reflected by the inner surface of the first lamp cover, and the plurality of detection beams are provided from the opening toward the object to be tested, and the light beam is partially passed out. a first image capturing unit; the first image capturing unit and the first straight image folder 1 obtain a first image through the first image capturing port toward the to be tested, and the first image has an eccentric _ dark The second lighting unit includes a second lamp cover 'and a plurality of second light sources' with the opening facing downwards. The second lamp cover top forms a second image capturing port, and the third image capturing port a portion of the light beam emitted from the second light source that is offset from the first line toward the other side. The portion of the light beam emitted by the second light source is reflected by the inner surface of the second lamp cover to provide a plurality of detection beam portions from the opening toward the object to be tested. Two taking image port; 豸 second image capturing unit and the first A second image is obtained from the second image capturing port toward the object to be tested by the second image capturing port. The second image has an eccentric second dark area, and the first image overlaps with the second image. In the case, the first dark area and the second dark area do not overlap. The object to be tested is imaged separately at a first station and a second station, and the first illumination unit and the first image capturing unit are disposed above the first station and the second illumination unit and the second The second image capturing unit is disposed above the second station. Preferably, further comprising a transport unit for transporting the object to be tested from the first station to the second station, the transport unit being one of a conveyor belt, a turntable or a mechanical arm. Further, the detecting device of the present invention can also prevent the side object from moving, and the first lighting unit and the first image capturing unit are moved to achieve the detection purpose. That is, the detecting device includes a first lighting unit and a first image capturing unit; the first lighting unit includes a first lamp cover that is open above the object to be tested, and a plurality of first light sources, the first A first 201018895 image capturing port is formed through the top of the lampshade, and the first lampshade can be wounded such that the displacement is such that the first image capturing port is offset from the first position of the first straight line with a ^AA ^ Deviating from the first side to the other side, the first portion of the first-to-one-to-star conversion is performed. The portion of the light beam emitted by the first light source is reflected by the detection light St to provide a plurality of initial beams from the opening toward the object to be tested. Partially exiting the first image capturing port; the first image capturing unit is displaceably (4) above the first lampshade, and when the first image capturing port is located at the first image capturing unit and the first image capturing port _ a linear clip 1 angularly passes through the first image taking port to obtain a first image toward the object to be tested, the first image has a eccentric inner-dark region; when the first image capturing port is located at the second position The first image capturing unit and the first straight line folder are oriented through the first image capturing port toward the first image capturing port The second image is obtained by the object, and the second image has an eccentric second dark region, and in the case where the first image and the second image are overlapped, the first dark region and the second dark region do not overlap. The first dark area does not exceed a first reference line passing through a center of the first image, and the second dark area does not exceed a second reference line passing through a center of the second image, in the first image and the second image The first reference line and the second reference line are coincident with each other without being rotated. In other words, the first and second lamp covers are substantially hemispherical, and the first light sources are rings. The second light source is disposed in an annular arrangement around the inner edge of the second lamp cover adjacent to the opening thereof. The method for obtaining the detected image includes the following Steps: (A) Obtaining an image of 201018895 from an image capturing port obliquely passing through a light cover toward the object to be tested, the image has an eccentric first dark area; (B) obliquely passing the image through the other side The mouth obtains another image toward the object to be tested, and the other image exists An eccentric second dark area, and in the case of the two image overlays, the first dark area and the second dark area do not overlap; and (C) combining the portions of the two images that do not contain the dark area, obtaining a The invention can determine whether the defective image is detected by the first illumination unit and the first illumination unit, and combines the first image and the second image without the dark region to avoid false detection. The above and other technical contents, features and effects of the present invention will be apparent from the following detailed description of the preferred embodiments of the preferred embodiments. Reflective illumination is performed, and the object to be tested is taken. However, if the position of the image capturing device is between the illumination unit and the object to be tested, the light that is irradiated onto the object to be tested is hidden, and if the image is viewed from the side When the side object is taken, the image on the far side of the side object cannot be obtained due to the angle. Therefore, the image capturing device is preferably located above the line connecting the illumination unit and the object to be side, and above the illumination unit, the image can be taken at a better angle without obscuring the light. Therefore, the present invention allows the lamp cover of the illumination unit to open the image taking port for the image taking device to take an image of the object to be side through the image taking port. However, the setting of the image taking port causes a portion of the light to be reflected onto the object to be viewed to form a dark region in the acquired image, and therefore, the design of the present invention has been developed. 201018895 FIG. 3 and FIG. 4 show a preferred embodiment of the detecting device 1 of the present invention for 'detecting the object to be tested 1' and including a conveying unit 2, a first lighting unit 3, and a first image capturing unit. The unit 4 is a second lighting unit $ and a second image capturing unit 6. The conveying unit 2 of the embodiment is exemplified by a conveyor belt, which is arranged horizontally to 'send the object to be tested 1 to two predetermined detection stations in a direction indicated by an arrow in the conveying direction (3) 3, that is, a first station. 21 and a second station 22. The first lighting unit 3 and the first image capturing unit 4 are disposed above the first station η , , and the second lighting unit 5 and the second image capturing unit 6 are disposed above the second station 22 . Further, the transport unit 2 may be a turntable, a robot arm or the like, as long as it is capable of achieving the effect of successively transporting it to two predetermined detection stations. Referring to FIG. 5, the first illumination unit 3 is a hemispherical diffused light source type illumination device, including a first lampshade 31 and a plurality of _th-lights; the original & first lampshade 31 is hemispherical and has a downward-opening σ 311 And a circular first-to-image port 312 extending through the top, and the first light sources 32 are arranged in a ring shape on the inner edge of the first lampshade, adjacent to the opening σ311. Define a vertical straight through the center of the object to be tested 1 in the direction of the first straight 绫, and the silver U in the orientation of the ® 5, when the object to be tested! When the first station 21 is located, the first image σ3ΐ2 is located at a first position that is offset from the first line U to the right side, and the light beam 321 emitted by the first light source 32 is partially reflected by the inner surface of the first lamp cover 31. The opening m provides a plurality of detection beams 322 toward the object to be tested 1, and partially passes through the first image taking port (as shown by the hatched area in FIG. 5). The first image capturing unit 4 is disposed above the first light-emitting cover 31 and obtains a first image 41' of the first image 41 toward the object to be tested 1 through the first image capturing port 312 at an angle α with the first straight line 10 201018895 U. There is an eccentric first dark area 42 (shown in Figure 6). The reason why the eccentric first dark region 42 is formed is because the portion of the light beam 321 that passes through the first image capturing port 312 is not reflected by the inner surface of the first lampshade 31, and thus is generated on the object to be tested 1. a shadow area having a relatively low illumination brightness. The shape of the shadow area corresponds to the shape of the image taking port 312.
形。在圖6中，是以一偏右的第一暗區42為例，但事實上 暗區的位置及大小’會受到取像口 312與第一直線U的偏 離程度、半球形第一燈罩31的曲率半徑或取像口 312與待 測物1的相對距離影響，有關這些條件的變化於後一併討 論0 再配合參閱圖7,第二照明單元5亦為半_擴散光源 式照明裝置，包括第二燈罩51及複數第二光源52，第二燈 罩51呈半球形，並具有一朝下的開口 511及貫穿頂部之圓 形的第二取像口 512 ’而該等第二光源52是環狀排列地設 於該第二燈罩51内邊緣鄰近其開口 511處。以圖7的方位 來說，當待測物i位於第二站22時，第二取像口 512是位 於朝左側偏離該第一直線11的第二位置，而如同第一昭明 單元3的第-光源32,該等第二光源52發出的光束部分經 该第一燈罩51内表面反射而由開口 5"朝向待測物i提供 複數檢測光束，部分穿出該第二取像口 512。 第二取像單元6設於該第二燈罩51上方並與第一直 線11夾1角度地通過該第二取像口犯朝該待測物丄取 201018895 得一第二影像61 ’該第二影像61存在一偏心的第二暗區 _ 62 (如圖8所示）’上述偏心的第二暗區62的成因與第— 暗區42之成因相同，而在圖8中，是以一偏左的第二暗區 62為例。 有關第一取像單元4及第二取像單元6的種類，可以 是一面陣型電荷粞合式影像棟取裝置（area CCD)或一感測陣 列裝置（CMOS)等，其皆為所屬技術領域中具有通常知識 者所熟知，不再贅述。 很明顯地’第一影像41與第二影像61不能各自單獨 〇 作為檢測依據，須進行適當選取再配合組合，即將第一影 像41與第二影像61疊置，當第一暗區42與第二暗區62 不相互重疊時’便可分別取第一影像41不含第一暗區42 的部分與第二影像61不含第二暗區62的部分，組合而得 到無暗區的完整檢測影像，以供判讀是否有缺陷。更進一 步來說，雖然本實施例以偏右的第一暗區42及偏左的第二 暗區62為例，事實上只要如上述達到不相互重疊即使第 一暗區42及第二暗區62都偏左、都偏右或一大一小形成❹ 皆可，差別在於如何組合第一影像41與第二影像61的方 式而已。 以下基於前述偏右的第一暗區42及偏左的第二暗區62 ，詳細說明第一影像41與第二影像61組合的方式參閱 圖6及圖8,首先分別界定一通過該第一影像41的中心的 第一基準線411 ’及一通過該第二影像61的中心的第二基 準線611，在本實施例中，第一基準線411及第二基準線 12 201018895 611疋平行輸送單元2的輸送方向，換句話說，該第—影像 41與第二影像61在不旋轉而疊置情況下該第一基準線 411與第二基準線611重合。藉由第一基準線4ιι可將第一 衫像41平分成左半部412及右半部413，而同樣地藉由第 二基準線611 V將第二影冑61 +分成左半冑612及右半部 613。备第一暗區42不超過第一基準線411，落在第一影像 41的右半部413,且第二暗區62不超過第二基準線611， 落在第二影像61的左半部612時，組合第一影像41的左 半部412與第二影像61的右半部613，便可得到無暗區的 完整檢測影像（參閱圖9 )。 基於上述組合第一影像41與第二影像61的方式，所 以前面提到的第一暗區42與第二暗區62必須偏心形成， 但是，若第一暗區42與第二暗區62其中一非偏心形成， 只要能滿足如上述不相互重疊的條件’亦可組合出無暗區 的完整檢測影像。但如果第一暗區42與第二暗區62皆非 偏心形成，則必產生重疊狀況，而重疊暗區處便難以檢測 出缺陷。 以下再配合圖4舉例說明如何調整各元件的相對位置 來達到上述結果。本實施例第一燈罩31及第二燈罩51外 型概呈半球形，本實施例以尺寸相同的半球形舉例說明。 定義分別通過第一取像口 312及第二取像口 512中心的直 線1及L2,並使1^及L2在垂直輸送單元2的輸送方向上 分別等距地位於第一直線u兩側，且Li及L2分別通過待 測物1 ;接下來，使第一取像單元4及第二取像單元6對稱 13 201018895 第一直線π地取像，即使点=—α，便可取得符合前述要 求的第一影像41及第二影像61。至於〇；與点的實際大小， 以第一取像單元4及第二取像單元6能夠取得整個待測物工 表面的影像為原則❶同樣地，開口 311及511的尺寸，基本 上是越小越好，但也是以能夠取得整個待測物丨表面的影 像為原則。另外，以第一取像單元4為例來說，其位置要 避開光束321穿出該第一取像口 312的部分，即是不可在 如圖5的斜線區域内，再調整找出具有較佳拍攝效果的相 對位置即可。 ❿ 有關上述將第一影像41及第二影像61平分並組合的 動作，是藉後端的影像處理系統（圖未示）來進行。另外 ，第一取像口 312及第二取像口 512的尺寸相對於待測物ι 表面應越小越好，藉此控制第一暗區42與第二暗區62的 大小，使暗區不會過大。 另外，除了上述將待測物1從第一站21輸送至第二站 22的檢測方式外，亦可使待測物！不動，藉由移動第一照 明單元3及第一取像單元4來進行檢測，並可將輸送單元2 Q 第一照明單元5及第一取像單元6省略不用。參閲圖1〇 ’首先使第-燈罩31的第-取像口 312是位於朝右侧偏離 待測物1的第一直線11之第一位置，而第一取像單元4與 第一直線11夾一《角度地通過第一取像口 312朝待測物工 取得一存在偏心第一暗區42的第一影像41 ;接下來，如圖 中箭頭方向所示，將第一燈罩31相對第一直線丨丨往反方向 移動，使第一取像口 312是位於朝左侧偏離該第一直線u 14 201018895 的第一位置，然後再使第一取像單元4與第一直線11炎一 /5角度地通過第一取像口 312朝待測物丨取得一存在偏心 第二暗區62的第二影像61，且在該第一影像41與第二彩 像61疊置時，第一暗區42與第二暗區62不重疊。如此〆 來，組合第一影像41不含第一暗區42的部分與第二影像 61不含第二暗區62的部分，便可同樣地得到可供判讀是否 有缺陷的檢測影像。 綜合言之，本發明提供一種取得檢測影像的方法包 含如圖11所示之以下步驟： 步驟Si—配合參閲圖10，使第一取像單元4由一側斜 通過第一燈罩31的第一取像口 312朝待測物丨取得第—影 像41，該第一影像41存在一偏心的第一暗區42; 步驟S2使第一取像單元6由另一側斜通過第—取像 口 312朝待測物1取得第二影像61，該第二影像61存在一 偏心的第二暗區62，且在第一影像41及第二影像61疊置 情況下’第一暗區42與第二暗區62不重疊； 步驟S3—組合第一影像41及第二影像61分別不含第 一暗區42與第二暗區62的部份，獲得一可供判讀是否有 缺陷的檢測影像。 綜上所述，藉由第一照明單元3及第一取像單元4，戈 再配合第二照明單元5及第二取像單元6，以兩個角度斜向 朝待測物1取得存在偏心的第一暗區42的第一影像4卜^ 存在偏心的第二暗區62的第二影像61，然後組合第一影像 41不含第一暗區42的部分與第二影像61不含第二暗區吣 15 201018895 的部分’便得到無暗區的完整檢測影像以供判讀是否有 缺陷’可避免第一暗區42或第二暗區62遮蔽缺陷而導致 誤檢’故確實能達成本發明之目的。 惟以上所述者，僅為本發明之較佳實施例而已，當不 能以此限定本發明實施之範圍’即大凡依本發明申請:利 範圍及發明說明内容所作之簡單的等效變化與修飾皆仍 屬本發明專利涵蓋之範圍内。 【圖式簡單說明】 圖1是一示意圖，說明一習知鏡片檢測裝置； ◎ 圖2是一示意圖，說明另一習知鏡片檢測裝置； 圖3是一立體示意圖，說明本發明之檢測裝置； 圖4是圖3中IV-IV剖面的元件結構示意圖； 圖5是一第一照明單元及一第一取像單元相對一待測 物位置的剖面示意圖； 圖6是該第一取像單元對該待測物所取得一第一影像 的示意圖； 圖7是一第二照明單元及一第二取像單元相對該待測❹ 物位置的剖面示意圖； 圖8是該第二取像單元對該待測物所取得一第二影像 的示意圖； 圖9是組合圖6中該第一影像及圖8中該第二影像分 別不含暗區的示意圖； 圖10是使該第一照明單元及一第一取像單元相對待測 物移動以取得二影像的示意圖；及 16 201018895 圖11是一流程圖，說明一取得檢測影像的方法之步驟 〇shape. In FIG. 6, the first dark area 42 of the right side is taken as an example, but in fact, the position and size ' of the dark area is affected by the degree of deviation of the image taking port 312 from the first straight line U, and the hemispherical first lampshade 31 The radius of curvature or the relative distance between the image taking port 312 and the object to be tested 1 is affected. The change of these conditions is discussed later. Referring again to FIG. 7, the second lighting unit 5 is also a semi-diffusion light source type lighting device, including a second lamp cover 51 and a plurality of second light sources 52. The second lamp cover 51 has a hemispherical shape and has a downward opening 511 and a circular second image taking port 512 ′ extending through the top portion. The second light sources 52 are rings. The inner edge of the second lampshade 51 is disposed adjacent to the opening 511 thereof. In the orientation of FIG. 7, when the object to be tested i is located at the second station 22, the second image taking port 512 is located at a second position deviating from the first straight line 11 toward the left side, and like the first portion of the first illuminating unit 3 The light source 32, the light beam emitted by the second light source 52 is partially reflected by the inner surface of the first lamp cover 51, and the plurality of detection beams are provided from the opening 5" toward the object to be tested, and partially exits the second image taking port 512. The second image capturing unit 6 is disposed above the second lamp cover 51 and is angled with the first straight line 11 through the second image capturing port to obtain a second image 61 'the second image from the object to be tested. There is an eccentric second dark area _ 62 (shown in FIG. 8). The cause of the eccentric second dark area 62 is the same as that of the first dark area 42, and in FIG. 8, it is a left side. The second dark area 62 is taken as an example. The type of the first image capturing unit 4 and the second image capturing unit 6 may be an array type charge-coupled image capturing device (area CCD) or a sensing array device (CMOS), etc., all of which are in the technical field. It is well known to those of ordinary knowledge and will not be described again. Obviously, the first image 41 and the second image 61 cannot be individually used as detection basis, and must be appropriately selected and combined, that is, the first image 41 and the second image 61 are overlapped, when the first dark area 42 and the first image When the two dark regions 62 do not overlap each other, the portions of the first image 41 that do not include the first dark region 42 and the portions where the second image 61 does not contain the second dark region 62 can be respectively taken to obtain a complete detection without dark regions. Image for interpretation of defects. Further, although the present embodiment is exemplified by the first dark area 42 and the second dark area 62 which are leftward, in fact, as long as the above does not overlap each other even the first dark area 42 and the second dark area 62 is all left, right or small, and the difference is in how the first image 41 and the second image 61 are combined. The following describes the manner in which the first image 41 and the second image 61 are combined based on the first right dark area 42 and the left second dark area 62. Referring to FIG. 6 and FIG. 8 , firstly, a first pass is defined. a first reference line 411 ' at the center of the image 41 and a second reference line 611 passing through the center of the second image 61. In the present embodiment, the first reference line 411 and the second reference line 12 201018895 611 are transported in parallel. The conveying direction of the unit 2, in other words, the first image line 41 and the second image 61 overlap with the second reference line 611 without being rotated. The first shirt image 41 can be divided into the left half 412 and the right half 413 by the first reference line 4 ι, and the second image 61 + is divided into the left half 612 by the second reference line 611 V and Right half 613. The first dark area 42 does not exceed the first reference line 411, falls on the right half 413 of the first image 41, and the second dark area 62 does not exceed the second reference line 611, and falls on the left half of the second image 61. At 612, the left half 412 of the first image 41 and the right half 613 of the second image 61 are combined to obtain a complete detection image without dark areas (see FIG. 9). Based on the above manner of combining the first image 41 and the second image 61, the aforementioned first dark area 42 and second dark area 62 must be formed eccentrically, but if the first dark area 42 and the second dark area 62 are A non-eccentricity formation, as long as it satisfies the conditions that do not overlap each other as described above, can also be combined to form a complete detection image without dark areas. However, if both the first dark area 42 and the second dark area 62 are not eccentrically formed, an overlap condition will occur, and it is difficult to detect the defect at the overlapping dark areas. The following will be exemplified by how to adjust the relative positions of the components in conjunction with FIG. 4 to achieve the above results. In this embodiment, the first lampshade 31 and the second lampshade 51 are generally hemispherical in shape, and the embodiment is illustrated by a hemispherical shape of the same size. Lines 1 and L2 respectively passing through the centers of the first image taking port 312 and the second image taking port 512 are defined, and 1 and L2 are respectively equidistantly located on both sides of the first straight line u in the conveying direction of the vertical conveying unit 2, and Li and L2 respectively pass the object to be tested 1; next, the first image capturing unit 4 and the second image capturing unit 6 are symmetrically taken by the first line π, and even if the point = -α, the above requirements can be obtained. The first image 41 and the second image 61. As for the actual size of the dots, the first image capturing unit 4 and the second image capturing unit 6 can obtain the image of the entire surface of the object to be tested. Similarly, the sizes of the openings 311 and 511 are basically The smaller the better, but also the principle of being able to obtain the image of the entire surface of the object to be tested. In addition, taking the first image capturing unit 4 as an example, the position thereof should avoid the portion of the light beam 321 that passes through the first image capturing port 312, that is, it cannot be adjusted in the oblique line region of FIG. The relative position of the better shooting effect is sufficient. ❿ The above-described operation of halving and combining the first image 41 and the second image 61 is performed by an image processing system (not shown) at the back end. In addition, the size of the first image taking port 312 and the second image taking port 512 should be as small as possible with respect to the surface of the object to be tested ι, thereby controlling the size of the first dark area 42 and the second dark area 62 to make the dark area. Not too big. Further, in addition to the above-described detection method of transporting the object 1 to be transferred from the first station 21 to the second station 22, the object to be tested can also be made! The movement is performed by moving the first illumination unit 3 and the first image capturing unit 4, and the first illumination unit 5 and the first image capturing unit 6 can be omitted. Referring to FIG. 1A, first, the first image taking port 312 of the first lampshade 31 is located at a first position away from the first straight line 11 of the object to be tested 1 toward the right side, and the first image capturing unit 4 and the first straight line 11 are sandwiched. The first image 41 having the eccentric first dark area 42 is obtained angularly through the first image taking port 312 toward the object to be tested; next, the first light cover 31 is opposite to the first straight line as indicated by the direction of the arrow in the figure. Moving in the opposite direction, the first image taking port 312 is located at a first position deviating from the first line u 14 201018895 toward the left side, and then causing the first image capturing unit 4 to be at an angle of /5 to the first straight line 11 Obtaining a second image 61 having the eccentric second dark region 62 toward the object to be tested through the first image taking port 312, and when the first image 41 and the second color image 61 are overlapped, the first dark region 42 is The second dark areas 62 do not overlap. In this way, by combining the portion of the first image 41 that does not include the first dark region 42 and the portion of the second image 61 that does not include the second dark region 62, it is possible to obtain a detection image for which a defect can be read. In summary, the present invention provides a method for obtaining a detected image including the following steps as shown in FIG. 11: Step Si - Referring to FIG. 10, the first image capturing unit 4 is obliquely passed from one side to the first light cover 31 The image capturing port 312 obtains the first image 41 toward the object to be tested, and the first image 41 has an eccentric first dark region 42; and the step S2 causes the first image capturing unit 6 to obliquely pass through the first image capturing device from the other side. The port 312 obtains the second image 61 toward the object to be tested 1, and the second image 61 has an eccentric second dark area 62, and in the case where the first image 41 and the second image 61 are stacked, the first dark area 42 is The second dark area 62 does not overlap; Step S3—combining the first image 41 and the second image 61 respectively without the portions of the first dark area 42 and the second dark area 62, and obtaining a detection image for determining whether the defect is defective . In summary, the first illumination unit 3 and the first image capturing unit 4 cooperate with the second illumination unit 5 and the second image capturing unit 6 to obtain an eccentricity toward the object to be tested 1 at two angles obliquely. The first image of the first dark area 42 has a second image 61 of the second dark area 62 of the eccentricity, and then the portion of the first image 41 that does not include the first dark area 42 and the second image 61 do not contain the first image. The second dark area 吣15 part of 201018895 'gets a complete test image without dark area for interpretation of whether it is defective' can avoid the first dark area 42 or the second dark area 62 to cover the defect and cause false detection. The purpose of the invention. However, the above is only the preferred embodiment of the present invention, and the scope of the present invention is not limited thereto. All remain within the scope of the invention patent. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic view showing a conventional lens detecting device; FIG. 2 is a schematic view showing another conventional lens detecting device; FIG. 3 is a perspective view showing the detecting device of the present invention; 4 is a schematic structural view of an element of the IV-IV section of FIG. 3; FIG. 5 is a schematic cross-sectional view of a first illumination unit and a first image capturing unit with respect to a position of the object to be tested; FIG. 7 is a schematic cross-sectional view of a second illumination unit and a second image capturing unit relative to the object to be tested; FIG. 8 is the second image capturing unit A schematic diagram of a second image obtained by the object to be tested; FIG. 9 is a schematic diagram of combining the first image in FIG. 6 and the second image in FIG. 8 respectively without a dark region; FIG. 10 is a first lighting unit and a first lighting unit; A schematic diagram of the first image capturing unit moving relative to the object to be tested to obtain two images; and 16 201018895 FIG. 11 is a flowchart illustrating a method of obtaining a method for detecting an image〇
17 201018895 【主要元件符號說明】 100..........檢測裝置 1 .............待測物 11 ...........第一直線 2 .............輸送單元 21 ...........第一站 22 ...........第二站 3 .............第一照明單元 31 ...........第一燈罩 311、511··開口 312..........第一取像口 32 ...........第一光源 321 ..........光束 322 ..........檢測光束 4 .............第一取像單元 41 ...........第一影像17 201018895 [Explanation of main component symbols] 100..........Detection device 1 .............Test object 11 ........... First straight line 2.................. conveying unit 21 ........... first station 22 ........... second station 3 . ............first lighting unit 31 ...........first lampshade 311, 511·· opening 312.......... first Taking the image port 32 ...........the first light source 321 ..........the light beam 322 ..........detecting the light beam 4 ... .......the first image capturing unit 41 ...........the first image
411.......... 第一基準線 412、612.. 左半部 413 > 613·· 右半部 42............ 第一暗區 5 ............. 第二照明單元 51............ 第二燈罩 512.......... 第二取像口 52............ 第二光源 6............. 第二取像單元 61............ 第二影像 611.......... 第二基準線 62............ 第二暗區 Lj ' L2 ···.· 直線 a ' β ..... 夾角角度411.......... First reference line 412, 612.. Left half 413 > 613·· Right half 42............ First dark area 5 ............. Second lighting unit 51............ Second lampshade 512.......... Second image capturing port 52 ............ Second light source 6............. Second image capturing unit 61............ Second image 611.......... Second reference line 62............ Second dark area Lj ' L2 ····· Line a ' β ..... Angle angle
1818