I3436!6 九、發明說明： ‘【發明所屬之技術領域】 本發明關於-種太陽能電池（s〇lar(：ell)或其晶圓（驗r) 檢測方法及m尤指—種針對太陽能電池⑸⑹⑽）或 其晶圓（Wafer)於製程中檢測之技術。 【先前技術】I3436!6 Nine, invention description: '[Technical field to which the invention belongs] The present invention relates to a solar cell (s〇lar (:ell) or its wafer (inspection r) detection method and m especially refers to a solar cell (5) (6) (10)) or its wafer (Wafer) technology for testing in the process. [Prior Art]
由於太陽能具有豐富、清潔、安全1保之特性，而且屬於 難以壟斷之自然界公共財，自切來即被廣泛彻，*近代更成 為產生電力之新興輯，尤其是所謂太陽能電池更被普遍興起。 按’太陽能電池是敎陽能轉触電㈣裝置，不需要透過電解 質來傳遞導電離子，而是改採半導體產生PN結來獲得電位。當 半導體X到太陽光的照射時，大量的自由電子伴隨而生，而此電 子的移動又產生了電流，也就是在pN結處產生電位差。 按’太陽能產業鏈中’最上游的原料為矽（Si)，由單晶矽 或多晶石夕長晶成晶棒（Ing〇t)，然後再切割成晶圓（Wafer)， 成為太陽能電池（Solar Cell)的主要原料，再加以組裝成模組 C Module )’最後搭配負責直/交流電轉換電之調節器 (Inverter)，，置成太陽能光電系統(pv System)。 目前在晶圓與太陽能電池之製程方面，檢測事項甚多，除基 本之計算數量’尚包括產品是否缺角？是否磨亮？是否形成碎 4.或者監看疋否有晶圓與太陽能電池進料等狀態。然而’目前 由於晶圓與太陽能電池均具有「薄」（例如目前之太陽能電池 μ的56 、Cel1 ’其晶圓Wafer之厚度普遍已降至180#m，甚至更薄），及 「 易脆裂」之特性，因此在各工作流程中，容易因為「接觸」而 導致震動、碰損*使得晶圓或者太陽能電池成為不良品。因此如 何降低接觸所導致之傷損，更為製程以及檢測過程中必須提升之 技術。 ; 賴業界與學界曾提^ ίττ-lcd面版之自動化檢測，運用機 、器視覺’針對TFT-LGD面版上的勤丨電晶體，進行圖形追縱、定 鲁位、檢測及線寬、線距等相關製程量測工作。然而該技術主要係 針對娜之影像進行灰階判讀，亦屬—種「虛擬追縱」，而非針 對直接觀測，因此其真確性’往往必須透過較多次之模擬與分析、 比對。且’其與太陽能電池（s〇lar Ce⑴或其晶圓檢 測方法及其系統就目的性以及特徵性並不相同。 【發明内容】 。有#於先前麟巾，對於太陽能電池（SQlar㈤）或其晶 圓（Wafer)(亦有稱石夕晶片）之檢測，所差生人工檢測容易因為 *，觸而產生損害之問題’本發明者認為應有—種改善之道，尤其 $利用非接觸」、「自動化」與「效率化」之方式以及系統達到 前開之改善。 <· 本發明解決先制題之技術手段，主要在於個光學影像掏 取裝置擷取該太陽能電池⑽ar Cell)或其晶圓（耐打）之影 2取得擷取之純部位後’彻影像自動辨識軟體對於該操取 影像為處理’並顯示制結果’降低甚至免除人瑞觸式計算數 量 常够之失’啊配合各種所需要之檢測項目，建置標準或 爾’透過擷取之影像與標準或常態之影像資料庫 =^影像自動辨識軟體以及處理器能夠得到—個計量值、 訊號（例如警示語音、震動等），而終極達到自動化、 效率化、大幅降低損耗之檢測目的。 光ztr料求触前述綠，更進—步設計㈣統，包括一 旦/予衫雜裝置，設於檢測平台之鄰邊；一檢測平台，盘光學 =崎置觀聰；—紛_單元，連麟絲影像搁取 、4像處理單元包括—影像自動辨識軟體，—常態影像資 、及處理n ’藉由常⑮影像資料庫與擷取影像，經由處理 器比對與判斷影像意義。 【實施方式】 以下藉由圖式說明本發明之方法、系統、特色以及實施例， 俾使貴審查委員對於本發明有更進一步之瞭解。 _ 按’該太陽能電池（▲ Gell)係依設計所f要的電流進 -行晶圓⑽er)切割後’焊上箱條導線再將許多焊好的晶圓 .(Wafer)用4條串聯成一組，再和驗、與低鐵質強化 破璃層層資邊’ -同放入層壓機（laminate )的機台上做真空 封裝’製成module (plane /panel )稱之為模組或稱太陽能板。 而由於太陽能電池（Solar Cell)或其晶圓（wafei·)之特性 與屬性均適驗本發明’因此本發明將二者列為可實施之標的。 如第四圖單純檢測太陽能電池（s〇larCell),第六圖單純檢測其 I3436?6 ;··晶圓（Wafer)。第三圖則顯示皆可檢測太陽能電池（solar Cell) . 以及檢測其晶圓（Wafer)。 請參閱第-圖與第二圖所示，本發明係關於—種太陽能電池 . (Solar Cel1)或其晶圓（Wafer)檢測方法其實施例係以「計算 為太陽能電池（Solar Cell)或其晶圓（Wafer)之片數」為例， 步驟在於： .一'係利用光學影像擷取裝置（1)(例如第一圖之圖式中之電荷 ’ 耦合照像機CCD或CMOS)，自動擷取多片疊合之太陽能電池 (Solar Cell)或其晶圓（Wafer)側邊之中央或任一部位， 甚至可以為自動對焦（光學變焦），降低環境燈光干擾，增強 檢測穩定性。 二、 如第二圖所示，然後將擷取該太陽能電池（Solar· Cell)或 其晶圓（Wafer)之影像，傳遞至影像自動辨識軟體（3 1) 處理。 三、 由該影像自動辨識軟體（3 n計算其片數（圖例為計算片 _ 數’尚有其他之檢測項目）。 - 而本發明之該影像自動辨識軟體（3 1)可以如第六圖所示， -連接設一圖形使用者介面(GUI) ( 3 4 )以顯示結果。此外，該影 像自動辨識軟體（3 1)係採特殊演算法可避免因為太陽能電池 (Solar Cell)或其晶圓（wafer)之翹曲現象而造成計數錯誤。 又’該影像自動辨識軟體（3 1)介面可做參數微調，符合各製成 產品的厚薄度修正^且該檢測規格尺寸可依客製需求進行不同精 密度之設定。 本發明之鮮祕綠裝置⑴aD，亦可由具有喊光二極 體感光it賴祕獅影像，而此_絲二極體祕元件可以 例如以互補性氧化金屬半導體⑽s)照相機為之。 又’本發明之檢測方法，該光學影像擷取裝置⑴，盆取得 影像之方向，可以為X、γ、z軸任—軸向或χ、Y、Z轴任一 軸之兩兩組合之影像。例如由x軸、z軸方向取得側邊之分層疊 設之太陽能電池（SQlar cell)或其晶圓（Wafer)之影像，方便 觀察其層級而可以求得其片數之計算。亦可以由Μ (如俯視或 仰視），_之疊層絲，與常誠標特料庫她而得知其 層級數量（即片數）。 至於本發明較為具體之步驟，可為： A、 各裝置定位； B、 掏取由2 5㈣〇片疊合之複數太陽能電池（SQiar⑽） 或其晶圓（Wafer) X軸（側向）影像，可則數多寡而調整 擷取影像之相關步驟與機構； 儿自動辨4軟體（3 1)為—種計量影像自動辨識軟體， ㈣叶量影像自動辨雌體騎影料合所代表之數量。亦 =以，早片或複數片而檢測其他檢測項目，例如「缺角」、「磨 儿」碎4」、「進料狀態」，而檢測不同項目則該影像自動辨 2體改縣各軟體，例如「缺角影像自動辨識軟體」、「磨 心像自動辨識軟體」（如第四圖之檢測狀態）、「 ^識軟體」、「進料狀_自動辨識軟體」（例如:進料 s’膜包裝之>{數自動辨識軟體，以進行進料_包裝之片 1343m 數计算）以及所需要之比對資料庫形成比對。Because solar energy is rich, clean, and safe, it belongs to the natural wealth that is difficult to monopolize. It has been widely cut since its inception. * Modern times have become an emerging series of power generation, especially the so-called solar cells are generally emerging. Press 'Solar battery is a device that can be switched to electric (4). It does not need to transmit conductive ions through the electrolyte. Instead, it uses a semiconductor to generate a PN junction to obtain the potential. When the semiconductor X is irradiated with sunlight, a large amount of free electrons accompany it, and the movement of this electron generates a current, that is, a potential difference is generated at the pN junction. According to the raw material in the 'solar industry chain', the most upstream raw material is bismuth (Si), which is formed into a crystal rod (Ing〇t) by single crystal germanium or polycrystalline crystal, and then cut into wafers (Wafer) to become a solar cell. (Solar Cell)'s main raw materials, and then assembled into a module C Module) 'Last with the regulator responsible for the direct / AC conversion, (Inverter), set into a solar photovoltaic system (pv System). At present, there are many inspection items in the process of wafers and solar cells. Except for the basic calculation number, does the product include a missing corner? Is it polished? Whether it is broken 4. Or monitor whether there is a state such as wafer and solar cell feed. However, 'currently, both wafers and solar cells are "thin" (for example, the current solar cell μ of 56, Cel1 'the thickness of wafer Wafer has generally dropped to 180 #m, or even thinner), and "easy to crack Because of the characteristics, in each workflow, it is easy to cause vibration or damage due to "contact"*, making wafers or solar cells defective. Therefore, how to reduce the damage caused by contact, the process and the technology that must be improved during the inspection process. Lai industry and academics have raised the automatic detection of ^ ίττ-lcd panel, using machine and device vision 'for the diligent transistor on TFT-LGD panel, for graphic tracking, setting Lu, detection and line width, Line spacing and other related process measurement work. However, this technique mainly involves gray-scale interpretation of Na's image. It is also a kind of "virtual tracking" rather than direct observation, so its authenticity often has to be simulated and analyzed through multiple comparisons. And 'the solar cell (s〇lar Ce (1) or its wafer detection method and its system are not the same in terms of purpose and characteristics. [Inventive content] There are # previously lining towels, for solar cells (SQlar (5)) or Wafer (also known as Shixi wafer) detection, the manual detection of the poor is easy to cause damage due to *, the inventor believes that there should be a way to improve, especially $ non-contact, The method of "automation" and "efficiency" and the improvement of the system have been improved. <RTIgt; The present invention solves the technical problem of the first problem, mainly because an optical image capturing device takes the solar cell (10) ar Cell) or its wafer ( After the capture of the pure part, the image of the image is automatically processed for the operation of the image and the display result is reduced or even eliminated. Need to test the project, build standards or 'through the captured image and standard or normal image database = ^ image automatic identification software and the processor can get - a measured value , signal (such as warning voice, vibration, etc.), and ultimately achieve automation, efficiency, and greatly reduce the loss of detection purposes. Light ztr is required to touch the aforementioned green, and further into the step-by-step design (four) system, including the once/preventing device, located at the neighboring side of the detection platform; a detection platform, disk optics = Qishang Guancong; - _ unit, even The Linsi image shelving and 4-image processing unit includes an image automatic identification software, a normal image resource, and a processing n'. The image is compared and judged by the processor through the image database and the captured image. [Embodiment] Hereinafter, the method, system, features, and embodiments of the present invention will be described in the following drawings, and the reviewing committee will further understand the present invention. _ According to 'the solar cell (▲ Gell) according to the design of the current into the wafer (10) er) after cutting 'welding the box wire and then a number of welded wafers (Wafer) with 4 in series The group, the re-inspection, and the low-iron reinforced layer of the lining layer - the vacuum-packed on the machine that is placed on the laminator is made into a module (plane / panel) called a module or Called solar panels. Since the characteristics and attributes of the solar cell or its wafer (wafei) are suitable for the present invention, the present invention has been listed as an implementable subject. As shown in the fourth figure, the solar cell (s〇larCell) is simply detected. The sixth figure simply detects its I3436?6; Wafer. The third diagram shows that both solar cells can be detected and wafers (Wafer) are detected. Referring to the first and second figures, the present invention relates to a solar cell. (Solar Cel1) or its wafer (Wafer) detection method is described as "calculated as a solar cell (Solar Cell) or For example, the number of wafers (Wafer) is as follows: . 'Using an optical image capturing device (1) (for example, the charge in the pattern of the first figure 'coupling camera CCD or CMOS), automatically Capture the center or any part of the side of the laminated solar cell (Solar Cell) or its wafer (Wafer), even for auto focus (optical zoom), reduce ambient light interference and enhance detection stability. 2. As shown in the second figure, the image of the solar cell (Solar Cell) or its wafer (Wafer) is then transferred to the image automatic identification software (3 1) for processing. 3. The image is automatically recognized by the image (the number of slices is calculated by 3 n (the figure is the calculated slice _ number 'there are other detection items). - The automatic image recognition software (3 1) of the present invention can be as shown in the sixth figure. As shown, - a graphical user interface (GUI) (3 4) is connected to display the results. In addition, the automatic image recognition software (3 1) adopts a special algorithm to avoid the solar cell or its crystal. The warping phenomenon of the wafer causes the counting error. Also, the image automatic identification software (3 1) interface can be fine-tuned to meet the thickness correction of each manufactured product, and the inspection specification size can be customized. The setting of different precisions. The fresh green device (1) aD of the present invention may also be a camera with a shattering diode sensitization, which may be, for example, a complementary oxidized metal semiconductor (10) s) camera. For it. Further, in the detecting method of the present invention, the optical image capturing device (1) can obtain the image in the direction of the image, and can be an image of any combination of any one of the X, γ, and z axes, either the axial direction or the χ, Y, or Z axis. For example, an image of a solar cell (SQlar cell) or a wafer (Wafer) laminated on the side is obtained from the x-axis and the z-axis direction, and the number of slices can be calculated by observing the hierarchy. It is also possible to know the number of levels (ie the number of sheets) by Μ (such as looking down or looking up), _ the laminated wire, and Chang Cheng's special material library. For a more specific step of the present invention, it may be: A. Positioning each device; B. Taking a plurality of solar cells (SQiar(10)) or wafers (Wafer) X-axis (lateral) images superimposed by 25(4) cymbals, The number of steps and mechanisms for adjusting the image can be adjusted. The automatic identification software (3 1) is a kind of automatic image recognition software, and (4) the number of leaf images automatically identifies the number of female elephants. Also, detect other test items, such as "corner angle", "grinding" 4", "feeding status", and detect the different items. For example, "Automatic Identification Software for Notch Image", "Automatic Recognition Software for Grinding Image" (such as the detection status of Figure 4), "Mechanical Software", "Feeding_Automatic Identification Software" (for example: Feed s' The film packaging > {number automatic identification software for the calculation of the feed_package piece 1343m number) and the required comparison database to form an alignment.
•.―而决由於拍攝過程當巾，影像資料若有澤;屯不清，或影像階層 不谷易爱清’本發明亦可以加利用輔助照明器具（1 2)，或者虛 擬打光來強化影像對比。 / I . 至於本發明所使用之系統，請參閱第六騎示，包括·· 一光學影像擷取裝置（1 ): 設於檢測平台⑵之鄰邊；該光學影像娜裝置⑴可以 ，為销光二極親光元件難機，例如電綠合照相機⑽）， 或者互雛氧化金屬半導體（CM0S)照相機為之 操取裝置⑴係輕於-承載桿（11),且於承栽^ =位移砂變其高度。該辟影侧取裝置（1)之本體或 鄰邊設一輔助照明器具（1 2 )。 一檢測平台（2): 與光學影像擷取裝置⑴形成對應；該檢測平台⑵上方 設一載具（2 1)。 、匕；上乃 • 一影像處理單元（3): .連接該光學影細取裝置⑴，該影像處理單元（3 )包括一 -，像自動辨識軟體（31)，—常態影像資料庫（32)以及一 =(，3)，藉由常態影像資料庫(32)與一 由處理益（3 3)比對與_影像意義。 用光學影像擷取敦置⑴操取該太陽能娜— Cell)或其晶圓（wafer)之寻) 料*聲α ，之〜像’可以利用取得之影像部位，然 〜 職軟體（3 1 )對於該娜影像為處理，同時 1343感6 配合各種所需要之制項目，建置標準或常態之影像資料庫（3 •2 ) ’透過麻之影像與鮮料態之影像:#,解㈣，而令彩像 自動辨識軟體（3 1)以及處理器（3 3 )能夠制一個計量值、 異常值或者其訊號’並透過圖形使用者介面_ ( 3 4 )以顯示 結果，而終極達到自動化、效率化之檢測效果。 綜上所述’本發明確實符合產業彻性，且未於申請前見於 ^物或公開個’絲為公眾所知悉，且具有非顯㈣知性，符 5可專利之要件，爰依法提出專利申請。 惟上賴陳，為本發·紅—健實關，舉凡依本發明 請專利範圍所作之均等變化，皆屬本案訴求標的之範脅。 I343S»6 - •【圖式簡單說明】 • 第一圖係本發明之實施例方法流程圖 第二圖係本發明之實際影像擷取以及計量狀態示意圖 ' 第三圖係本發明使用於太陽能電池（Solar Cell)以及其晶圓 (Wafer)之實際影像擷取之狀態示意圖 第四圖係本發明使用於太陽能電池（Solar Cell)磨亮狀態實際 檢測示意圖 第五圖係本發明使用於太陽能電池之晶圓（Wafer)之實際影像擷 取之狀態示意圖 第六圖係本發明之系統立體示意圖 第七圖係本發明之系統平面示意圖 【主要元件符號說明】 (1 ):光學影像擷取裝置 (1 1 ):承載桿 _ (12):輔助照明器具 - (2):檢測平台 (2 1 ):載具 (3 ):影像處理單元 (3 1 ):影像自動辨識軟體 ) (3 2):常態影像資料庫 (3 3):處理器 (3 4 ):圖形使用者介面（GUI)•. ― —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— Compared. / I. As for the system used in the present invention, please refer to the sixth riding instructions, including an optical image capturing device (1): disposed adjacent to the detecting platform (2); the optical image device (1) can be a pin The light-polar two-pole optical component is difficult to operate, such as an electric green camera (10), or the mutual oxidized metal semiconductor (CMOS) camera is operated by a device (1) that is lighter than a carrier rod (11), and is supported by a ^=displacement sand. Change its height. An auxiliary lighting fixture (12) is disposed on the body or adjacent side of the imaging side picking device (1). A detection platform (2): corresponding to the optical image capturing device (1); a carrier (2 1) is disposed above the detection platform (2). , an image processing unit (3): is connected to the optical shadowing device (1), the image processing unit (3) includes a -, like automatic identification software (31), a normal image database (32) And a = (, 3), by the normal image database (32) and a processing benefit (3 3) comparison and _ image meaning. Use optical image to capture Dunhuang (1) to fetch the solar energy - Cell) or its wafer (wafer) to find material * sound α, ~ like 'can use the obtained image parts, then ~ job software (3 1) For the Na Na image processing, at the same time 1343 sense 6 with a variety of required projects, the establishment of a standard or normal image database (3 • 2) 'through the image of the image and fresh material image: #, solution (four), The color image automatic recognition software (3 1) and the processor (3 3 ) can generate a measurement value, an abnormal value or a signal thereof and display the result through the graphical user interface _ ( 3 4 ), and finally achieve automation, Efficiency detection. In summary, the present invention does meet the industry's thoroughness, and is not known to the public or publicly known before the application, and has the knowledge of the public, and has the non-obvious (four) intellectuality, the patentable element of the symbol 5, and the patent application according to law. . However, Shang Lai Chen, the chief of the hair, the red-health, and the equal changes in the patent scope of the invention are all the threats of the subject matter of the case. I343S»6 - • [Simplified illustration of the drawings] • The first diagram is a flowchart of the method of the present invention. The second diagram is a schematic diagram of the actual image capture and measurement state of the present invention. The third diagram is the solar cell used in the present invention. (Solar Cell) and the actual image capturing state of the wafer (Wafer). The fourth drawing is a schematic diagram of the actual detection of the solar cell in the polished state of the solar cell. The fifth figure is used in the solar cell. FIG. 6 is a schematic view of the system of the present invention. The seventh diagram is a schematic diagram of the system of the present invention. [Main component symbol description] (1): Optical image capturing device (1) 1): Carrier bar _ (12): Auxiliary lighting fixture - (2): Detection platform (2 1 ): Vehicle (3): Image processing unit (3 1 ): Image automatic identification software) (3 2): Normal Image Database (3 3): Processor (3 4): Graphical User Interface (GUI)