TWI807536B - Inspection system and parameter setting method thereof - Google Patents

Inspection system and parameter setting method thereof Download PDF

Info

Publication number
TWI807536B
TWI807536B TW110146853A TW110146853A TWI807536B TW I807536 B TWI807536 B TW I807536B TW 110146853 A TW110146853 A TW 110146853A TW 110146853 A TW110146853 A TW 110146853A TW I807536 B TWI807536 B TW I807536B
Authority
TW
Taiwan
Prior art keywords
parameters
inspection
parameter
panels
found
Prior art date
Application number
TW110146853A
Other languages
Chinese (zh)
Other versions
TW202326115A (en
Inventor
孫培真
李文進
Original Assignee
國立高雄師範大學
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 國立高雄師範大學 filed Critical 國立高雄師範大學
Priority to TW110146853A priority Critical patent/TWI807536B/en
Application granted granted Critical
Publication of TWI807536B publication Critical patent/TWI807536B/en
Publication of TW202326115A publication Critical patent/TW202326115A/en

Links

Images

Landscapes

  • Electrotherapy Devices (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Testing Of Optical Devices Or Fibers (AREA)

Abstract

The present invention provides a parameter setting method of an inspection system. The inspection system includes a processor and an optical detection device. The parameter setting method includes: obtaining multiple object parameters about multiple panels through the optical detection device, and setting multiple inspection parameters; determining whether each panel is defective according to the object parameters and the inspection parameters, and determining whether all defective products can be found from the panels; if all the defective products cannot be found, the inspection parameters will be added with random variables to generate child parameters; and determine whether each panel is defective according to the object parameters and the child parameters, and determine whether all the defective products can be found; output the child parameters if all the defective products can be found.

Description

檢測系統與其參數設定方法Detection system and its parameter setting method

本揭露是關於顯示面板的檢測系統與其參數設定方法。The present disclosure relates to a detection system of a display panel and a parameter setting method thereof.

在面板的生產線中常用自動光學檢查(Automated Optical Inspection，AOI)裝置來判斷面板是否有瑕疵。然而，即使取得關於面板的各種參數，例如亮度、線間距、像素間距等等，還是要根據每一個參數是否超出一臨界值來判斷是否為瑕疵，這些臨界值仰賴人為經驗來設定。此外，對於不同的產品或同一產品不同的批次，可能也要設定不同的臨界值來符合實際檢測的結果。如何自動化的找到這些臨界值，為此領域技術人員所關心的議題。Automated Optical Inspection (AOI) devices are commonly used in the panel production line to determine whether the panel has defects. However, even if various parameters about the panel are obtained, such as brightness, line pitch, pixel pitch, etc., it is still necessary to judge whether it is a defect according to whether each parameter exceeds a threshold, and these thresholds are set by human experience. In addition, for different products or different batches of the same product, different thresholds may also be set to conform to the actual detection results. How to automatically find these critical values is a topic of concern to those skilled in the art.

本揭露的實施例提出一種檢測系統的參數設定方法，此檢測系統包括處理器與光學檢測裝置。參數設定方法由處理器執行，包括：透過光學檢測裝置取得關於多個面板的多個物件參數，並且設定多個檢查參數；根據物件參數與檢查參數判斷每一個面板是否有瑕疵，並且判斷是否可以找到面板中所有的瑕疵品；如果無法找到面板中所有的瑕疵品，將檢查參數加上隨機變數以產生子代參數；以及根據物件參數與子代參數判斷面板是否有瑕疵，並且判斷是否可以找到面板中所有的瑕疵品，若是的則輸出子代參數。Embodiments of the disclosure provide a method for setting parameters of a detection system, and the detection system includes a processor and an optical detection device. The parameter setting method is executed by the processor, including: obtaining multiple object parameters on multiple panels through an optical detection device, and setting multiple inspection parameters; judging whether each panel has defects according to the object parameters and inspection parameters, and judging whether all defective products in the panel can be found; if not all defective products in the panel can be found, adding random variables to the inspection parameters to generate offspring parameters;

在一些實施例中，物件參數包括像素的亮度與像素的間距，檢查參數包括亮度臨界值與間距臨界值。In some embodiments, the object parameters include pixel brightness and pixel pitch, and the inspection parameters include brightness threshold and pitch threshold.

在一些實施例中，隨機變數是對應至物件參數，隨機變數的平均值為0，隨機變數的標準差是相同於對應的物件參數的標準差。In some embodiments, the random variable is corresponding to the object parameter, the mean value of the random variable is 0, and the standard deviation of the random variable is the same as the standard deviation of the corresponding object parameter.

在一些實施例中，參數設定方法還包括：將隨機變數乘上控制因子，此控制因子隨著產生子代參數的次數而遞減。In some embodiments, the parameter setting method further includes: multiplying the random variable by a control factor, and the control factor decreases with the number of generation of offspring parameters.

在一些實施例中，上述判斷是否可以找到面板中所有的瑕疵品的步驟包括：判斷真陽性率與真陰性率是否為1，若是的話判斷已經找到面板中所有的瑕疵品。In some embodiments, the step of determining whether all the defective products in the panel can be found includes: determining whether the true positive rate and the true negative rate are 1, and if so, determining that all the defective products in the panel have been found.

以另外一個角度來說，本揭露的實施例提出一種檢測系統，包括光學檢測裝置與處理器。光學檢測裝置用以取得關於多個面板的多個物件參數。處理器通訊連接至光學檢測系統，用以設定多個檢查參數，根據物件參數與檢查參數判斷面板是否有瑕疵，並且判斷是否可以找到面板中所有的瑕疵品，如果無法找到面板中所有的瑕疵品，將檢查參數加上隨機變數以產生子代參數。處理器還用以根據物件參數與子代參數判斷面板是否有瑕疵，並且判斷是否可以找到面板中所有的瑕疵品，若是的則輸出子代參數。From another point of view, the embodiments of the present disclosure provide a detection system including an optical detection device and a processor. The optical detection device is used for obtaining multiple object parameters related to multiple panels. The processor is connected to the optical inspection system by communication to set multiple inspection parameters, judge whether the panel is defective according to the object parameters and inspection parameters, and judge whether all defective products in the panel can be found, if not all defective products in the panel can be found, add random variables to the inspection parameters to generate offspring parameters. The processor is also used for judging whether the panel is defective according to the object parameter and the descendant parameter, and judging whether all the defective products in the panel can be found, and outputting the descendant parameter if yes.

在一些實施例中，處理器還用以將隨機變數乘上控制因子，控制因子隨著產生子代參數的次數而遞減。In some embodiments, the processor is further configured to multiply the random variable by a control factor, and the control factor decreases with the number of generation parameter generation.

在一些實施例中，處理器還用以在判斷面板是否有瑕疵以後，判斷真陽性率與真陰性率是否為1，若是的話判斷已經找到面板中所有的瑕疵品。In some embodiments, the processor is further configured to determine whether the true positive rate and the true negative rate are 1 after determining whether the panel has defects, and if so, determine that all defective products in the panel have been found.

為讓本發明的上述特徵和優點能更明顯易懂，下文特舉實施例，並配合所附圖式作詳細說明如下。In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail together with the accompanying drawings.

圖1是根據一實施例繪示檢測系統的示意圖。檢測系統100包括了光學檢測裝置110與處理器120。光學檢測裝置110例如為自動光學檢查(Automated Optical Inspection，AOI)裝置，其中包含了感光耦合元件(Charge-coupled Device，CCD)感測器、互補性氧化金屬半導體(Complementary Metal-Oxide Semiconductor)感測器或其他可見光或不可見光的感光元件。處理器120可為中央處理器、微處理器、微控制器、特殊應用積體電路等，處理器120會通訊連接至光學檢測裝置110，此通訊連接可用任意的有線或無線通訊手段來達成。在此是要檢查多個面板131~133是否有瑕疵，這些面板131~133是中間產品。具體來說，面板的製造程序包括了透過沉積、顯影、蝕刻等程序在基板上形成線路與元件，也包括貼上配向膜、上下基板的組立、注入液晶、將大板切割為多個小板、清洗、貼附偏光片、組裝驅動電路等等。本揭露的檢測是在清洗之後並在貼附偏光片之前，但本揭露所提出的方法也可以用在任意的階段，本揭露並不在此限。FIG. 1 is a schematic diagram illustrating a detection system according to an embodiment. The detection system 100 includes an optical detection device 110 and a processor 120 . The optical detection device 110 is, for example, an Automated Optical Inspection (AOI) device, which includes a Charge-coupled Device (CCD) sensor, a Complementary Metal-Oxide Semiconductor (Complementary Metal-Oxide Semiconductor) sensor, or other photosensitive elements for visible or invisible light. The processor 120 can be a central processing unit, a microprocessor, a microcontroller, an application-specific integrated circuit, etc. The processor 120 is communicatively connected to the optical detection device 110 , and this communication connection can be achieved by any wired or wireless communication means. Here, it is necessary to check whether there are defects in the plurality of panels 131-133, and these panels 131-133 are intermediate products. Specifically, the panel manufacturing process includes forming circuits and components on the substrate through deposition, development, etching and other procedures, as well as pasting alignment films, assembling upper and lower substrates, injecting liquid crystals, cutting large panels into multiple small panels, cleaning, attaching polarizers, assembling drive circuits, etc. The detection in the present disclosure is after cleaning and before attaching the polarizer, but the method proposed in the present disclosure can also be used in any stage, and the present disclosure is not limited thereto.

光學檢測裝置110用以擷取面板131~133的影像，從這些影像中取得多個物件參數，這些物件參數例如包括像素的亮度與像素的間距，可參照圖2，圖2繪示了關於面板上像素排列的示意圖。面板上包含了多個像素(例如像素211、212)。在此可以用任意的影像處理方法或機器學習方法來辨識出影像中的像素，並且取得像素之間的間距，例如間距B x、P x、B y、P y等，另外也可以取得線路之間的間距，或者是線路與像素之間的間距。透過感光元件也可以取得每個像素的亮度。本揭露並不限制物件參數的內容。 The optical detection device 110 is used to capture the images of the panels 131-133, and obtain a plurality of object parameters from these images. These object parameters include, for example, the brightness of the pixels and the pitch of the pixels. Refer to FIG. 2, which shows a schematic diagram of the arrangement of pixels on the panels. A plurality of pixels (eg, pixels 211, 212) are included on the panel. Here, any image processing method or machine learning method can be used to identify the pixels in the image, and the distance between pixels can be obtained, such as the distance B x , P x , By y , P y , etc., and the distance between lines, or the distance between lines and pixels can also be obtained. The brightness of each pixel can also be obtained through the photosensitive element. The present disclosure does not limit the content of the object parameters.

另一方面，光學檢測裝置110必須設定多個檢查參數，這些檢查參數是用來判斷上述的物件參數是否異常(有瑕疵)。例如，檢查參數包括了亮度臨界值與間距臨界值，如果某個像素的亮度大於(或小於)亮度臨界值，則這個像素可能是亮點(或暗點)。如果像素或線路之間的間距小於間距臨界值，則可能會發生短路的現象因此會判斷為瑕疵。除了上述瑕疵以外，光學檢測裝置110還可用來檢查成像不均勻(mura)、彩色濾光片上的色斑、液晶分子排列上的錯位、黑色矩陣(black matrix)的破洞、彩色區域短路、微小顆粒或ITO膜破洞、斷線等等，每種瑕疵都有對應的檢查標準，因此也有對應的檢查參數。在一些實施例中，檢查參數還包括光學檢測裝置110的控制參數，例如調整光線的強度等等，本揭露並不限制物件參數與檢查參數的內容。On the other hand, the optical inspection device 110 must set a plurality of inspection parameters, and these inspection parameters are used to determine whether the above object parameters are abnormal (defective). For example, the inspection parameters include a brightness threshold and a distance threshold, and if the brightness of a pixel is greater (or smaller) than the brightness threshold, the pixel may be a bright spot (or a dark spot). If the spacing between pixels or lines is smaller than the spacing threshold, a short circuit may occur and thus be judged as a defect. In addition to the above-mentioned defects, the optical detection device 110 can also be used to inspect imaging unevenness (mura), color spots on the color filter, dislocations in the arrangement of liquid crystal molecules, holes in the black matrix, short circuits in colored areas, tiny particles or holes in the ITO film, broken wires, etc. Each defect has a corresponding inspection standard, so there are also corresponding inspection parameters. In some embodiments, the inspection parameters also include control parameters of the optical detection device 110 , such as adjusting the intensity of light, etc. The present disclosure does not limit the content of the object parameters and inspection parameters.

在經過光學檢測裝置110的檢查之後，每一塊面板還會在後續程序進行測試，例如將每一條訊號線接上測試的訊號以檢查每個像素的顯示是否異常，後續產生的檢查結果可以當作真實結果(ground truth)，用以判斷光學檢測裝置110的檢查是否準確。在此實施例中，處理器120會執行一個參數設定方法，用以對於每一個批次的面板適應性的決定檢查參數，藉此讓光學檢測裝置110可以檢查出所有的瑕疵。After being inspected by the optical detection device 110, each panel will be tested in subsequent procedures, such as connecting each signal line to a test signal to check whether the display of each pixel is abnormal. In this embodiment, the processor 120 executes a parameter setting method for adaptively determining inspection parameters for each batch of panels, so that the optical inspection device 110 can inspect all defects.

圖3是根據一實施例繪示參數設定方法的流程圖。請參照圖3，圖3的左半邊是關於光學檢測裝置110的操作流程，而右半邊是處理器120的計算流程。在步驟311，處理器120產生檢查參數的初始值，在一些實施例中可以用隨機變數來產生初始值，或者也可以用人為經驗產生的預設值。在步驟301，光學檢測裝置110載入這些檢查參數，在步驟302進行檢測，也就是擷取面板的影像。在步驟303，取得面板的物件參數，這些物件參數可以再傳送給處理器120。在步驟312，根據物件參數與檢查參數判斷每一塊面板是否有瑕疵，並且計算錯誤率。由於處理器120可以取得真實輸出，因此也可以計算真陽性(true positive，TP)、偽陽性(false positive，FP)、真陰性(true negative，TN)、偽陰性(false negative，FN)，在此還會計算真陽性率TPR=TP/(TP+FN)、真陰性率TNR=TN/(TN+FP)以及錯誤率 (FP+FN)/(TP+FP+TN+FN)。FIG. 3 is a flowchart illustrating a parameter setting method according to an embodiment. Please refer to FIG. 3 , the left half of FIG. 3 is the operation flow of the optical detection device 110 , and the right half is the calculation flow of the processor 120 . In step 311, the processor 120 generates an initial value of the inspection parameter. In some embodiments, a random variable may be used to generate the initial value, or a preset value generated by human experience may also be used. In step 301, the optical inspection device 110 loads these inspection parameters, and in step 302 performs inspection, that is, captures an image of the panel. In step 303 , the object parameters of the panel are obtained, and these object parameters can be sent to the processor 120 . In step 312, it is judged whether each panel has a defect according to the object parameter and the inspection parameter, and the error rate is calculated. Since the processor 120 can obtain the real output, it can also calculate true positive (true positive, TP), false positive (false positive, FP), true negative (true negative, TN), false negative (false negative, FN), and also calculate the true positive rate TPR=TP/(TP+FN), true negative rate TNR=TN/(TN+FP) and error rate (FP+FN)/(TP+FP+TN+FN) ).

在步驟313，產生子代參數。具體來說，可以將一個隨機變數乘上一個控制因子，再將檢查參數加上隨機變數與控制因子的乘積，可以表示如以下數學式1。 [數學式1] In step 313, descendant parameters are generated. Specifically, a random variable can be multiplied by a control factor, and then the inspection parameter can be added to the product of the random variable and the control factor, which can be expressed as the following mathematical formula 1. [mathematical formula 1]

其中 代表第i個檢查參數的親代， 代表第i個檢查參數的子代。 為隨機變數，此隨機變數的平均值(期望值)為0，每個檢查參數都有對應的物件參數，而隨機變數 的標準差是相同於對應的物件參數的標準差 。在此共有N個檢查參數，N為正整數，因此每個檢查參數對應的隨機變數的標準差都不相同。在產生所有檢查參數的子代參數以後，這些子代參數可以視為一個個體，在一些實施例中可以同時產生多個個體進行平行處理，藉此可以減少演算法所需要的時間。此外，c為控制因子，在一些實施例中控制因子c是隨著產生子代參數的次數而遞減，也就是說控制因子可以當作是更新率，在演算法剛開始執行時採用較大的更新率，接著會逐漸降低更新率。在一些實施例中，上述的控制因子也可以省略。如果檢查參數包括了光學檢測裝置110的控制參數，子代的檢查參數也可以傳送至光學檢測裝置110重新進行檢測。 in represents the parent of the i-th checked parameter, Represents the child of the i-th checked parameter. It is a random variable whose average value (expected value) is 0, each check parameter has a corresponding object parameter, and the random variable The standard deviation of is the same as the standard deviation of the corresponding object parameter . Here, there are N inspection parameters in total, and N is a positive integer, so the standard deviations of the random variables corresponding to each inspection parameter are different. After generating descendant parameters of all checking parameters, these descendant parameters can be regarded as an individual, and in some embodiments, multiple individuals can be generated simultaneously for parallel processing, thereby reducing the time required for the algorithm. In addition, c is a control factor. In some embodiments, the control factor c decreases with the number of generation of offspring parameters. That is to say, the control factor can be regarded as an update rate. When the algorithm starts to execute, a larger update rate is used, and then the update rate is gradually reduced. In some embodiments, the above-mentioned control factors can also be omitted. If the inspection parameters include control parameters of the optical detection device 110 , the inspection parameters of the offspring may also be transmitted to the optical detection device 110 for re-detection.

在步驟314中，計算子代參數的錯誤率，也就是用子代的檢查參數與物件參數來判斷每個面板是否有瑕疵，再根據真實結果(ground truth)計算出對應的錯誤率。In step 314, the error rate of the offspring parameters is calculated, that is, the inspection parameters and object parameters of the offspring are used to determine whether each panel has defects, and then the corresponding error rate is calculated according to the ground truth.

在步驟315中，判斷子代參數是否比較好(錯誤率較低)，如果否的話則捨棄這些子代參數(步驟316)，並回到步驟313產生新的子代參數。如果子代參數比較好，在步驟317中用子代參數取代親代參數。在步驟318中，判斷是否找到面板中的所有瑕疵品，在此實施例中是判斷真陽性率與真陰性率是否都為1，如果不是的話回到步驟313，根據更新後的親代參數繼續產生子代參數。如果真陽性率與真陰性率都為1，這表示找到所有的瑕疵品，在步驟319中結束流程並且輸出目前的子代參數，之後光學檢測裝置110便可以根據輸出的檢查參數進行檢測。In step 315, it is judged whether the offspring parameters are better (lower error rate), if not, discard these offspring parameters (step 316), and return to step 313 to generate new offspring parameters. If the offspring parameters are better, in step 317 the offspring parameters are used to replace the parent parameters. In step 318, it is judged whether all defective products in the panel are found. In this embodiment, it is judged whether the true positive rate and the true negative rate are both 1. If not, return to step 313, and continue to generate offspring parameters according to the updated parent parameters. If both the true positive rate and the true negative rate are 1, it means that all defective products have been found, the process ends in step 319 and the current offspring parameters are output, and then the optical detection device 110 can perform detection according to the output inspection parameters.

根據上述實施例的系統與方法，可以快速地取得檢測參數的設定值，也可以準確的找到瑕疵品。此外，上述逐漸降低控制因子的做法可以再搜尋開始的時候放大參數變動的範圍以在較大的空間中搜尋，可較快速的逼近最佳解，都越靠近最佳解時則在較小的空間中搜尋，避免過大的變動而錯過最佳解。另外，隨機變數 的平均為0可以做為控制突變強度的機制，避免因為大的標準差導致不穩定的震盪現象。 According to the system and method of the above-mentioned embodiments, the setting values of the detection parameters can be quickly obtained, and defective products can also be accurately found. In addition, the above method of gradually reducing the control factor can enlarge the range of parameter changes at the beginning of the search to search in a larger space, and can approach the optimal solution more quickly. When they are closer to the optimal solution, search in a smaller space to avoid missing the optimal solution due to excessive changes. In addition, the random variable The average of 0 can be used as a mechanism to control the intensity of mutations, avoiding unstable oscillations caused by large standard deviations.

雖然本發明已以實施例揭露如上，然其並非用以限定本發明，任何所屬技術領域中具有通常知識者，在不脫離本發明的精神和範圍內，當可作些許的更動與潤飾，故本發明的保護範圍當視後附的申請專利範圍所界定者為準。Although the present invention has been disclosed as above with the embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the technical field may make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be defined by the scope of the appended patent application as the criterion.

100:檢測系統 110:光學檢測裝置 120:處理器 131~133:面板 Bx,Px,By,Py:間距 211,212:像素 301~303,311~319:步驟 100: Detection system 110: Optical detection device 120: Processor 131~133: panel Bx,Px,By,Py: Spacing 211,212: pixels 301~303,311~319: steps

圖1是根據一實施例繪示檢測系統的示意圖。 圖2是根據一實施例繪示關於面板上像素排列的示意圖。 圖3是根據一實施例繪示參數設定方法的流程圖。 FIG. 1 is a schematic diagram illustrating a detection system according to an embodiment. FIG. 2 is a schematic diagram illustrating an arrangement of pixels on a panel according to an embodiment. FIG. 3 is a flowchart illustrating a parameter setting method according to an embodiment.

110:光學檢測裝置 120:處理器 301~303,311~319:步驟 110: Optical detection device 120: Processor 301~303,311~319: steps

Claims (6)

一種檢測系統的參數設定方法，該檢測系統包括一處理器與一光學檢測裝置，該參數設定方法由該處理器執行，該參數設定方法包括：透過該光學檢測裝置取得關於多個面板的多個物件參數，並且設定多個檢查參數；根據該些物件參數與該些檢查參數判斷每一該些面板是否有瑕疵，並且判斷是否可以找到該些面板中所有的瑕疵品；如果無法找到該些面板中所有的該瑕疵品，根據以下數學式產生每一該些檢查參數的一子代參數
Figure 110146853-A0305-02-0011-1
.c,i=1,2,...N其中x i 代表該些檢查參數中的第i個檢查參數的親代，
Figure 110146853-A0305-02-0011-2
代表該第i個檢查參數的該子代參數，a(0,σ i )為一隨機變數，該隨機變數的平均值為0，每一該些檢查參數對應至該些物件參數的其中之一，該隨機變數a(0,σ i )的標準差是相同於對應的該物件參數的標準差σ i ，N為正整數，c為一控制因子，該控制因子隨著產生該子代參數的次數而遞減；以及根據該些物件參數與該些子代參數判斷每一該些面板是否有瑕疵，並且判斷是否可以找到該些面板中所有的該瑕疵品，若是的則輸出該些子代參數。 The parameter setting method of a detection system includes a processor and a optical detection device. The setting method of this parameter is executed by the processor. The parameter setting method includes: to obtain multiple object parameters about multiple panels through the optical detection device, and set multiple check parameters; according to the object parameters to determine whether each panel is defective and judge whether each panel is defective and judge whether the panels are flawed and judge You can find all the defects in this panel; if you cannot find all the defects in the panel, generate each sub -generation parameter of each of the check parameters according to the following mathematical format
Figure 110146853-A0305-02-0011-1
. c,i =1 , 2 , ... N where x i represents the parent of the i-th inspection parameter among these inspection parameters,
Figure 110146853-A0305-02-0011-2
Representing the offspring parameter of the i-th inspection parameter, a (0 , σ i ) is a random variable whose average value is 0, each of the inspection parameters corresponds to one of the object parameters, the standard deviation of the random variable a (0 , σ i ) is the same as the standard deviation σ i of the corresponding object parameter, N is a positive integer, c is a control factor, and the control factor decreases with the number of times the offspring parameter is generated; and according to the object parameters and the object parameters The offspring parameter judges whether each of the panels is defective, and judges whether all the defective products in the panels can be found, and if so, outputs the offspring parameters. 如請求項1所述之參數設定方法，其中該些物件參數包括像素的亮度與像素的間距，該些檢查參數包 括亮度臨界值與間距臨界值。 The parameter setting method as described in Claim 1, wherein the object parameters include pixel brightness and pixel spacing, and the inspection parameters include Including Brightness Threshold and Spacing Threshold. 如請求項1所述之參數設定方法，其中判斷是否可以找到該些面板中所有的瑕疵品的步驟包括：判斷真陽性率與真陰性率是否為1，若是的話判斷已經找到該些面板中所有的該瑕疵品。 The parameter setting method as described in Claim 1, wherein the step of judging whether all the defective products in the panels can be found includes: judging whether the true positive rate and the true negative rate are 1, and if so, judging that all the defective products in the panels have been found. 一種檢測系統，包括：一光學檢測裝置，用以取得關於多個面板的多個物件參數；以及一處理器，通訊連接至該光學檢測系統，用以設定多個檢查參數，根據該些物件參數與該些檢查參數判斷每一該些面板是否有瑕疵，並且判斷是否可以找到該些面板中所有的瑕疵品，如果無法找到該些面板中所有的該瑕疵品，根據以下數學式產生每一該些檢查參數的一子代參數，
Figure 110146853-A0305-02-0012-3
.c,i=1,2,...N其中x i 代表該些檢查參數中的第i個檢查參數的親代，
Figure 110146853-A0305-02-0012-4
代表該第i個檢查參數的該子代參數，a(0,σ i )為一隨機變數，該隨機變數的平均值為0，每一該些檢查參數對應至該些物件參數的其中之一，該隨機變數a(0,σ i )的標準差是相同於對應的該物件參數的標準差σ i ，N為正整數，c為一控制因子，該控制因子隨著產生該子代參數的次數而遞減，其中該處理器還用以根據該些物件參數與該些子代參數判斷每一該些面板是否有瑕疵，並且判斷是否可以找到該 些面板中所有的該瑕疵品，若是的則輸出該些子代參數。 A detection system, comprising: an optical detection device, used to obtain a plurality of object parameters related to a plurality of panels; and a processor, communicatively connected to the optical detection system, for setting a plurality of inspection parameters, judging whether each of the panels has defects according to the object parameters and the inspection parameters, and judging whether all the defective products in the panels can be found, and if not all the defective products in the panels can be found, a descendant parameter of each of the inspection parameters is generated according to the following mathematical formula,
Figure 110146853-A0305-02-0012-3
. c,i =1 , 2 , ... N where x i represents the parent of the i-th inspection parameter among these inspection parameters,
Figure 110146853-A0305-02-0012-4
代表該第i個檢查參數的該子代參數， a (0 ,σ i )為一隨機變數，該隨機變數的平均值為0，每一該些檢查參數對應至該些物件參數的其中之一，該隨機變數a (0 ,σ i )的標準差是相同於對應的該物件參數的標準差σ i ，N為正整數， c為一控制因子，該控制因子隨著產生該子代參數的次數而遞減，其中該處理器還用以根據該些物件參數與該些子代參數判斷每一該些面板是否有瑕疵，並且判斷是否可以找到該些面板中所有的該瑕疵品，若是的則輸出該些子代參數。 如請求項4所述之檢測系統，其中該些物件參數包括像素的亮度與像素的間距，該些檢查參數包括亮度臨界值與間距臨界值。 The inspection system according to claim 4, wherein the object parameters include pixel brightness and pixel pitch, and the inspection parameters include brightness threshold and pitch threshold. 如請求項4所述之檢測系統，其中該處理器還用以在判斷每一該些面板是否有瑕疵以後，判斷真陽性率與真陰性率是否為1，若是的話判斷已經找到該些面板中所有的該瑕疵品。The detection system as described in claim 4, wherein the processor is further used to determine whether the true positive rate and the true negative rate are 1 after determining whether each of the panels has defects, and if so, to determine that all the defective products in the panels have been found.
TW110146853A 2021-12-15 2021-12-15 Inspection system and parameter setting method thereof TWI807536B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW110146853A TWI807536B (en) 2021-12-15 2021-12-15 Inspection system and parameter setting method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW110146853A TWI807536B (en) 2021-12-15 2021-12-15 Inspection system and parameter setting method thereof

Publications (2)

Publication Number Publication Date
TWI807536B true TWI807536B (en) 2023-07-01
TW202326115A TW202326115A (en) 2023-07-01

Family

ID=88147604

Family Applications (1)

Application Number Title Priority Date Filing Date
TW110146853A TWI807536B (en) 2021-12-15 2021-12-15 Inspection system and parameter setting method thereof

Country Status (1)

Country Link
TW (1) TWI807536B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW200844429A (en) * 2007-05-15 2008-11-16 Chi-Hao Yeh An automatic optical inspection approach for detecting and classifying the surface defects on coating brightness enhancement film
US20090306922A1 (en) * 2006-04-27 2009-12-10 Sharp Kabushiki Kaisha Method and System for Classifying Defect Distribution, Method and System for Specifying Causative Equipment, Computer Program and Recording Medium
US20170091671A1 (en) * 2015-09-25 2017-03-30 Canon Kabushiki Kaisha Classifier generation apparatus, classifier generation method, and storage medium
CN107677679A (en) * 2017-09-22 2018-02-09 武汉精测电子技术股份有限公司 Sorting technique and device the defects of L0 pictures in a kind of AOI detection

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090306922A1 (en) * 2006-04-27 2009-12-10 Sharp Kabushiki Kaisha Method and System for Classifying Defect Distribution, Method and System for Specifying Causative Equipment, Computer Program and Recording Medium
TW200844429A (en) * 2007-05-15 2008-11-16 Chi-Hao Yeh An automatic optical inspection approach for detecting and classifying the surface defects on coating brightness enhancement film
US20170091671A1 (en) * 2015-09-25 2017-03-30 Canon Kabushiki Kaisha Classifier generation apparatus, classifier generation method, and storage medium
CN107677679A (en) * 2017-09-22 2018-02-09 武汉精测电子技术股份有限公司 Sorting technique and device the defects of L0 pictures in a kind of AOI detection

Also Published As

Publication number Publication date
TW202326115A (en) 2023-07-01

Similar Documents

Publication Publication Date Title
CN102279189B (en) Defect detecting device, defect restoring device and defect detecting method
TWI665655B (en) Optical compensation apparatus applied to panel and operating method thereof
JP4436810B2 (en) Display multi-angle measuring system and method
TWI758609B (en) Image generation device and image generation method
US10510141B2 (en) Method and apparatus for determining illumination intensity for inspection, and method and apparatus for optical inspection
WO2013078708A1 (en) Automatic optical detection method and automatic optical detection device
KR100842616B1 (en) Method and Apparatus For Detecting Flat Panel Display By Vision Model
JP5645142B2 (en) Unevenness automatic inspection device and inspection method for polarizing plate using color difference analysis
CN112070762A (en) Mura defect detection method and device for liquid crystal panel, storage medium and terminal
CN109712115B (en) Automatic PCB detection method and system
TWI807536B (en) Inspection system and parameter setting method thereof
JP2005283583A (en) System for fault sensing prevention in visual inspection of film or tape shaped printed circuit board, and its processing method
KR101286548B1 (en) System and method for testing liquid crystal display device
CN115546140A (en) Display panel detection method and system and electronic device
CN115167021A (en) Detection method and detection device for display panel
CN113933309A (en) Method for retesting defects in blind hole by AOI machine
TWI788822B (en) Electronic device and method for inspecting defect in display area of display
TWI843045B (en) Image processing and detection method, computer device, and storage medium
KR100566772B1 (en) method for detecting line defects in a light-related plate element for a flat panel
KR20080006202A (en) Apparatus and method of inspecting display substrate
WO2019109408A1 (en) Detection method and device for display panels
CN115225885A (en) Electronic device and method for inspecting defect of display area of display
CN111722422B (en) Liquid crystal panel detection method and liquid crystal panel
TW202331643A (en) Image processing and detection method, computer device, and storage medium
WO2012115003A1 (en) Inspection device and inspection method for liquid crystal panels