TW200921087A - Apparatus for determining defect position of panel - Google Patents

Abstract

The purpose of the present invention is to provide an apparatus which determines a defect position of a panel with high precision. The apparatus 10 of the present invention includes: a holding base 14 of a panel 12; a signal generator 16 for driving all pixels of the panel 12; a camera 18 for photographing the panel 12; a laser pointer 20 for irradiating a laser onto the center of an area to be taken by the camera 18; a biaxial (XY) stage 22 for moving the camera 18 and the laser pointer 20; a scaler (not shown) for determining the position (indicated as the first position as below) of the camera 18 on the biaxial stage 22; a computer 24 for performing various controls or data processing; and a database 26 for storing coordinates of a defective sub-pixel or the like.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for specifying a defect position of a panel in which a liquid crystal panel or the like is arranged in a row and a column. [Prior Art] Inspection is usually carried out during the manufacture of a liquid crystal display (refer to the following patent document). The result of the inspection is repaired if a defect is found. ^ For example, after the manufacture of the liquid crystal cell (LCD CELL), the mounting of the driving π can be checked. The inspection disclosed in the patent is to simultaneously drive all the pixels of the liquid crystal 7L at the same time, and visually find the defects. Image processing is used to specifically locate the defective pixel. Specifically, it is repaired by laser or the like. Patent Document 1 performs a special action of a pixel to move a liquid crystal unit to a different device after finding a defect of the liquid crystal cell. It is a question of whether A, defective pixel specific or repair by laser can be performed correctly. Since the image processing-sentence description is only 'therefore, the algorithm for image processing is unclear' is not clear as to how accurately the defect pixel is specified. Patent Document 2 projects a pointer onto a panel and causes a defect to be directed to the device. When the mouse is clicked, the coordinates of the moment are used as the defective seat. Since the system is performed by visual inspection, it is a question whether the missing coordinates can be correctly specified. The harder it is for $ # ^. ...the fine panel, the defect of the specific patent document 3, the defect of the photographing panel, the lack of the image processing, the lack of 200921087, the patent document, the patent document, the patent document, the patent document 3, the 特本特开平10-325780, 曰本特开2001- Japanese Patent Application No. 9422737 [Abstract] The object of the present invention is to provide an apparatus for trapping a position with good precision. ～ 4 缺 主 主 τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ τ Holding the voltage; applying a voltage to the eve + κ, π * and # through the peripheral portion of the panel: = driving the signal generator of the sub-pixel; for lifting the surface J. a second camera, which is used to illuminate the area of the camera to be photographed, and is mounted on the holding table, having a vertical axis and a horizontal axis, so that the photographic/machine plaque is tender;); raw child # + ζ ι , Prenatal pregnancy, the original version of the camera is multi-shifted on the 2-axis stage above the panel; the middle of the area captured by the East Cinema is used as the first position of the vertical axis and the horizontal axis of the camera. a means for determining a distance between the axis of the peripheral axis of the panel: the means of the target mark at the first position; a means for determining the distance of the image Π:::: the marker 1; and using the method for determining the distance of the f 1 The first distance, the first position is converted into the first and the first ^ proud < the second place of the camera; the camera is photographed to find s missing pixels; the center of the shirt and the extraction The hand 200921087 segment of the second distance of the defective sub-pixel; and the means for using the second position and the second distance to extract the position of the defective sub-pixel in the panel. The panel is placed on the holding table, and the panel is driven by the signal generator. If there is a defect, the display of the panel is taken by the camera. The camera is moved by the light that is incident on the panel. The position where the camera mover is the carrier and the machine is read by the scaler provided on the 2-axis stage. The defect position of the panel is obtained by image processing. The image processing system moves the camera to the target mark provided on the panel to capture the target mark. Extract the target mark from the image. Find the distance between the heart and the target mark (the "large"). From the position of the "large" and the scaler (1st position), 箆 'no) converts the 1st position to the panel The position of the camera (2nd position). After shooting the target mark and correcting the relative position of the panel to the camera, move the camera over the defect to capture the defect. Extract the defective sub-pixel from the captured image. Find the image center and The distance between the defective sub-pixels (the second distance is obtained from the second position and the second distance, and the position of the defective sub-pixel in the panel based on the target mark is obtained. According to the present invention, by improving the accuracy of the scaler or the like, The accuracy of the position of the defective sub-pixel. The image of the defective sub-pixel is obtained by image processing. ^ ν ή The position of the precision pixel of the field. Since the position of the defective sub-pixel is obtained with good precision, the panel can also be improved. The repair accuracy can reduce the probability of panel shipment. [Embodiment] 7 200921087 曰 The embodiment of the device of the present invention is described using a schematic diagram. The panel of the φ board is mounted with a liquid crystal cell. In this description, the same symbol and liquid crystal cell may be used. As shown in the figure, the device of the present invention has a holding table of the panel 12. 14. A signal generator 16 for driving all the pixels of the panel 12, using: a camera 18 of the photographing panel 12, and irradiating a laser to the photographing of the photographing (4). The laser at the center of the field is directed at 1 2 〇 to make the camera 18 and laser pointing::::: two-axis (XY) stage 22 for moving, a scaler (not shown) for determining the position of each of the two pumping stations", "or" The computer 24 for the poor material processing and the database 26 for storing the coordinates of the defective sub-pixels, etc. ', the computer 24 communicates with the cameras 18 and 2. For example, _通1°, and the target t through the system RS232C interface and other access rods. In the operation of the household media player 18, it will be monitored in the future. ^ Φ 4 machine image is displayed on the computer 24 < melon state. Also, computer 24 control, 2 蚰哉 ^ 仃The processing system includes the control of the 2-axis stage 22 of the camera 18, and will be described in detail later. (4) The processing of each pixel is performed in the same shape by red, green, and sub-pixels. In the present description, the order of green 'blue' is arranged in the horizontal direction 夂, the human pixel is red, and the mode is the same, in other orders or The direction, the order of the color arrangement in each material (for example, the red bottle r system., the order of the )) is repeated like the 200921087. The lightness of the (4) _ 12 is when the panel 12 is a liquid crystal cell. If the moonlight. The reason for the state. Also, if it is not liquid 曰Μ - and r ... the first thing is that the driver's liquid can not be confirmed early and the organic EL is the mouth. The slave is not required to be backlit. The illuminating device is as known as A, in the liquid crystal unit 1 2

The Leifu, 兮n door edge has been used to mount the drive 1C at the time of ......also r early 兀*12° signal generator will be applied to the electrode. The electric extension 12 敕锕, the — 人 仰 仰 仰 ’ ’ ’ ’ , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Dark spots are easy to observe when displayed in white, and bright spots are easy to observe when displayed outside of white. Because of the 笠 gr gr heart +, not θ A know the array tester to find: the drive circuit or terminal, so the structure is simple and easy to repair. Photograph: The area photographed by machine 18 is the part of panel 12. The camera 18 of the photographing panel 12 is a CCD or cm-inch camera. For example, 12 one image, the measured field of view is h7 with x2.4_. Zoom in on the panel 1 2 to shoot. When fjt Xin Shizhuang # & Er Tian point qing women's mouse on the computer 24, I: line shooting. Also, it can be taken by operating the keyboard. The image taken by the camera 18 is transmitted to the computer 24. Ray: Pointer 2. The machine side W attached to the camera 18 is photographed from the vertical direction with respect to the panel 12, and the laser light is irradiated to the panel 12 from the oblique direction. Lei... The lateral side of the β machine 18 illuminates the center of the photographing area of the camera 18. The position at which the laser beam L is illuminated is recognized at the center of the area photographed by the camera 18. The operator can easily perform the shooting of the target mark or defect of the panel 12 instead of the laser L, so that the light of any light source can be focused by the lens in the lens 200921087 on the panel [2. Even if the laser beam L is irradiated to the mark so that the defect position is located at the center to capture the port, the image will be described later:::: If the laser irradiation is stopped, the laser 1 will not be irradiated until the mouth 22 has the vertical axis. (Y_2y and the horizontal axis (again, you are attached. The camera 18 and the laser pointer 2 are mounted on the vertical axis 22y. By == vertical axis 22y and camera i" multi-action. Control of linear motor" 曰 from computer 24 The signal is carried out. The invention is remote from the computer 24, and the code _ + A is the finest mouse or keyboard for the field, and the output is used to control the linear motor due to the use of the linear motor', so the movement of the camera 18 and the laser pointer 2 The accuracy is higher. 22x: 2V: The vertical axis % of the 2-axis stage 22 of Fig. 1 is 1 and the horizontal drawing is 2, but the ❹% is 2 and the horizontal axis ❿ At this time: (4) Camera 18 and The laser pointer 20 is mounted on the horizontal axis 疋 branch. The 疋 盗 ,, for example, is a linear scale. When the 2 camera 18 is mounted with the 4 camera 18 moving, the position information is made to the center of the camera 18 Become a scaler: =:=, computer 24 series (four) ^ or the second shot, for the computer 24 in order to find the defect What is being done: The description of the information processed by the computer 24 is from the image taken by the camera and the position data transmitted by the linear scale.: The means are realized by software, hardware, or both. The storage device such as a hard disk provided on the computer 24 can be used to divide the processing of the computer 24 into two items. One item is the processing for finding the reference point of the surface 10 200921087 board U, and the other item is to find the defect. The processing of the position f first describes the processing of the reference point of the face anti-12. In order to obtain the reference point of the panel, the computer 24 has the following function, that is, the target mark is extracted from the image taken by the camera 18. In addition, 11) finds the distance between the image and the extracted target mark (the first segment 32, and the position at the first distance (the first position) / the distance of the brother 1 is used. The means 34 is converted into the position (second position) of the camera. The 卞 panel ^ ^ does not have + 复 at a plurality of portions of the peripheral portion of the panel 12. This is the target mark 36. In the circle 4, at 3 Set near the corner

Si Yu. The target mark 36 is the reference point of the panel 12. By panel 12

38. mi2, the target position has a target head 36, a valid area human pixel, and the like. The relative position of the camera 18 relative to the panel 12 can be found by finding the position of the target P, "where the niobium titanium crucible 30 is on the linear scale". Knowing a number of alpha clamps sets the target mark 36, M ^ 乜 positive panel 1 2 relative line! The position of the photographic plate can be obtained, and finally the center of the two shots can be found to be at the surface of the surface. The means for extracting the target mark 3 6 m , ^ , , , , . ^ ^ 衩%, including the means for binarizing the captured image, the processing of the binary image, the cat stop, the image Hand-held, pre-stored target mark 36 in the form of M έ<« i® ^ rfe size ' 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用The means for extracting the center of the target mark 36 from the portion of the target value of the extracted value is extracted from ^^ Ψ 〇^. Before you binarize the image, you can also go to the screening process. For example, remove the noise generated by shooting in 200921087 or delete unnecessary areas. By binarizing the image, it becomes an image of black and white. When the binary is performed, the target mark is colored in advance to a color different from the color of the panel. In the image 40 of Fig. 5, the panel 12 is black and the target mark 36 is white. Since the image 40 has been binarized, one of the pixels of the image 4 can be labeled "丨" and the other color is labeled ",", ,, . For example, label white as "1" and black as , , 〇, ,. Alternatively, only the color of the target mark 36 may be numbered. Since the target mark 36 uses the same target mark 36 for each type of the panel 12, the shape and size of the target mark 36 are stored in advance. The size or the like stored in advance is the size of the target mark 36 of the image 40 or the like. A portion from the labeled image 40 which has been labeled as "丨,, and satisfies the size of the eye mark 36, etc., can be extracted as the target mark %. For example, Z, like 40, satisfies the vertical 1〇〇 The part of the condition of the 〇 pixel, the horizontal 〇〇〇 pixel, and the line width of 150 pixels + is the target mark %. Therefore, in the image complement, the portion indicated by the symbol 42 is not extracted. The matching panel η = size Etc., the allowable range is set in advance for extracting the size of the target mark 36. For example, an allowable range of 2 to 5 pixels is set in advance. ° "Bu' As in the prior art, one pixel of the image 40 and the panel η 1 pixel are not necessarily It must be consistent. Depending on the structure of the camera 18 or the _ shooting conditions, the value will change with the value. Therefore, according to the panel 2, each object set on the panel 12 is pre-determined in the image 40 ϋ, or distance, and the like. (4) Exist, for example, the size of the desired object is 36 or sub-pixels, from the target mark 36 to the effective area 12 200921087 & This is equivalent. If the target mark % mark 30 can be extracted from the image 40 , . P _ coordinate. The coordinate system of the center, from the longitudinal direction of the image 40 + stalks; ^shenggu a does not show. The coordinates of the direction of the direction of 3 6 are calculated by calculation. For example, the vertical direction will be the big coordinates The minimum coordinates are added and then the heart coordinates are removed. If the target is in the image 40, the target mark 3 = the center of the image 4. The target mark 36: = ILVr) (1 is as follows) The camera 18 used can know the difference in the relative position of the position of the target mark 36 of the i-position disc panel 12 of the (i) position of the linear scale. By == converting the first position to _/18 with respect to the panel 12. = § Ten different positions). After finding the position of the defect ^ (after the 2nd, the linear scale is sent to the younger; the position is converted to the 2nd position) is used to find the position of the defect. The processing of the defect position of the panel 12 is explained. In order to obtain the defect position of the surface 12, the computer 24 has the following functions (Fig.), that is, the method of extracting defective pixels from the p+U image taken by the camera 18. 46, = means of the shirt as the center and the second distance of the extracted defective pixels And the second position and the second distance are used, and the position of the pixel of the mountain is 48. The defect in the panel 12: the camera 18 of the shooting target mark 36 moves to take a defect. By the aid of the L, the defective sub-pixel is photographed. Scope. Camera Η 糸 Hunting by PLC (Programmable L〇gic c〇 nt r: programmable logic 13 200921087 controller) automatically moves to the defect secondary image assist to make the missing image (four) f image like the t # author to the laser The defective sub-pixel 52 of L will definitely become; however, as shown in Fig. 8, the position of the position 5 读出 read by the sex scale is shown. Since the line must find the position of the defective sub-pixel 52〇*520, it is configured. In addition, the m2R two-two reaches the sub-pixel 52G to display the green sub-pixel, the human pixel, and the sub-pixel. The sub-pixel 52B is displayed. The image of the 5G image taken by the blue self-examination. From the _##5〇, the defect is automatically extracted like the symplectic 52d and the image 50

Passing the method of ^ _ _ 仙 仙 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 , 52G, 52B is the means to extract the range of defects; pre-stored the information of the shape of the secondary image 52R qr · d, the shape of the 2B, extracting all the sub-pixels 52R, 52G, 52B to obtain each pixel training , the method of fine image data; find the difference between the image f material obtained by the sub-pixel (10), 52G, 52B of the sub-pixel (52) which is located at the pixel of the sub-pixel of the pixel. And means for determining that the defective sub-pixel 52 is defective when the determined difference exceeds a threshold value stored in advance. The minimum coordinates of the sub-pixels 52R, 52G, 52B of the image are not exceeded, and the upper left coordinate A is shown in FIG. Further, the maximum coordinate is the lower right coordinate Ββ, and the average value of the image data of the entire image is a gradation value. The image 50 is binarized with a grading value as a threshold. Pre-stored secondary image 14 200921087 Prime 52R, 52G, 52B 2 ·») «tt , B shape. Size, from the image 5 sequence search and sub-pixel 52R, plus, Μ from the top left, for example, the coordinates on the image (1) Miscellaneous. The linear scale of the third position can also be converted to panel 12, depending on the type of panel U, the second like Xin Zhidi 2 position. I. Your Department 垚 a often 2R, 52G The shape of 52B. The size position is constant. The element (4) can be converted into the sub-image 52 of the seat (4) 〇, 1 〇 on the coordinate of the panel 12. :: The image is 5° from the bottom right and looks for the shape and size of the sub-pixel feed. For example, the coordinates (620, 470) on the image are the largest coordinates. The seat of the 冢 之 之 之 ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( Images that can be included in the minimum coordinates and the largest coordinates are detected defects

range. Located in the 50th week of the image, Qiu AA

The Bureau of the Edge. The sub-pixels 52R, 52G, 52B that are missing are not included in the detection defect range. The sub-pixels 52R, 52G, 52B that are partially deficient due to the right are included in the horse by m + ,

When the image processing is performed, the sub-pixels 52R and 52G 52B are judged to be defective. ‘If you have decided to extract the range of defects, you will extract all the secondary image from the image 5,, 52G, MB, and obtain the image information of the sub-pixel 5, 52匕 (4). Because the minimum coordinates are known,

^ Take the shape of the large coordinates, sub-pixels 52R, 52G, 52B. The size and the size of the black matrix are 'prematurely Φ shape. Therefore, the sub-pixel rule is extracted from the range determined according to the size and the like, for example, 52B. For example, from the top left coordinate A to the lower right, the pressure heart is sequentially extracted according to the size of the sub-pixels 52R, 52G, 52Β, etc., and the second pixel '52 pixels 52H, 52G, 52Β are extracted, and the extracted is obtained. Image data of the location. Image data contains deleted values. When pattern matching is performed from 15 200921087 to 50, the sub-pixel 520 having a dark dot cannot be extracted, but the method of the present invention can be extracted. The difference between the sub-pixels 52R, 52G located in the - pixel and the sub-pixels 52R, 52g, and MB of the same color displayed in the pixel is obtained. In Fig. 9, the sub-pixel 52R1 of the red pixel of the - pixel can grow the difference between the red sub-pixel 52R2 of the surrounding pixel. When the difference knife exceeds the threshold value stored in advance, it is determined as the defective sub-pixel 52D. When the difference is obtained, the image with the highest value in the image of the surrounding pixel f MR2 (4) is removed, and the lowest image material is removed. Further, the image data of the sub-pixel 52D determined to be defective is also removed. 2 When the difference is obtained, the extreme image data can be removed to make a correct judgment. There is no need to store absolute image data in advance. Although the absolute value of the brightness of the panel 12 will be the same, it can be checked without considering the difference. # Find the distance between the center 影像 of the image 50 and the second sub-pixel 52D of the extracted defect (Fig. 1A). Using the second position, the second distance, and the imaging magnification of the camera 18, etc. Pixel plus position. Since the size of the panel 12 or the size of the sub-pixels 52R, 52G, 52B has been previously determined, the position of the defective sub-pixel 52D can be converted into a coordinate at the panel 。. The coordinates are corrected when the conversion is made. In order to correct the coordinates of the panel, the means 48 for determining the position of the defective sub-pixel includes means for converting the obtained position into the coordinates of the sub-pixels 52R, 52G, 52B of the panel ;; The coordinates of the sub-pixels 52R, 52G, and 52B are divided by the sub-pixels of the pixels, 52 feet, and the means of the color number of 16 200921087 52B; the decision is made according to the color of the pixels 52R, 52G, 52B to complete the division operation. The remainder is compared with the obtained color according to the color of 52G and 52B of the sub-pixels of the surrounding pixels, and the comparison is made with the obtained color, and the guest s A # , . If the colors are not right, the coordinates are corrected to the coordinates of the image data and the coordinates of the roots of the child's fh that are the same as the color of the remainder. According to the panel 12, the arrangement of the colors of the 52I 52R, 52G, and 52B is determined. For example, it is arranged in the order of red, 、, and I in the pixel. The entire color of the panel 12 is repeated in the entire system. therefore,

The mouth-to-person pixel 52R, 52G, 52B X coordinate (horizontal direction) divided by the number of subtle pixels 52R, 52G, 52Β color number is red, 1, green is 2, blue sound Α Λ 4th Pumping y _ I bar is 0. Based on the image data of the sub-pixels 52R h . , 2B of the pixels located around the defect coordinates, it is determined which color is missing from the defective sub-pixel 52D, and compared with the color known from the remainder of the division. If both are 4s Γ = 1 έ ε A. I u Feng the same color does not need to be corrected. If the two are inconsistent, they will be stunned according to the defect. s) «The color of the humanoid 52D will be corrected to the color - the coordinates next to it. The accuracy of the coordinates obtained can be improved. Further, when the defective sub-pixel 52D is judged, the sub-pixel 52R2' of the surrounding pixel used is determined to be unable to determine the correct color based on the image data of the sub-pixel which is determined to be defective. The obtained defect is the data of the coordinates of the 次 Λ + + + tstt 曰 像素 像素 52D is stored in the database 26 and the like. Use this information to repair the τ + 面板 panel 1 2 . Alternatively, the data can be transferred directly to the repair unit. The procedure for determining the position of the defective sub-pixel 5 using the above-described apparatus will be described. 17 200921087 (1) The panel 12 to be mounted is placed on all the sub-pixels 52R, 52G of the holding stage u, and is thinned to visually perform the defect inspection panel. The defect is sent to the next process. '-Right..., if the defect is seen on the moving face, the mouse of the computer 24 is operated to cause the camera 18 to shoot above the target mark 36 of the board 12. Shoot when the target mark 36 torpedo strikes L-induced. At least one percent of the panel 12 is taken, and all of the target marks 36 are taken. - Target mark (take the image taken 4G to extract the center of the target mark 36, Hummer you ^ 丄 ^ ma ^ ^ Continued... The distance from the center of the 4 。. Find the relative of the panel 12, the relative degree of the spring Position: (:) Move the camera 18 over the defect to shoot the defect. At the time of shooting, the position is determined by the linear scale reading ^ $ ^ w. The position of the defect is not the position of the defect. In the range, the coordinates of the defective sub-pixel 52D are obtained by extracting the missing image. The sub-pixels 52r, 52G, 52b located at the partial defect of = are detected from the defect: =, to improve the detection accuracy, and the same 5 进入 is entered into the detection range. , 52B, compares the image data of the sub-pixel 52R1 and the sub-pixel 52R2 located around. The difference between the data is judged as a defect. (6) The panel I2 is sent to the repair process, and the repair is performed. Find the defective sub-pixel to perform the repair of the defect 52D.

In the above manner, the defective sub-pixel 52D is processed by various data of the present invention to improve the accuracy of the position sought. With this high precision, it can also be greatly increased. 18 200921087 Reduces the concern of laser cutting off the wrong sub-pixel electrode during laser repair. Although the embodiments of the present invention have been described above, the present invention is not limited to the embodiments. For example, if the right is checked when the defect is checked, the defect to be detected is determined. Therefore, the method of finding and -... can not find the coordinates of the shirt like the binary ‘...沄 find the human pixel 52R, 52G, 52B. Therefore, the position of the linear scale is converted into the position of the panel 12 by using the target mark 36 to find the reference.

Class, sub-pixels 52R, 52G, 52B are certain. The image of the 50 夂 夂 庐 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - t pixel (four), gamma, the top left of the coffee. If the sub-pixels 52R, 52 = up, then the sub-pixels 52r, 52g, 52b of the minimum coordinates can be defined. No A (four) mode can also be derived from the bottom right of the image in order to determine whether it is the person: plain coffee, such as, 523 right lower. If it is a sub-pixel coffee, the lower right side of = can be defined as the sub-pixel 52r, 52g, 52b of the largest coordinate. = The image contained in the smallest coordinate and the largest coordinate is the defect to be detected =: Even when the black display is performed, the blue 1匕1 of the defect to be detected can be obtained. If the black display is performed, the classification value is Find and compare the car to the parent. If it is below the threshold, it is judged to be black. Next, the region was obtained by the above method. When the predetermined values are extracted from the pixels 52R, 52G, and 52B, all of the extracted sub-pixels 52R, 52G, and 52B can be obtained by the processing of the defective sub-pixel 52D in the subsequent processing of the sub-pixels 52R, 52G, and 52B of the panel 12. You can find the coordinates of the panel 12 at 19 200921087. In the above description, the position, the distance, and the coordinates are variously converted, but the type used for the conversion is not limited. As long as the conversion can be performed as described, the person skilled in the art can appropriately change it. In addition, the present invention can be applied in various forms of improvement, correction, and modification in accordance with the knowledge of the industry within the scope of the gist of the invention. For example, the present invention can also be applied to an organic electroluminescent display or Panel inspection of plasma display. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a view showing the construction of a device for a specific defect position of the present invention. Fig. 2 shows the configuration of a camera and a laser pointer. Fig. 3 shows the configuration for determining the reference point of the panel. Figure 4 is an example of a panel. Fig. 5 is a view showing an image processing of a target mark. Fig. 6 shows the first distance. Fig. 7 shows the configuration for determining the position of the defect. Fig. 8 is a view showing an example of a defect of the photographing panel. Fig. 9 is a view showing the difference between the sub-pixel of one pixel and the sub-pixel of the surrounding pixel. Fig. 10 shows the second distance. [Main component symbol description] 20 200921087 ίο Device with specific defect position 12 Panel

14 16 18 20 22 24 26 30 32 34 36 38 40, 50 44 46 48 52R, 52G, 52B 52D Holder signal generator camera laser pointer 2 axis carrier computer database to extract the target mark to find the first The means for calculating the distance from the first position to the second position means the target mark effective area image extracting the defective sub-pixel means obtaining the second distance means determining the position of the defective sub-pixel by the sub-pixel defective sub-pixel 21

Claims (1)

200921087 X. Patent application: [Special device for defect location] In order to specify the defect position of the panel of the display device arranged by a plurality of secondary images and pixels, the panel includes: a holding plate of the panel; An electrode disposed at a peripheral portion of the panel, a signal generator for driving the element; a light source for photographing the portion of the panel _ = a light source that illuminates the center of the region captured by the camera; and a panel and a camera and a light source u Find the center of the area photographed by the camera with your helmet & m + the vertical axis and the birch axis um or medium ~ as the scaler of the camera in the first position of the enamel; set on the periphery of the panel There is a record, the image taken from the camera. The target of the reference position of the plate: the target mark is obtained; the center of the image and the pomelo means are owed; /, the use of the first distance of the target mark When the ith distance is obtained, the first position is converted to the first position, and the first position and the first distance are obtained, and the image taken from the second position of the panel is extracted; the image captured by the pain is extracted to obtain the image. Hand &, means the center of the mountain m pixels; and a second defective sub-pixel distance means of using the second position and from the second position. 'Defective sub-pixels in the panel 22 200921087 2. In the special position of the defect position of the i-th patent application scope, the means for extracting the target mark includes: , ^ means for binarizing the captured image; The method for labeling the processed image; the processed image in which the shape and size of the target mark are stored in advance is extracted as a target segment; and the target mark extracted from the hand of the private note The specific means for determining the defect position of the first or second range of the center of the target mark, the means for extracting the defect includes: finding the minimum coordinate of the sub-pixel that does not exceed the image and including the minimum coordinate Take a big seat, and use the range of means to not take the sub-pixel of the big coordinate as the pre-stored sub-picture sound of the defect to be extracted. ## &&&&&&&&&&&& And the means for obtaining the image data of each sub-pixel; the difference between the image data obtained by the sub-pixel located in the pixel and the sub-pixel located around it: the sub-pixel of the color The means of:: the difference between the member and the non-pixel = the difference between the pre-stored values is determined as the defect, and the difference is: the specific device of the defective position of the patented three brothers, the image of which is The second method is to eliminate the surrounding image data before the difference is obtained, and the lowest image data in the image data of the human pixel. Take the threshold 诼 诼 诼 与 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 A specific device for a defect position, wherein the means for determining the position of the defective sub-pixel includes: means for converting the obtained position into a coordinate of a sub-pixel of the panel, and dividing the obtained sub-pixel coordinate by a pixel Means of the number of colors of the sub-pixels included; 'Determining the remainder after dividing the colors of the successive pixels, dividing the color of the image data of the sub-pixels of the surrounding pixels, and performing the de-calculation The color obtained by comparing the remainder is used as a means for comparison; and if the color of the comparison result is different, the coordinate is corrected to the color of the image and the color of the wall corresponding to the color obtained by the remainder ~ Ten, FIG formula: such as hypophosphorous page 24