JP4454434B2 - Image processing method - Google Patents

Description

The present invention relates to a method of manufacturing a display device in which a plurality of display cells such as a plasma display and a liquid crystal display are arranged in a matrix , and image data when inspecting the deposition state of a material deposited on a substrate. the about the image processing how to divide the data corresponding to each display cell.

  FPDs (Flat Panel Displays) such as PDPs (Plasma Display Panels) and LCDs (Liquid Crystal Displays), which are currently in practical use, have cell structures as display pixels on the substrate. A plurality of display cells are arranged in a matrix.

  Hereinafter, a conventional method for manufacturing a PDP will be described as an example of an FPD. First, a scanning electrode, a sustain electrode, a dielectric layer, a black stripe, a color filter, and the like are formed on the surface of a glass substrate to produce a front substrate. On the other hand, a data substrate, a dielectric layer, a partition, a phosphor layer, and the like are formed on the surface of another glass substrate to produce a back substrate. Then, the front substrate and the rear substrate are overlapped so that the direction in which the scan electrode and the sustain electrode extend and the direction in which the data electrode extends are orthogonal to each other, and both the substrates are sealed. Thereafter, the discharge space formed between the front substrate and the rear substrate is evacuated and filled with a discharge gas. Thereby, PDP is produced.

  In the above-described PDP manufacturing method, when the front substrate is manufactured, in the process of forming the scan electrode, the sustain electrode, the black stripe, and the color filter, when the rear substrate is manufactured, the data electrode, the barrier rib, and the fluorescence are formed. In the step of forming the body layer, a matrix or stripe pattern is formed. Hereinafter, a method of forming a matrix-like or stripe-like pattern will be described taking a color filter forming step as an example.

  When forming a color filter for a PDP, red (R), green (cells) are formed in cells partitioned in a grid pattern at each closest point between the discharge region and the data electrode corresponding to the scan electrode and the sustain electrode. A material containing a pigment of G) or blue (B) is applied to the substrate by printing or coating. And after drying this material, it inspects. In general, the color filter pattern is formed of three colors of R (red), G (green), and B (blue) individually. For this reason, it is necessary to perform the deposition process, the drying process, and the inspection process three times each.

  In the scan electrode forming process, the sustain electrode forming process, and the data electrode forming process, these patterns are automatically inspected (see, for example, Patent Document 1). FIG. 8 is a block diagram showing a conventional inspection apparatus. As shown in FIG. 8, in a conventional inspection apparatus, a substrate 101 that is an inspection object is installed, and the substrate 101 is placed on the opposite side of the surface to be inspected, that is, the formation surface of electrodes or the like. A transmissive illumination 102 for irradiating light to 101 is provided, and a light source 103 is connected to the transmissive illumination 102. A CCD camera 104 is provided on the surface to be inspected of the substrate 101. An image processing apparatus 105 is connected to the CCD camera 104. The image processing device 105 is provided with an image storage device 106 that temporarily stores image data, and a processing program 107 that extracts a defective portion by performing processing on the image data.

  Next, an inspection method using the above-described inspection apparatus will be described. When inspecting the pattern formed on the substrate 101, the substrate 101 is irradiated with light by the transmitted illumination 102, and the pattern on the surface of the substrate 101 is photographed by the CCD camera 104. The image of the captured pattern is subjected to image processing in the image processing apparatus 105, and defects are extracted.

  9 and 10 are plan views showing a conventional defect inspection method in the order of steps. As shown in FIG. 9, inspected patterns 111a and 111b such as electrodes are formed on the substrate 101. The inspected pattern 111a has a defect 112a, and the inspected pattern 111b has a defect 113a. Existing. An image processing device 105 shown in FIG. 1 stores image data input from the CCD camera 104 in an image storage device 106 and performs image processing on the image data by a processing program 107. At this time, the image processing apparatus 105 superimposes the comparison pattern 114a on the inspection pattern 111a and also superimposes the comparison pattern 114b on the inspection pattern 111b. For example, the luminance gradation of the corresponding comparison pattern 114a is subtracted from the luminance gradation of the pattern to be inspected 111a to extract a portion where the luminance remains. Further, the luminance gradation of the corresponding comparison pattern 114b is subtracted from the luminance gradation of the pattern to be inspected 111b to extract a portion where the luminance remains.

  Next, as shown in FIG. 10, a portion where the luminance gradation is significantly different between the two superimposed patterns is extracted. At this time, the defect 112a is extracted as a portion 112b having a low gradation, and the defect 113a is extracted as a portion 113b having a high gradation. Thereby, the defect 112a is recognized as a white defect, and the defect 113a is recognized as a black defect. In this way, defects can be detected. As described above, many of the defect inspections in the manufacturing process of the FPD having the cell structure are performed by a method in which the entire substrate is imaged by the line sensor camera and defects are extracted by the adjacent comparison method or the like.

  However, in the color filter forming process, a method for automatically inspecting is not established, and an inspector inspects visually. Since the color filter is disposed at the forefront position with respect to the viewer, any state that gives the viewer a sense of discomfort is a defect. Color filter defects range from color unevenness, color loss, color misregistration, color mixture, and adhesion of foreign matter. In addition, the characteristics of defects to be inspected range from those with a high contrast difference to those with a low contrast difference and a wide range. Up to a wide variety. For this reason, it is difficult to deal with all of these defects by the conventional automatic inspection method described above.

JP 2002-207995 A

  However, the conventional techniques described above have the following problems. Display devices such as LCDs and PDPs, in particular PDPs, are baked at a high temperature a plurality of times in the production process, and substrate shrinkage is inevitably caused to a degree that does not affect the product. For this reason, as disclosed in Patent Document 1, in the image processing method for pattern comparison of image data in a wide range, a shift occurs between the comparison pattern and the actual pattern, and accurate pattern comparison is performed. Is difficult.

  The inspection method described in Patent Document 1 is effective for a fine and high-contrast defect. However, a defect that is not recognized as a defect in a minute region but is recognized as a defect over a wide area is detected. Can not. For this reason, as described above, it is impossible to inspect all the defects to be inspected in the color filter forming process for PDP. Accordingly, the color filter is inspected by the inspector by visual observation. However, the visual observation increases the manufacturing cost due to the labor cost required for the inspection, and the determination is made quantitatively with the determination being varied by the inspector. There is a problem that it is difficult.

The present invention was made in view of the above problems, at the test how that can automatically detect all defects in deposited material on a substrate definitive method of manufacturing a display device, images and an object thereof is to provide an image processing how to divide the data corresponding to data to individual display cells.

Engaging Ru images processing method according to the present invention, the image data of the display device substrate has a plurality of display cells arranged in a matrix, a first arrangement direction of the display cell the gradation of each position of the substrate To obtain a first gradation addition profile corresponding to the position in the second arrangement direction of the display cells intersecting the first arrangement direction, and the gradation at each position on the substrate A step of obtaining a second gradation addition profile corresponding to a position in the first arrangement direction by adding in two arrangement directions, and a standard deviation of a fixed length range for each position in the first gradation addition profile Calculating a first standard deviation profile corresponding to the position in the second arrangement direction, calculating a standard deviation of a fixed length range for each position in the second gradation addition profile, First arrangement A step of obtaining a second standard deviation profile corresponding to a position in the direction; and a coordinate in the second arrangement direction of the display cells as a position having a minimum value in the first standard deviation profile, and the first arrangement The position of the display cell is specified by setting the coordinate in the direction to a position having a minimum value in the second standard deviation profile, and the image data of the substrate is determined based on the specified position of the display cell. Dividing the data into data corresponding to the individual display cells.

  In the present invention, the position of the display cell is specified, and based on this, the image data of the entire display device substrate can be divided into the image data of the individual display cells, so that the defect inspection can be performed for each display cell. Thus, by summing up the results for each display cell, even if each display cell is not recognized as a defect, it is possible to detect a state where it is recognized as a defect when it occurs over a wide range. In addition, since the position of the display cell is specified by obtaining the minimum value of the standard deviation of the gradation addition profile, the display cell and the cell whose size is different from the display cell to be inspected on the substrate Even when cells that are not to be inspected are mixed, only display cells to be inspected can be detected. As a result, even if the substrate contracts, the position of each display cell can be accurately grasped, and an accurate inspection can be performed.

  It is preferable that the method further includes a step of obtaining a boundary line between a display cell formation region and a display cell non-formation region in the substrate based on the first and second gradation addition profiles. As a result, even when the image data includes display cell non-display area data, only the data corresponding to the display cell display area can be extracted and the image data can be divided into data corresponding to individual display cells. it can.

  Further, the fixed length selected in the first gradation addition profile is set as the design length of the display cell in the second arrangement direction, and the fixed length selected in the second gradation addition profile is The design length of the display cell in the first arrangement direction is preferable. Thereby, a display cell can be distinguished from other cell structures, and the position of a display cell can be recognized more correctly.

  According to the present invention, since image data can be accurately divided for each display cell, each display cell can be individually inspected. As a result, all defects can be automatically detected. At this time, since the display cell is recognized by taking the minimum value of the standard deviation, the display cell can be recognized with high accuracy.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the present embodiment, a PDP manufacturing method will be described as a display device manufacturing method. First, the data electrode and the dielectric layer are formed on the surface of the transparent substrate, and then, for example, a cross-shaped partition wall is formed. The partition wall partitions the display cells in a matrix. Then, a phosphor layer is formed on the inner surface of the display cell to produce a rear substrate.

  Meanwhile, a plurality of scan electrodes and sustain electrodes are formed on the surface of another transparent substrate so as to be alternately arranged in parallel with each other. Next, a dielectric layer is formed so as to cover the scan electrode and the sustain electrode, and a cross-shaped black stripe is formed on the dielectric layer. The display cells are partitioned in a matrix by the black stripe. A color filter is formed by forming a pattern on the display cell by individually printing or applying a material containing a pigment of R (red), G (green) or B (blue), and then drying it. Form. In this color filter forming step, the color filter of each color is subjected to application of material by application or printing, drying of the applied material, and inspection of the material after drying. That is, if the color filter has three colors of red, green and blue, the above-described deposition-drying-inspection process is repeated three times. Thereby, a front substrate is produced.

  As a result, the display cells 11 (see FIG. 2) are arranged in a matrix on the surfaces of the front substrate and the rear substrate. The display cells 11 are arranged in two directions orthogonal to each other, that is, in the horizontal direction and the vertical direction. At this time, cells 12 are formed between the display cells 11. The cells 12 are arranged alternately with the display cells 11 in the horizontal and vertical directions, for example, but do not contribute to display.

  Next, the front substrate and the rear substrate thus manufactured are overlapped and sealed, and the discharge space formed between the front substrate and the rear substrate is evacuated, and then the discharge is discharged into the discharge space. Fill with gas. Thereby, PDP is manufactured.

  Hereinafter, the inspection method in the above-described color filter forming step will be described in detail. FIG. 1 is a block diagram showing a color filter inspection apparatus according to the present embodiment, and FIG. 2 is a diagram showing image data of the front substrate of the PDP acquired in the present embodiment and its gradation addition profile. . In this embodiment, a method for inspecting a color filter will be described. However, a scan electrode, a sustain electrode, a step of forming a black stripe in a front substrate manufacturing step, and a data electrode, a partition wall, and a phosphor layer in a rear substrate are formed. In the step of forming, the inspection can be performed by a similar method.

  As shown in FIG. 1, in the inspection apparatus of this embodiment, a conveyor (not shown) for moving the front substrate 1 of the PDP in one direction is provided, and the substrate 1 is coated with a color filter and dried. It is placed on this conveyor so that the finished surface becomes the upper surface. As described above, display cells are arranged in the horizontal direction and the vertical direction on the surface of the substrate 1. A color filter is formed for each column of display cells arranged in the vertical direction. For example, one display cell column in which a red (R) color filter is formed and one column in which a color filter is not formed are repeatedly arranged in this order, or a red (R) color filter is formed. A row in which a green (G) color filter is formed and a row in which no color filter is formed are arranged in this order, or a row in which a red (R) color filter is formed, and green The column in which the color filter (G) is formed and the column in which the blue (B) color filter is formed are repeatedly arranged in this order.

  Two illumination devices 2 for illuminating the inspection target area of the substrate 1 from above are provided above the substrate 1. The illumination device 2 is, for example, a fluorescent lamp, a halogen lamp, an LED (Light Emitting Diode), or the like. Further, a camera 3 that captures an image of the substrate 1 and obtains its image data is provided immediately above the inspection target area of the substrate 1. The camera 3 is a one-dimensional CCD camera, for example.

  An image processing apparatus 4 is connected to the camera 3. The image processing apparatus 4 receives digital image data 21 (see FIG. 2) from the camera 3 and divides the image data 21 into data corresponding to each display cell formed on the substrate 1. In the image processing apparatus 4, a gradation adding unit 5, a boundary line calculating unit 6, a standard deviation calculating unit 7, and a data dividing unit 8 are provided.

  The gradation adding unit 5 receives image data 21 from the camera 3, and in this image data 21, the gradation addition profile 22 corresponding to the position in the horizontal direction by adding the gradation at each position of the substrate 1 in the vertical direction. (See FIG. 2), and the gradation at each position of the substrate 1 is added in the horizontal direction to obtain a gradation addition profile 23 (see FIG. 2) corresponding to the position in the vertical direction.

  The boundary line calculation unit 6 obtains a boundary line 28 between the display cell formation region 26 and the display cell non-formation region 27 in the substrate 1 based on the gradation addition profiles 22 and 23 calculated by the gradation addition unit 5. Is.

  The standard deviation calculation unit 7 calculates a standard deviation in a range corresponding to the design length of the display cell 11 for each position in the horizontal direction at a portion corresponding to the display cell formation region 26 in the gradation addition profile 22, and The standard deviation profile 32 (see FIG. 3) corresponding to the position in the direction is obtained, and the design length of the display cell 11 at each position in the vertical direction in the portion corresponding to the display cell formation region 26 in the gradation addition profile 23. Is calculated to obtain a standard deviation profile (not shown) corresponding to the position in the vertical direction.

  The data dividing unit 8 receives the standard deviation profile calculated by the standard deviation calculating unit 7 and uses the horizontal position where the minimum value is obtained in the horizontal standard deviation profile 32 as the horizontal coordinate of the display cell. The position of the display cell is specified by setting the vertical position where the minimum value is obtained in the standard deviation profile as the vertical coordinate of the display cell, and the image data of the substrate 1 is obtained based on the position of the specified display cell. The data is divided into data corresponding to individual display cells.

  In addition, a determination device 9 is connected to the image processing device 4. The determination device 9 receives data divided for each display cell from the data division unit 8 of the image processing device 4 and investigates the state of the display cell based on this data. Then, the determination device 9 determines the presence or absence of defects based on information indicating the state of each display cell, reconstructs (maps) this information, and based on the number of display cells having a problem and the arrangement thereof, The presence or absence of defects in the substrate 1 is determined.

  Next, the operation of the inspection apparatus according to the present embodiment configured as described above, that is, the substrate inspection method will be described. FIG. 3 is a graph showing a method of obtaining a standard deviation profile from a horizontal gradation addition profile, with the horizontal axis representing the horizontal coordinate and the vertical axis representing the gradation, and FIG. FIG. 6 is a graph showing a horizontal standard deviation profile with horizontal coordinates taken and the vertical axis taken standard deviation.

  First, the substrate 1 after the color filter is formed is placed on a conveyor (not shown) with the surface to be inspected on which the color filter is formed as an upper surface, and is moved relative to the camera 3 by the conveyor at a constant speed. At the same time, the illumination device 2 irradiates the inspection target portion of the substrate 1 with light. And the camera 3 images the board | substrate 1 in response to a conveyor. At this time, the conveyor moves the substrate 1 at a constant speed, and the entire substrate 1 passes directly under the camera 3, thereby imaging the entire surface of the substrate 1 and obtaining image data 21. Then, the camera 3 outputs the image data 21 to the gradation adding unit 5 of the image processing device 4.

  Next, the gradation adding unit 5 adds the gradation at each position of the substrate 1 in the vertical direction with respect to the image data 21 and generates a gradation addition profile 22 corresponding to the position in the horizontal direction for each row of the display cells 11. Ask for. Further, for the image data 21, the gradation at each position on the substrate 1 is added in the horizontal direction, and a gradation addition profile 23 corresponding to the position in the vertical direction is obtained for each column of the display cells 11. At this time, in the gradation addition profiles 22 and 23, the position corresponding to the display cell 11 and the cell 12 becomes a peak portion 24 because the gradation is high, and the portion corresponding to the black stripe is a valley portion because the gradation is low. 25. Further, since the vertical length of the display cell 11 is equal to the vertical length of the cell 12, the peaks 24 are arranged at equal intervals in the vertical gradation addition profile 23, but the horizontal direction of the display cell 11 Is longer than the length of the cell 12 in the horizontal direction, the arrangement of the peaks 24 in the horizontal gradation addition profile 22 is not evenly spaced, and the length of the peaks 24 corresponding to the display cells 11. However, it becomes longer than the length of the peak portion 24 corresponding to the cell 12. Then, the gradation addition unit 5 outputs the gradation addition profiles 22 and 23 to the boundary line calculation unit 6.

  Next, the boundary line calculation unit 6 obtains a boundary line 28 between the display cell formation region 26 and the display cell non-formation region 27 in the substrate 1 based on the gradation addition profiles 22 and 23. That is, the gradation addition profiles 22 and 23 are compared with the threshold value 29, and the position where the gradation of the gradation addition profiles 22 and 23 matches the threshold value 29, that is, the intersection 30 of the gradation addition profiles 22 and 23 and the threshold value 29. Ask for. Next, the boundary line 28 is set by drawing a line segment connecting the intersections 30 adjacent to each other. The inside of the boundary line 28 is a display cell formation region 26 and the outside is a display cell non-formation region 27.

  Next, as shown in FIG. 3, the standard deviation calculation unit 7 corresponds to the design length of the display cell 11 for each position in the horizontal direction in the portion corresponding to the display cell formation region 26 in the gradation addition profile 22. The standard deviation of the range to be calculated is calculated. That is, the standard deviation area 31 is set as a range including the number of pixels corresponding to the design length of the display cell 11. Then, the standard deviation is calculated at each position while the standard deviation area 31 is sequentially moved from the position A to the positions B, C, and D. The standard deviation calculated when the standard deviation area 31 is located at positions A to D is set as the standard deviation at positions A to D, respectively. Thereby, a standard deviation profile 32 as shown in FIG. 4 is obtained. Note that A, B, C, and D shown on the horizontal axis of FIG. 4 are the same as the positions A, B, C, and D shown in FIG.

  As shown in FIGS. 3 and 4, in the standard deviation profile 32, when the standard deviation area 31 is located at the position A, both the top portion and the shoulder portion of the peak portion 24 are included in the standard deviation area 31. Therefore, the standard deviation is slightly higher. When the standard deviation area 31 is located at the position B, since only the top of the peak portion 24 is included in the standard deviation area 31, the gradation in the standard deviation area 31 has a substantially constant value. Becomes lower. Further, when the standard deviation area 31 is located at the position C, the standard deviation is high because the standard deviation area 31 includes one valley 25 and a part of the peaks 24 on both sides thereof. Furthermore, when the standard deviation area 31 is located at the position D, the two valleys 25 in the standard deviation area 31 and the peaks 24 between the valleys 25 and the peaks on both sides of the two valleys 25 are provided. Since a part of 24 is included, the gradation in the standard deviation area 31 varies greatly, and the standard deviation becomes extremely high. On the other hand, in the standard deviation profile in the vertical direction, the standard deviation is relatively low at positions corresponding to the display cells 11 and 12 and the standard deviation is relatively high at other positions. Then, the standard deviation calculating unit 7 outputs this standard deviation profile to the data dividing unit 8.

  Next, the data dividing unit 8 uses the horizontal position B in the horizontal standard deviation profile 32 as the horizontal coordinate of the display cell 11 and takes the minimum value in the vertical standard deviation profile. The position of the display cell 11 is specified by setting the position of the display cell 11 as the vertical coordinate of the display cell 11. At this time, in the standard deviation profile 32 in the horizontal direction, the position corresponding to the cell 12 does not become a minimum value, so that the cell 12 is not erroneously recognized as the display cell 11. In this way, the coordinates of all the display cells 11 formed on the substrate 1 are determined. Based on the specified position, the image data 21 of the substrate 1 is divided into data corresponding to the individual display cells 11. Then, the data dividing unit 8 outputs this data to the determination device 9.

  Next, the determination device 9 investigates the state of the display cell 11 based on this data. The determination device 9 determines the presence / absence of defects in the individual display cells 11 based on the information indicating the state of each display cell, and reconstructs (maps) this information to determine the number of display cells having a problem and the number of display cells. Based on the arrangement, the presence or absence of a wide range of defects is determined. At this time, since the data dividing unit 8 does not recognize the cell 12 that does not contribute to the display as the display cell 11, the determination device 9 does not inspect the cell 12. Hereinafter, this inspection method will be specifically described.

  FIG. 5 is a flowchart showing a cell state determination method. As shown in FIG. 5, in the determination method according to the present embodiment, the states of a plurality of types of display cells are defined by the luminance distribution state, and a plurality of state determinations are performed for each display cell. If the state of the display cell conforms to a certain state (YES), the state and size of the display cell are converted into data. That is, first, as shown in step S101, data corresponding to a certain display cell 11 is prepared. Next, as shown in step S102, state determination 1 is performed, and if it becomes conformity (YES), the process proceeds to step S201 and is converted into [state 1 + size] as data. At this time, the area ratio of specific luminance in one display cell is defined as size. In addition, when the condition determination 1 is non-conforming (NO), the process proceeds to step S103, and the condition determination 2 is performed. Then, when it becomes conformity, the process proceeds to step S202 and is converted into data as [state 2 + size], and when it is unsuitable, the process proceeds to the next state determination (not shown). In this way, after performing (n-1) times (n is a natural number) of state determinations, if all the determinations are not satisfied, the process proceeds to step S104, where state determination n is performed. Then, when it becomes conformity, the process proceeds to step S203, and is converted into [state n + size], and when it is unsuitable, the process proceeds to step S105. Also, when the process proceeds to steps S201 to 203, the process proceeds to step S105. In step S105, the data for each display cell is converted into data representing [state + size]. This process is performed for all display cells.

  In the above-described display cell state determination, each cell is classified into an attached cell and an unattached cell. FIGS. 6A to 6E are schematic diagrams showing states of attached cells, and FIGS. 7A to 7D are schematic diagrams showing states of unattached cells. The attached cells are shown in, for example, an appropriate state 34 shown in FIG. 6A, an overcolored state 35 shown in FIG. 6B, an undercolored state 36 shown in FIG. 6C, and FIG. 6D. They are classified into five types: a color loss state 37 and a foreign substance adhesion state attached 38 shown in FIG. On the other hand, the unattached cell includes, for example, the unprocessed state 39 shown in FIG. 7A, the large color mixture state 40 shown in FIG. 7B, the small color mixture state 41 shown in FIG. 7C, and FIG. It is classified into four types of foreign matter adhesion state empty 42 shown in d).

  Next, the determination device 9 constructs a data array (map) based on the data of each display cell so as to be the same as the array on the substrate 1 based on the coordinate data. That is, mapping is performed. For this map, defect determination is performed for each display cell at a location different from the preset basic map state. Then, whether or not the micro defect is good is determined from the total number of display cells determined as defects in the map and the number of portions in which adjacent display cells are both defective.

  Further, the map is divided into a plurality of blocks including a certain number of display cells. Each block does not become a defect in the above-described micro defect determination, but is in a normal state, that is, a state other than “appropriate 34” shown in FIG. 6A and “unprocessed 39” shown in FIG. If there are more than a certain percentage of display cells, it is determined that a macro defect exists in this block. Then, the determination result of the micro defect determination and the determination result of the macro defect determination are compared with the defect standard, and a comprehensive substrate quality determination is performed. Thereby, the inspection operation is completed.

  Next, the effect of this embodiment will be described. In the present embodiment, by dividing the image data 21 of the substrate 1 into data corresponding to each display cell, both micro defects and macro defects can be determined based on the state of each display cell. As a result, it is possible to easily quantify the defect that gives the viewer a sense of incongruity and to automate the appearance inspection.

  In addition, since the coordinates of the display cell 11 are obtained from the minimum value of the standard deviation profile, even when the display cell 11 to be inspected and the cell 12 not to be inspected are mixed on the substrate 1, the display cell 11 Only can be extracted and inspected. That is, it is conceivable to use the gradation addition profiles 22 and 23 shown in FIG. 2 to identify the peak 24 of the profile as one display cell and specify the coordinates of the display cell 11. Therefore, it is impossible to distinguish between the display cell 11 that performs the inspection and the cell 12 that does not perform the inspection. In the method of using the peak portion 24 of the gradation addition profile as one display cell, the peaks of the peak portion 24 and the valley portion 25 often become unstable due to manufacturing variations. In many cases, it cannot be identified well. In such a case, since the display cell is recognized by the design data, an error due to thermal contraction of the substrate occurs. On the other hand, in the present embodiment, since the coordinates of the display cell 11 are obtained from the minimum value of the standard deviation profile, the display cell 11 can be distinguished from the cell 12.

  Further, the boundary line calculation unit 6 can inspect only the display cell formation region 26 by obtaining the boundary line 28 between the display cell formation region 26 and the display cell non-formation region 27 for the image data 21 of the substrate 1. . For this reason, even when the image data 21 acquired by the camera 3 includes data of the display cell non-formation region 27, the display cell 11 in the display cell formation region 26 can be cut out without using a specific marker. . As a result, it is possible to cope with a plurality of types of PDPs having different substrate layouts.

  In the present embodiment, the example in which the length of the standard deviation area 31 is the design length of the display cell 11 has been shown. However, the present invention is not limited to this, and the length of the standard deviation area 31 is the standard deviation. When the profile is created, the standard deviation profile is set to a value that takes a minimum value at a position corresponding to the display cell 11 to be inspected and does not take a minimum value at a position corresponding to the cell 12 that is not to be inspected. do it.

  In the present embodiment, the method for manufacturing the PDP has been described. However, the present invention is not limited to this, and a flat display device in which a plurality of display cells are arranged in a matrix form may be used. The present invention can also be applied to display devices other than the PDP.

  Furthermore, in this embodiment, although the inspection method of the color filter formed in the transparent substrate was described, the inspection object of the present invention is not limited to this, and the cell on which the coloring material or the coloring material is deposited is applied. Any insulating substrate may be used, and for example, it can be used for inspection of a PDP phosphor layer.

  Furthermore, in the present embodiment, an example is shown in which an illumination device is provided on the upper surface of the substrate, that is, the surface to be inspected, the substrate is illuminated from above, and reflected light is imaged. It may be disposed on the opposite side of the surface and the transmitted light may be imaged.

  The present invention can be used in a method for manufacturing a display device in which a plurality of display cells such as a plasma display and a liquid crystal display are arranged in a matrix.

It is a block diagram which shows the inspection apparatus of the color filter which concerns on embodiment of this invention. It is a figure which shows the image data of the front substrate of PDP acquired in this embodiment, and its gradation addition profile. It is a graph which shows the method of calculating | requiring a standard deviation profile from a horizontal direction gradation addition profile, taking a horizontal coordinate as a horizontal axis and taking a gradation as a vertical axis | shaft. It is a graph which shows the horizontal direction standard deviation profile, taking a horizontal direction coordinate as a horizontal axis and taking a standard deviation as a vertical axis | shaft. It is a flowchart which shows the determination method of a cell state. (A) thru | or (e) is a schematic diagram which shows the state of the adhere | attached cell. (A) thru | or (d) is a schematic diagram which shows the state of an unattached cell. It is a block diagram which shows the conventional inspection apparatus. It is a top view which shows the conventional defect inspection method. It is a top view which shows the conventional defect inspection method, and shows a process next to FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1; Board | substrate 2; Illuminating device 3; Camera 4; Image processing device 5; Gradation addition part 6; Boundary line calculation part 7; Standard deviation calculation part 8; Data division part 9; Image data 22, 23; gradation addition profile 24; peak 25; valley 26; display cell formation area 27; display cell non-formation area 28; boundary line 29; threshold 30; intersection 31; standard deviation area 32; Deviation profile 34; Appropriate state 35; Overcolored state 36; Insufficient color state 37; Color loss state 38; Adhered foreign matter state 39; Untreated state 40; 101; substrate 102; transmitted illumination 103; light source 104; CCD camera 105; image processing device 106; image storage device 107; Turn 112a, 113a; defect 112b, 113b; portions 114a, 114b; comparison pattern A, B, C, D; position

Claims (4)

For image data of a substrate for a display device in which a plurality of display cells are arranged in a matrix, the gradation at each position on the substrate is added to the first arrangement direction of the display cells in the first arrangement direction. Obtaining a first gradation addition profile corresponding to the position of the intersecting display cells in the second arrangement direction; and adding the gradation of each position of the substrate in the second arrangement direction to obtain the first Obtaining a second gradation addition profile corresponding to the position in the arrangement direction, calculating a standard deviation of a fixed length range for each position in the first gradation addition profile, Obtaining a first standard deviation profile corresponding to the position; calculating a standard deviation of a fixed length range for each position in the second gradation addition profile; and calculating a first standard deviation profile corresponding to the position in the first arrangement direction. 2 A step of obtaining a quasi-deviation profile; the coordinates of the display cells in the second arrangement direction as positions where local minimum values are obtained in the first standard deviation profile; and the coordinates in the first arrangement direction as the second standard The step of specifying the position of the display cell by setting the position to take the minimum value in the deviation profile, and the image data of the substrate based on the specified position of the display cell is data corresponding to each of the display cells And an image processing method comprising the steps of: The image processing method according to claim 2 , wherein the second arrangement direction is orthogonal to the first arrangement direction. Based on the first and second gradation summing profile, to claim 1 or 2 characterized by having a step of finding a boundary line between the display cell forming region and the display cell non-formation region in the substrate The image processing method as described. The fixed length selected in the first gradation addition profile is the design length of the display cell in the second arrangement direction, and the fixed length selected in the second gradation addition profile is the first length. the image processing method according to any one of claims 1 to 3, characterized in that the design length of the display cells in the array direction.

Images