JP2007192625A - Flaw discrimination method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a flaw discrimination method which enables a discrimination especially between a streak flaw and a stain flaw in the displayed flaw, which is formed on an image display device, with high precision. <P>SOLUTION: In the flaw discrimination method, the inertial main axis (a) of a displayed flaw is calculated and the circumscribed rectangle, which is circumscribed about the displayed flaw is set by two circumscribed short side having a length A vertical to the inertial main axis (a) and circumscribed about the displayed flaw and two circumscribed long sides having a length B parallel to the inertial main axis (a) and circumscribed about the displayed flaw. A predetermined first threshold value α and a predetermined second threshold value β are respectively compared with the aspect ratio B/A of the circumscribed rectangle and the reciprocal A*B/S of the occupying ratio of the display flaw (area S) in the circumscribed rectangle (area A*B) and the kind of the displayed flaw is discriminated on the basis of the comparison result. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a defect determination method and a defect determination apparatus for determining the type of display defect of an image display apparatus.

  Conventionally, a quality inspection of an image display apparatus that has been performed by relying on human vision has been automatically performed by a machine in recent years. However, automatic inspection by machine has not yet achieved inspection accuracy comparable to human vision, and various improvements have been attempted even now.

  For example, paying attention to the fact that human visual perception differs depending on the type of display defect (for example, streak defect, spot defect, point defect) that occurs in the image display device, each of the relevant defects prior to the inspection of each display defect A type of display defect is determined (for example, see Patent Document 1). Then, for each type of display defect, an optimal inspection process taking into account human visual characteristics is set in advance, and an optimal inspection process according to the display defect type determination result is applied to the display defect. By doing so, attempts have been made to inspect display defects with the same degree of accuracy as human vision.

  The defect determination method of Patent Document 1 is a method of determining the type of each display defect included in an image by repeatedly performing expansion / contraction processing on the entire image, and the image includes a plurality of display defects. Even so, the expansion / contraction process may be performed collectively on the entire image, so that the processing can be speeded up.

Japanese Patent Laid-Open No. 9-288037

  However, on the contrary, in the defect determination method of Patent Document 1, since the types of individual display defects included in the image are determined based on the expansion / contraction process performed on the entire image, the individual inevitably The discrimination accuracy for the display defect must be reduced. Therefore, there is a problem that the type of the display defect that is difficult to discriminate is erroneously discriminated, and as a result, the inspection accuracy is lowered.

  An object of the present invention is to provide a defect discriminating method and a defect discriminating apparatus capable of discriminating, in particular, a streak defect or a spot defect among display defects generated in an image display device with high accuracy.

  The defect determination method of the present invention is a defect determination method for determining the type of display defect of an image display device, wherein a defect extraction step of extracting a display defect from data of an image displayed by the image display device, and the display An inertia main axis calculating step for calculating an inertia main axis of the defect; two circumscribed short sides that are perpendicular to the inertia main axis and circumscribe the display defect; and two circumscribed long sides that are parallel to the inertia main axis and circumscribe the display defect; The circumscribed rectangle setting step for setting a circumscribed rectangle circumscribing the display defect, the aspect ratio of the circumscribed rectangle, and the display defect occupation ratio in the circumscribed rectangle are determined. A defect determination step.

According to the defect determination method of the present invention as described above, the main axis of inertia is calculated along the direction in which the display defect to be determined extends, and the display defect is determined using the circumscribed rectangle set along the main axis of inertia. Can be discriminated with high accuracy. Specifically, the type of display defect is determined based on the aspect ratio of the circumscribed rectangle and the display defect occupation ratio in the circumscribed rectangle. For example, when the aspect ratio of the circumscribed rectangle is large and the display defect occupancy ratio in the circumscribed rectangle is small, it is determined that the display defect is a streak defect, the aspect ratio of the circumscribed rectangle is small, and the display defect in the circumscribed rectangle is small. When the occupation ratio is large, the display defect is determined to be a spot defect, and otherwise, the display defect is determined to be a spot defect and a type of display defect other than the spot defect.
Note that, in the defect determination method of the present invention, unlike the defect determination method of Patent Document 1 in which determination processing (expansion / contraction processing) is performed on the entire image, the determination processing is performed for each display defect to be determined. Therefore, it is possible to improve the discrimination accuracy for each display defect.

  In the defect determination method of the present invention, a threshold value setting step for setting a threshold value α1 (> 1) and a threshold value β1 (> 1) for determining the type of the display defect is provided. In the defect determination step, the circumscribed short The length of the side is A, the length of the circumscribed long side is B, the area of the display defect is S, the aspect ratio B / A of the circumscribed rectangle, and the reciprocal of the occupation ratio of the display defect in the circumscribed rectangle A · B / S is compared with the threshold value α1 and the threshold value β1, respectively. If at least α1 ≦ B / A and β1 ≦ A · B / S are satisfied, the display defect is It is preferable to determine that it is a streak defect.

  According to such a defect determination method of the present invention, it is possible to determine, in particular, a line defect with high accuracy among display defects generated in the image display device based on the threshold value α1 and the threshold value β1. Here, the threshold value α1 is a threshold value for the aspect ratio B / A (≧ 1) of the circumscribed rectangle. For example, when the display defect is elongated and the aspect ratio B / A of the circumscribed rectangle is large, α1 ≦ B / A and the discriminant by the threshold α1 is satisfied. The threshold value β1 is a threshold value for the reciprocal number A · B / S (≧ 1) of the display defect (area S) occupation ratio in the circumscribed rectangle (area A · B). For example, when the display defect occupancy ratio in the circumscribed rectangle is small and the reciprocal number A · B / S is large, β1 ≦ A · B / S is satisfied and the discriminant by the threshold value β1 is satisfied. In the defect determination method of the present invention, it is determined that the display defect is a streak defect when at least both of the discriminants are satisfied. Here, the information indicating that the aspect ratio of the circumscribed rectangle of the display defect is large obtained by satisfying the discriminant by the threshold α1, and the display defect in the circumscribed rectangle obtained by satisfying the discriminant by the threshold β1. The information that the occupation ratio is small is information that suggests that the display defect has a streak shape. Therefore, according to the threshold value α1 and the threshold value β1, the streak defect can be determined with high accuracy.

  In the defect determination method of the present invention, in the threshold setting step, a threshold γ1 larger than the threshold α1 is set. In the defect determination step, first, the aspect ratio B / A of the circumscribed rectangle is compared with the threshold γ1. When γ1 ≦ B / A is satisfied, it is preferable to determine that the display defect is a streak defect without performing defect determination based on the threshold value α1 and the threshold value β1.

  In such a defect determination method of the present invention, the aspect ratio B / A of the circumscribed rectangle of the display defect is larger than the threshold value γ1 larger than the threshold value α1, and it is suggested that the display defect extends longer than a predetermined level. In this case, the determination based on the threshold value α1 and the threshold value β1 is omitted, and it is determined that the display defect is a streak defect. As described above, if the discriminant expression based on the threshold value γ1 is satisfied, the discrimination process can be completed by omitting the discrimination based on the threshold value α1 and the threshold value β1, so that the process can be speeded up.

  In the defect determination method of the present invention, a threshold value setting step for setting a threshold value α2 (> 1) and a threshold value β2 (> 1) for determining the type of the display defect is provided. The length of the side is A, the length of the circumscribed long side is B, the area of the display defect is S, the aspect ratio B / A of the circumscribed rectangle, and the reciprocal of the occupation ratio of the display defect in the circumscribed rectangle A · B / S is compared with the threshold value α2 and the threshold value β2, respectively, and if at least α2> B / A and β2> A · B / S are satisfied, the display defect is stained. It is preferable to determine that it is a spot-like spot defect.

  According to the defect determination method of the present invention as described above, it is possible to particularly determine a spot defect among display defects generated in the image display device based on the threshold value α2 and the threshold value β2. Here, the discriminant based on the threshold α2 and the discriminant based on the threshold β2 reverse the direction of the inequality sign of the discriminant based on the threshold α1 and the discriminant based on the threshold β1 described above with respect to the defect discriminating method of the present invention for discriminating streak defects. It is a thing. Therefore, in contrast to the defect determination method of the present invention for determining a streak defect, when the aspect ratio B / A of the circumscribed rectangle is small, the determination formula based on the threshold α2 is satisfied, and the occupation ratio of the display defect in the circumscribed rectangle is When it is larger, the discriminant by the threshold value β2 is satisfied. Then, the information indicating that the aspect ratio of the circumscribed rectangle of the display defect is small obtained from satisfying the discriminant by the threshold α2, and the occupation of display defects in the circumscribed rectangle obtained from satisfying the discriminant by the threshold β2. The information that the ratio is large is information that suggests that the display defect has a spot shape. Therefore, according to the threshold value α2 and the threshold value β2, the spot defect can be determined with high accuracy.

  In the defect determination method of the present invention, it is preferable that the threshold α1 is set to a value larger than 2 in the threshold setting step. Furthermore, it is more preferable if the threshold value α1 is set to a value larger than 2.5.

  In the defect determination method of the present invention, it is preferable that the threshold value β1 is set to a value larger than 2.5 in the threshold setting step. Furthermore, it is more preferable that the threshold value β1 is set to a value larger than 3.

  In the defect determination method of the present invention, it is preferable that the threshold α2 is set to a value larger than 2 in the threshold setting step. Furthermore, it is more preferable that the threshold α2 is set to a value larger than 2.5.

  In the defect determination method of the present invention, it is preferable that the threshold β2 is set to a value larger than 2.5 in the threshold setting step. Furthermore, it is more preferable that the threshold value β2 is set to a value larger than 3.

  The defect discriminating apparatus of the present invention is a defect discriminating apparatus that discriminates the type of display defect of the image display device, the defect extracting means for extracting a display defect from image data displayed by the image display device, and the display An inertia main axis calculating means for calculating an inertia main axis of the defect; two circumscribed short sides that are perpendicular to the inertia main axis and circumscribe the display defect; and two circumscribed long sides that are parallel to the inertia main axis and circumscribe the display defect; Based on the circumscribed rectangle setting means for setting a circumscribed rectangle circumscribing the display defect, the aspect ratio of the circumscribed rectangle, and the occupation ratio of the display defect in the circumscribed rectangle, the type of the display defect is determined. And a defect discriminating means.

  Since each defect discriminating apparatus according to the present invention has a configuration for carrying out each defect discriminating method of the present invention described above, the same operations and effects as the respective defect discriminating methods can be achieved.

  Next, embodiments of the present invention will be described with reference to the drawings.

(1) Configuration of Defect Determination Device FIG. 1 is a functional block diagram showing the configuration of the defect determination device.
The defect determination device is a device that determines the type of display defect of the image display device 1, and an image pickup unit 2 that picks up an image displayed by the image display device 1 and a determination control that performs general control in defect determination. Part 3.

The image display device 1 is a device that displays an image based on an image signal input from the display control unit 11, and is placed on an imaging table 12 by the imaging unit 2 when determining a defect.
The imaging unit 2 captures an image displayed by the image display device 1 with a large number of imaging pixels (CCD or the like) arranged in an alignment on the imaging surface.
The discrimination control unit 3 includes an image data acquisition unit 31, a defect extraction unit 32, a centroid calculation unit 33, an inertia main axis calculation unit 34, a circumscribed rectangle setting unit 35, a threshold setting unit 36, and a defect determination unit 37. It is configured with.

The image data acquisition unit 31 is a unit that displays a desired inspection image on the image display device 1 via the display control unit 11 and acquires the image data by causing the imaging unit 2 to capture the inspection image.
The defect extraction unit 32 is a unit that extracts display defects included in the image data acquired by the image data acquisition unit 31.
The centroid calculating unit 33 is a unit that calculates the centroid of the display defect extracted by the defect extracting unit 32.
The inertia main axis calculation means 34 is a means for calculating the inertia main axis of the display defect extracted by the defect extraction means 32.
The circumscribed rectangle setting means 35 includes two circumscribed short sides that are perpendicular to the inertia principal axis of the display defect calculated by the inertia principal axis calculation means 34 and circumscribe the display defect, and 2 that circumscribe the display defect parallel to the inertia principal axis. A circumscribed rectangle that circumscribes the display defect is set by two circumscribed long sides.
The threshold value setting means 36 is a means for setting threshold values α, β, and γ for determining the type of display defect extracted by the defect extraction means 32.
The defect discriminating means 37 is a means for discriminating the type of display defect extracted by the defect extracting means 32 based on the threshold values α, β, γ set by the threshold setting means 36, and discriminates streak-like streak defects. It comprises a streak defect discriminating means 371 for discriminating and a spot defect discriminating means 372 for discriminating a spot-like spot defect.

(2) Defect Determination Method Subsequently, a defect determination method performed by the defect determination apparatus having the above configuration will be described.
FIG. 2 is a flowchart showing the flow of defect determination.

  In S <b> 1, the threshold setting unit 36 sets thresholds α, β, and γ for determining the type of display defect that occurs in the image display device 1. In this embodiment, α = 2, β = 2.5, and γ = 5 are set. Here, the threshold value γ is set to a value larger than the threshold value α.

  In S <b> 2, the image data acquisition unit 31 causes the image display device 1 to display a desired inspection image via the display control unit 11, and causes the imaging unit 2 to capture the inspection image to acquire image data. Here, the inspection image displayed on the image display device 1 is preferably a uniform image without a pattern so that display defects are conspicuous.

In S3, the defect extraction unit 32 extracts display defects included in the image data acquired in S2.
First, the defect extraction unit 32 performs a defect enhancement process for enhancing the display defect included in the image data on the image data acquired in S2 (S3a). Note that details of the defect enhancement processing are disclosed in Japanese Patent Application No. 2004-163643, and thus description thereof is omitted here.
Subsequently, the defect extraction unit 32 binarizes the image data after the defect enhancement processing based on a predetermined binarization threshold (S3b). In the image data after the defect emphasis process, the pixel value of the portion corresponding to the display defect is larger than the pixel value of the other portion. Can be distinguished. With such a binarization threshold, it is possible to generate binary image data in which the pixel value corresponding to the display defect is 1 and the pixel value of the other part is 0.
Subsequently, the defect extraction means 32 concatenates each pixel (each defective pixel) having a pixel value of 1 adjacent to each other in the binary image data, and assigns a natural number of defect numbers starting from 1 to each connected pixel group. (S3c). Here, the reason why the number assigned to each connected pixel group is described as “defect number” is that each connected pixel group corresponds to one display defect included in the image data. . That is, if the number of display defects included in the image data is N (natural number), each connected pixel group corresponding to each display defect is assigned a defect number n of 1 to N, respectively. Hereinafter, for simplification of description, each display defect included in the image data is expressed as “display defect with defect number n” by a defect number n of 1 to N attached to each connected pixel group corresponding to each display defect. ".

  In S4, the display defect of defect number 1 is selected as the display defect of the type discrimination target. FIG. 3 shows an example of the selected display defect. In the example of this figure, pixels having a pixel value of 1 (defective pixels) indicated by hatching with diagonal lines are connected to form a curved streak defect.

In S5, the center of gravity calculation means 33 calculates the center of gravity of the display defect selected in S4. The coordinates i _G , j _G of the center of gravity of the display defect are binary image data f (i, j defined by the following equation 1 when the coordinates of each pixel are expressed as (i, j) by natural numbers i, j. j) = 0 (non-defective pixel), 1 (defective pixel) p + q order moments m _pq are expressed as i _G = m ₁₀ / m ₀₀ and j _G = m ₀₁ / m ₀₀ . FIG. 4 shows the center of gravity G (i _G , j _G ) calculated in the example of FIG.

  Note that Σ on the right side of Equation 1 means that a sum is taken for defective pixels (f = 1).

In S6, the inertia main axis calculation means 34 calculates the inertia main axis of the display defect selected in S4. The direction θ of the principal axis of inertia of the display defect is expressed by the following formula 3 by the centroid moment M _pq around the centroid G (i _G , j _G ) defined by the following formula 2.

  In Equation 3, the center of gravity G is calculated as the coordinate origin. FIG. 5 shows the inertia main axis a calculated in the example of FIG. As shown in this figure, the direction of the inertial main axis a is the direction in which the display defect extends.

  Subsequently, the inertia main axis calculation means 34 calculates the equation (expression 4) of the inertia main axis a and the equation (expression 5) of the axis b (see FIG. 5) perpendicular to the inertia main axis a with the center of gravity G as the coordinate origin. .

  In S7, the circumscribed rectangle setting means 35 includes two circumscribed short sides that are perpendicular to the inertia principal axis of the display defect calculated in S6 and circumscribe the display defect, and two circumscribed rectangles that are parallel to the inertia principal axis and circumscribe the display defect. A circumscribed rectangle circumscribing the display defect is set by the circumscribed long side.

  First, in order to set two circumscribed short sides, defective pixels farthest from the axis b are found on both sides of the axis b. Specifically, the distance D between the pixel at the coordinate (i, j) represented by the following expression 6 and the axis b is calculated for all defective pixels, and the defective pixel having the maximum distance D is calculated on the axis b. Find on each side. Note that k in Expression 6 below is the inclination of the axis b, and specifically, k = 1 / {− tan (90−θ)} as shown in Expression 5.

  Then, as shown in FIG. 6, straight lines b1 and b2 parallel to the axis b are drawn through the defective pixels found on both sides of the axis b.

  Similarly, in order to set two circumscribed long sides, defective pixels farthest from the inertial main axis a are found on both sides of the inertial main axis a. Specifically, the distance D between the pixel at the coordinates (i, j) represented by the equation 6 and the inertial main axis a is calculated for all defective pixels, and the defective pixel having the maximum distance D is determined as the inertial main axis a. Find on both sides of each. In the calculation of the distance D here, k in the equation 6 is the inclination of the inertial main axis a. Specifically, as shown in the equation 4, k = tan (90−θ). is there. Then, as shown in FIG. 6, straight lines a1 and a2 parallel to the inertial main axis a are drawn through the defective pixels respectively found on both sides of the inertial main axis a.

  As described above, the four straight lines a1, a2, b1, and b2 are drawn, and the circumscribed rectangle that circumscribes the display defect is set by the four straight lines. Hereinafter, as shown in FIG. 6, the length of the circumscribed short side of the circumscribed rectangle is A, and the length of the circumscribed long side is B. Depending on the shape of the display defect, the circumscribed rectangle may be set to a square. In this case, A = B.

  In S8, the defect determination unit 37 determines the type of display defect selected in S4 based on the threshold values α, β, and γ set in S1. FIG. 7 is a flowchart showing the flow of defect determination.

  In S9, the streak defect discriminating means 371 compares the threshold value γ = 5 set in S1 with the aspect ratio B / A of the circumscribed rectangle set in S7, and whether γ ≦ B / A is satisfied. Determine whether. When γ ≦ B / A is satisfied (Yes), it is determined that the display defect is a streak defect.

In S9, when γ ≦ B / A is not satisfied (No), in S10, the streak defect discriminating means 371 is set in S7 with the threshold α = 2 and the threshold β = 2.5 set in S1. The aspect ratio B / A of the circumscribed rectangle and the inverse number A · B / S of the display defect occupancy ratio in the circumscribed rectangle are respectively compared, and α ≦ B / A and β ≦ A · B / S are satisfied. It is determined whether or not. Here, S is the area of the display defect, and is expressed as S = m ₀₀ by the moment m _pq defined by the equation 1. When α ≦ B / A and β ≦ A · B / S are satisfied (Yes), it is determined that the display defect is a streak defect.

  In S10, when α ≦ B / A and β ≦ A · B / S are not satisfied (No), in S11, the spot defect determining means 372 determines that α> B / A and β> A. It is determined whether or not B / S is satisfied. When α> B / A and β> A · B / S are satisfied (Yes), it is determined that the display defect is a spot defect.

  In S11, when α> B / A and β> A · B / S are not satisfied (No), the display defect is a display defect of a type other than a streak defect and a stain defect (for example, a point defect). It is determined that

  When the defect determination process S8 as described above is completed, it is determined in S12 (FIG. 2) whether or not the types of display defects have been determined up to the defect number N (total number of display defects). Thereafter, until the type determination for all N display defects is completed (S12: Yes), the defect numbers of the display defects to be determined are incremented by 1 (S13), and the processes of S5 to S8 are repeated.

  As described above, when the determination of the types of all N display defects occurring in the image display device 1 is completed, the inspection process for each display defect whose type has been determined is performed. The optimum inspection processing is preset for each type of display defect (especially streak defect, spot defect) in consideration of human visual characteristics. Then, the display defect can be inspected with the same degree of accuracy as that of human vision by performing an optimal inspection process on the display defect according to the determination result of the type of display defect.

  According to the embodiment as described above, the inertia main axis along the direction in which the display defect to be discriminated extends is calculated, and the type of the display defect is determined using the circumscribed rectangle set along the inertia main axis. It can be determined with high accuracy.

  In the above embodiment, when α ≦ B / A and β ≦ A · B / S are satisfied in S10, it is determined that the display defect is a streak defect. Here, the information that the aspect ratio B / A of the circumscribed rectangle of the display defect is large obtained from satisfying the first equation, and the display defect in the circumscribed rectangle obtained from satisfying the second equation The information that the occupation ratio S / A · B is small is information that suggests that the display defect has a streak shape. Therefore, according to both equations, it is possible to discriminate streak defects with high accuracy.

  In the above embodiment, when α> B / A and β> A · B / S are satisfied in S11, it is determined that the display defect is a spot defect. Here, the information indicating that the aspect ratio B / A of the circumscribed rectangle of the display defect is small obtained from satisfying the first equation, and the display defect in the circumscribed rectangle obtained from satisfying the second equation The information that the occupation ratio S / A · B is large is information that suggests that the display defect has a spot-like shape. Therefore, according to both equations, it is possible to determine a spot defect with high accuracy.

  In the embodiment as described above, in S9, the aspect ratio B / A of the circumscribed rectangle of the display defect is larger than the threshold value γ (= 5) larger than the threshold value α (= 2), and the display defect exceeds a predetermined level. If it is suggested that the display is elongated, the determination based on the threshold value α and the threshold value β in S10 is omitted, and the display defect is determined as a streak defect. In this way, if γ ≦ B / A is satisfied in S9, the process in S10 can be omitted and the determination process can be completed, so that the process can be speeded up.

  In the embodiment as described above, since the discrimination process is performed for each display defect to be discriminated, the discrimination accuracy for each display defect can be improved.

  The present invention is not limited to the embodiment as described above, and any modification of the embodiment within the scope that can achieve the object of the present invention is included in the technical scope of the present invention.

  In the above-described embodiment, defect enhancement, binarization, and defect numbering are performed in order to extract display defects included in the image data in S3. However, other defect extraction methods may be employed. Good. That is, the main object of the present invention is to perform highly accurate defect determination processing on the extracted display defect, so any defect extraction method may be used, and there is no particular limitation.

  In the embodiment, the threshold for discriminating defects in S10 and the threshold for discriminating defects in S11 are set to the same value (α, β), but may be set to different values. . That is, α1 and β1 may be set as threshold values for determining streak defects in S10, and α2 and β2 may be set as threshold values for determining spot defects in S11. In this case, since the four threshold values α1, β1, α2, and β2 can be set to optimum values in accordance with the determination purpose, the type of display defect can be determined with higher accuracy.

  The present invention can be used to determine the type of display defect that occurs in an image display device.

It is a functional block diagram which shows the structure of a defect discrimination device. It is a flowchart which shows the flow of defect discrimination | determination. It is a figure which shows the example of a display defect. It is a figure which shows the gravity center of the display defect calculated in the example of FIG. It is a figure which shows the inertial principal axis of the display defect calculated in the example of FIG. It is a figure which shows the circumscribed rectangle of the display defect set in the example of FIG. It is a flowchart which shows the flow of defect discrimination | determination.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Image display apparatus, 2 ... Imaging part, 3 ... Discrimination control part, 11 ... Display control part, 12 ... Table, 31 ... Image data acquisition means, 32 ... Defect extraction means, 33 ... Gravity center calculation means, 34 ... Inertial spindle Calculation means, 35 ... circumscribed rectangle setting means, 36 ... threshold setting means, 37 ... defect determination means, 371 ... streak defect determination means, 372 ... blot defect determination means.

Claims (9)

A defect determination method for determining the type of display defect of an image display device,
A defect extraction step for extracting display defects from image data displayed by the image display device;
An inertial spindle calculating step of calculating an inertial spindle of the display defect;
A circumscribed rectangle circumscribing the display defect is set by two circumscribed short sides that are perpendicular to the inertia main axis and circumscribe the display defect, and two circumscribed long sides that are parallel to the inertia main axis and circumscribe the display defect. Circumscribed rectangle setting step;
And a defect determination step of determining a type of the display defect based on an aspect ratio of the circumscribed rectangle and an occupation ratio of the display defect in the circumscribed rectangle. The defect determination method according to claim 1,
A threshold value setting step for setting a threshold value α1 (> 1) and a threshold value β1 (> 1) for determining the type of the display defect;
In the defect determination step, the length of the circumscribed short side is A, the length of the circumscribed long side is B, the area of the display defect is S, the aspect ratio B / A of the circumscribed rectangle, and the circumscribed rectangle Are compared with the threshold value α1 and the threshold value β1, respectively, and at least α1 ≦ B / A and β1 ≦ A · B / S are satisfied. In this case, the defect determination method is characterized in that the display defect is determined as a streak-like streak defect. The defect determination method according to claim 2,
In the threshold setting step, a threshold γ1 larger than the threshold α1 is set,
In the defect determination step, first, the aspect ratio B / A of the circumscribed rectangle is compared with the threshold value γ1, and when γ1 ≦ B / A is satisfied, defect determination is performed based on the threshold value α1 and the threshold value β1. And determining that the display defect is a streak defect. In the defect determination method according to any one of claims 1 to 3,
A threshold setting step for setting a threshold α2 (> 1) and a threshold β2 (> 1) for determining the type of the display defect;
In the defect determination step, the length of the circumscribed short side is A, the length of the circumscribed long side is B, the area of the display defect is S, the aspect ratio B / A of the circumscribed rectangle, and the circumscribed rectangle The reciprocal number A · B / S of the display defect occupancy ratio is compared with the threshold value α2 and the threshold value β2, respectively, and at least α2> B / A and β2> A · B / S are satisfied. In such a case, the defect determination method is characterized in that the display defect is determined as a spot-like blot defect. In the defect determination method according to claim 2 or claim 3,
In the threshold value setting step, the threshold value α1 is set to a value larger than 2. In the defect determination method according to any one of claims 2, 3, and 5,
In the threshold value setting step, the threshold value β1 is set to a value larger than 2.5. The defect determination method according to claim 4,
In the threshold value setting step, the threshold value α2 is set to a value larger than 2. In the defect determination method according to claim 4 or 7,
In the threshold value setting step, the threshold value β2 is set to a value larger than 2.5. A defect determination device for determining the type of display defect of an image display device,
Defect extraction means for extracting display defects from image data displayed by the image display device;
An inertia main axis calculating means for calculating the inertia main axis of the display defect;
A circumscribed rectangle that circumscribes the display defect is set by two circumscribed short sides that are perpendicular to the principal axis of inertia and circumscribe the display defect, and two circumscribed long sides that are parallel to the principal axis of inertia and circumscribe the display defect. Circumscribed rectangle setting means;
A defect determination apparatus comprising: defect determination means for determining a type of the display defect based on an aspect ratio of the circumscribed rectangle and an occupation ratio of the display defect in the circumscribed rectangle.

Images