JP2006250876A - Method, device, and program for inspecting optical panel, and recording medium with program stored - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for inspecting optical panel which detects stripe defects, existing within an image displayed by an optical panel and which determines the rank, according to the state of the stripe-shaped defects. <P>SOLUTION: The method includes a stripe-shaped defect detection step (ST100) for detecting the stripe defects existing in the display image by a liquid crystal light valve; a continuity determination step (ST200) for determining whether the stripe defects detected by the stripe-shaped defect detection step (ST100) have continuity with each other; and a rank determination step (ST300), in which the stripe defects, having the continuity determined by the continuity determination step (ST200), are identified as a series of stripe-shaped defects and in which the rank of the liquid crystal light valve 112 is determined, on the basis of the quantitative evaluation value of the detected stripe defects. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an optical panel inspection method, an optical panel inspection apparatus, an optical panel inspection program, and a recording medium on which the program is recorded. For example, the present invention relates to an optical panel inspection method for detecting a streak defect present in an image displayed by an optical panel and performing rank determination according to the state of the streak defect.

As an optical panel for displaying an image, a liquid crystal panel having a liquid crystal cell arranged in each pixel on a display surface and in which display control is performed by a voltage applied to the liquid crystal cell is known.
In such a liquid crystal panel, an image display inspection is performed to inspect whether or not a display defect has occurred in the liquid crystal panel during the manufacturing process.
Conventionally, in an image display inspection of a liquid crystal panel, humans have actually visually checked the presence or absence of defects. However, in recent years, in order to increase the production amount and improve the inspection efficiency, the human eyes There is no automatic inspection going on.

Here, as one of the display defects to be inspected, there is an elongated streak-like defect appearing on the display screen, which is generally called a streak-like defect. FIG. 21 shows an example of a streak defect. And the stripe defect detection method which detects such a stripe defect 710,720 is conventionally known (nonpatent literature 1, patent document 1).
For example, a display image of a liquid crystal panel is picked up by a CCD camera, and this picked-up data is processed by a predetermined algorithm, thereby automatically detecting streak-like defects.
When the streak-like defects 710 and 720 are detected, factors such as the length, thickness, and contrast of the streak-like defects 710 and 720 are comprehensively determined to determine whether the liquid crystal panel is good or bad. Furthermore, rank determination of A to D is performed among non-defective products.

Such rank determination needs to be a determination that approximates an impression when an actual user looks at the display screen.
For example, even if the contrast of the streak-like defects 710 and 720 is low, if the length of the streak-like defects 710 and 720 is long, the long streak-like defects 710 and 720 are conspicuous in the user's appearance. On the other hand, even if the contrast of the streak-like defects 710 and 720 is somewhat high, if the length of the streak-like defects 710 and 720 is short, the streak-like defects 710 and 720 are not conspicuous, and thus are judged to be good.

For example, FIG. 22 is an example of a threshold curve set based on the relationship between the length of the streak defect and the contrast as such a determination threshold.
The lengths and contrasts of the streak defects 710 and 720 detected by the streak defect detection method are plotted on the graph (FIG. 22), and the liquid crystal panel to be inspected is in the non-defective range, or in the non-defective range. In some cases, it is determined which region from rank A to rank D corresponds.

"Image Analysis Handbook", Mikio Takagi, Yoshihisa Shimoda, The University of Tokyo Press (1991), p564-567 JP-A-10-240933 (first page, FIG. 1)

Here, although it is a series of streak-like defects, there are streak-like defects that are thinned or interrupted in the middle.
In the conventional streak defect detection method, even if each streak defect can be detected, if the streak defect is interrupted or thinned in the middle, it is natural that separate defects are detected. It was detected as.

However, if a human actually observes, even if it is interrupted or thinned in the middle, a series of streak-like defects are recognized as a series of defects.
In this way, the judgment of the length of the streak defect differs between human visual inspection and detection by automatic inspection. Therefore, human judgment and automatic inspection can be performed in good / bad judgment and ranking based on the detected streak defect. The problem arises that it is completely different from judgment.

For example, even in the threshold curve shown in FIG. 22, the judgment varies greatly depending on the length of the streak defect, and even from the same contrast, the judgment can be made from a defective product to a non-defective A rank depending on the length of the streak defect. It will be big. As a result, even if a product is determined to be a defective product when viewed by a human, it may be determined as a non-defective product by the automatic inspection.
As a result, even if streak-like defects can be detected with high accuracy by automatic inspection, the detection of length is different from human visual judgment. Therefore, good / bad judgment and rank based on streak-like defects detected by automatic inspection There was a problem that it was not reliable enough.

  An object of the present invention is to provide an optical panel inspection method, an optical panel inspection apparatus, an optical panel inspection program, and a recording medium on which the program is recorded, which can perform rank determination approximate to human visual determination in optical panel inspection. There is.

  The optical panel inspection method of the present invention is an optical panel inspection method for detecting a streak defect present in an image displayed by the optical panel and performing rank determination according to the state of the streak defect. A streak defect detection step for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing an image, and each streak shape detected in the streak defect detection step A continuity determination step for determining whether or not the streak defects have continuity with respect to the defect, and the streak defects determined to have continuity in the continuity determination step are regarded as a series of streak defects and detected. And a rank determination step for determining the rank of the optical panel based on the quantitative evaluation value of the streak-like defect.

  In such a configuration, first, in the streak defect detection step, a display image of the optical panel is picked up, and then a streak defect detection filter is applied to the picked-up image. Then, a streak defect generated in the display image of the optical panel is detected. At this time, since the process by the streak defect detection filter is a mechanical process, naturally, the separated streak defects are detected separately.

Next, in the continuity determination step, it is determined whether or not the streak defects detected separately in the streak defect detection step have continuity between the streak defects. For example, a series of streak defects that are connected to nature and are recognized as a series of streaks in the human eye when it is assumed that it has been interrupted or thinned, but extended without interruption. Judge that there is sex.
In the rank determination step, the streak-like defects determined to have continuity are regarded as a series of continuous streak-like defects, and a quantitative evaluation value of each streak-like defect, for example, length, width, contrast, etc. Based on the above, the rank of the optical panel is determined. As rank determination, in addition to good / bad rank determination, rank determination such as rank A to rank D in non-defective products is performed as necessary.

Since the streak defect detection filter only mechanically performs a streak defect detection process, a streak defect interrupted in the middle can only be detected separately.
However, human eyes have streak-like defects that appear to be connected to a series of lines that are interrupted. For this reason, if the rank determination process is performed while the streak-like defects that are recognized as a series by human beings are regarded as separate short defects, a rank determination result that is different from the person's visual impression is obtained. In this regard, in the present invention, since a continuity determination step is provided between the streak defect detection step and the rank determination step, it is determined whether there is a set having continuity among the detected streak defects. be able to.
Therefore, in the rank determination step, the streak-like defects having continuity can be handled as one defect to perform rank determination. As a result, an appropriate rank determination result can be obtained by performing rank determination that matches the impression seen by a person.

  In the present invention, the continuity determination step includes a separation distance determination step for determining continuity between the stripe defects based on a separation distance between the stripe defects detected by the stripe defect detection step, and a continuity. About one stripe-like defect to be judged and the other stripe-like defect, continuation of the one and the other stripe-like defects based on the degree of deviation of the other stripe-like defect with respect to the approximate straight line obtained by extending the one stripe-like defect It is preferable to include a deviation degree determination step for determining the property.

In such a configuration, when determining the continuity between the streak defects, a determination criterion is how far the streak defects are separated.
While it is difficult to find continuity even in human eyes if the streak defects are greatly separated, it can be judged that the streaky defects may have continuity if the separation distance is relatively small, If the distance between the streak-like defects is large, it can be clearly judged that there is no continuity between the streak-like defects.

  By determining whether there is continuity based on the separation distance between the streak defects in this way, it is determined that there is no continuity between the streak defects whose separation distance is too large, and is equivalent to the human impression Can be determined. Furthermore, even when a more detailed determination is made regarding the determination of continuity, a combination of large separation distances is excluded in advance in the separation distance determination step as having no continuity, thereby allowing subsequent processing. It can be simplified.

Further, in determining the continuity between the streak-like defects, one streak-like defect whose continuity is judged and the other streak-like defect, the other streak with respect to the approximate straight line obtained by extending one streak-like defect. The degree of deviation of the defects is used as a criterion. If it is assumed that the streak-like defect has been extended without interruption, it appears to the human eye that the streak-like defect is continuous if it is in a state of being connected naturally, and the line extending from one streak-like defect is the other If there is a large deviation from the streak-like defects, no continuity will be found in the human eye. In this regard, in the present invention, if the degree of deviation of the approximate straight line obtained by extending one streak defect with respect to the other streak defect is small, it can be determined that there is continuity between the streak defects. On the other hand, if the other streak defect is greatly deviated, it is determined that there is no continuity. For example, if an approximate straight line extending one streak defect extends toward the other streak defect, it can be determined that both are continuous, and conversely, an approximate straight line extending one streak defect is If the other stripe-shaped defect extends in a completely different direction, it can be determined that the two stripe-shaped defects are clearly not continuous.
Thus, by determining the continuity of the streak defect based on the degree of deviation between the approximate straight line obtained by extending one streak defect and the other streak defect, a determination equivalent to a human impression is performed. be able to.

  In the present invention, it is preferable that the deviation degree determination step is performed only on a pair of streak defects whose possibility of continuity is suggested by the determination of continuity in the separation distance determination step.

  According to such a configuration, more appropriate continuity determination can be performed by performing a two-stage determination of the separation distance determination step and the deviation degree determination step. First, by eliminating the combination of streak defects that are determined not to have continuity clearly in the separation distance determination step, the processing time of the deviation degree determination step is shortened, and the inspection efficiency of the optical panel is increased. Can be improved.

  In the present invention, the separation distance determination step includes a distance calculation step of calculating a separation distance between the streak defects, and a distance comparison step of comparing the calculated separation distance with a predetermined threshold to determine continuity. It is preferable to provide.

In such a configuration, the distance between the streak-like defects is calculated (distance calculation step), and then the calculated distance is compared with a predetermined threshold (distance comparison step). Determine.
Where the streak-like defects are greatly separated from each other, continuity cannot be found even in the human eye, by calculating the separation distance between the streak-like defects and comparing this separation distance with a predetermined threshold, While it can be determined that the streak defects having a small separation distance may have continuity, it can be determined that the streak defects that are largely separated beyond the threshold do not have continuity.
That is, in determining the continuity between streak-like defects, it is possible to make a determination equivalent to the appearance of a person by a calculation process such as a calculation process for obtaining a separation distance and a comparison process with a threshold value.

  In the present invention, the continuity determination step includes a thinning step for thinning each detected streak defect, an end point detection step for detecting an end point of the streak defect thinned in the thinning step, and It is preferable that the distance calculating step includes an end-to-end point distance calculating step of calculating a distance between the end points of the streak-like defects whose continuity is determined.

  In such a configuration, each detected streak-like defect is thinned in a thinning step, and then the end point of each thinned streak defect is detected (endpoint detection step). Then, the distance between the end points is calculated for one streak-like defect whose continuity is determined and the other streak-like defect (inter-endpoint distance calculating step). Thereafter, the continuity between the stripe defects is determined by comparing the calculated distance between the end points with a predetermined threshold value.

According to such a configuration, it becomes easy to find the end points of the streak-like defect by thinning, and the distance between the end points can be calculated easily.
Further, when determining continuity, it is considered desirable to determine based on the size of the interval between streak-like defects that were interrupted in the middle. The distance between end points, which is an interval, can be obtained. Then, by comparing the distance between the end points with a predetermined threshold value, it is possible to appropriately determine the continuity between the streak defects.

  In addition, as a threshold value for determining continuity, for example, a distance corresponding to a length of 2 pixels to 400 pixels is set as an example.

  In the present invention, the thinning process includes a pattern matching process for detecting a point matched when a reference pattern for detecting a point on a contour line excluding an end point is applied to a streak defect as a point on the contour line, and a pattern matching process. It is preferable to further include a contour deleting step of sequentially deleting the points on the contour line detected in this way.

  In such a configuration, a point on the contour line of the streak defect is detected by the pattern matching step, and further, the detected point on the contour line is deleted (contour deletion step). As a result, the contour line is sequentially cut, and a thin line connecting the end points of the streak defect is obtained. By sequentially deleting the points on the contour in this way, it is possible to thin the streak defect while maintaining the length and direction information of the streak defect.

  In the present invention, the continuity determination step includes a thinning step for thinning each detected streak defect, an end point detection step for detecting an end point of the streak defect thinned in the thinning step, and The deviation degree determination step connects the end points of the streak-like defects whose continuity is determined with an approximate straight line calculation step for calculating an approximate straight line reflecting a tendency within several pixels from the end points of the streak-like defects. It is preferable to include a sandwiching angle calculation step of calculating a sandwiching angle sandwiched between the end point connection line and the approximate straight line, and a sandwiching angle comparison step of comparing the sandwiching angle with a predetermined threshold to determine continuity.

In such a configuration, first, after the streak defect is thinned by the thinning process, the end point of the thinned streak defect is detected (endpoint detection process).
Next, an approximate straight line reflecting a tendency within several pixels from the end point is obtained for one streak-like defect (approximate straight line calculating step).
In obtaining the approximate straight line, since it is a straight line reflecting a tendency within several pixels from the end point, the directivity when the streaky defect is further extended from the end point is estimated by this approximate line. Further, a sandwiching angle between the end point connecting line connecting the end points of the streak defect and the approximate straight line is calculated (a sandwiching angle calculating step). Then, the continuity between the stripe defects is determined by comparing the sandwich angle with a predetermined threshold value.

Here, since the end point connecting line connecting the end points between two streak defects is a line connecting one streak defect to the other streak defect, if the two streak defects are continuous, If so, the end point connection line and the approximate line extend in the same direction, and it is considered that the approximate line and the end point connection line are close to each other without being far apart.
Therefore, when the clamping angle between the approximate straight line and the end point connection line is compared with a predetermined threshold value, and the opening angle between the approximate straight line and the connection line is not large and the opening between the approximate straight line and the connection line is not large, the streak-like defect is stretched as it is. It can be determined that the continuity is connected to nature.
On the other hand, if the sandwich angle exceeds the predetermined threshold and the opening between the approximate straight line and the end point connection line is large, it can be determined that there is no continuity between the streak defects. For example, the threshold value is set to +30 degrees to -30 degrees.
In determining the continuity between streak-like defects, it is possible to make a determination equivalent to a human impression by a calculation process of comparing the holding angle between the approximate straight line and the end point connection line with a predetermined threshold.

  In calculating the approximate straight line, in a thin line-like defect, a straight line passing through the end point starting from a point returned several pixels from the end point may be obtained, or, for example, within a range within several pixels from the end point You may obtain | require the regression line by the least squares method. Here, in calculating the approximate straight line, the range of “within several pixels from the end point” may be, for example, within 20 pixels from the end point or within 10 pixels from the end point in order to estimate the extension direction of the streak-like defect. A predetermined range that can calculate an approximate straight line reflecting the tendency of the vicinity of the end point of the streak-like defect may be set as appropriate.

  In the present invention, the approximate straight line calculating step calculates the approximate straight line for each of the one streak-like defect and the other streak-like defect whose continuity is determined, and the holding angle calculating step includes the endpoint connecting line. And a first sandwiching angle formed by the approximate straight line of the one streak-shaped defect and a second sandwiching angle formed by the end point connection line and the approximate straight line of the other streak-shaped defect, respectively, In the sandwiching angle comparison step, it is preferable to compare the first sandwiching angle and the second sandwiching angle with a predetermined threshold value.

In such a configuration, after calculating the sandwich angle between the end point connection line and the approximate straight line for both of the sets of streak defects whose continuity is determined (a sandwich angle calculation step), in the sandwich angle comparison step, Both calculated sandwiching angles are compared with a predetermined threshold value.
According to such a configuration, since the opening of the sandwiching angle is checked for both, for example, in a set of streak-like defects that do not have continuity, when one streak-like defect is seen from the other streak-like defect, it happens by chance Even if the clamping angle was small, it can be eliminated.
That is, since it is determined that there is continuity only when both the sandwiching angles are within the threshold, the continuity can be determined more appropriately.

  In determining the degree of deviation between one streak defect and the other streak defect in the deviation degree determination step, the sandwich angle between the approximate straight line and the end point connection line may be compared with a threshold value, For example, the shortest distance between the approximate straight line obtained by extending the streak defect and the other streak defect may be compared with a threshold value.

  In the present invention, the continuity determination step includes an identification number giving step for assigning different identification numbers to the respective streak defects detected by the streak defect detection step, and a streak defect determined to have continuity. It is preferable to include an identification number rewriting step of rewriting the identification numbers of the set to the same number.

In such a configuration, in the identification number assigning step, a different identification number is assigned to each of the streak defects detected in the streak defect detection step, and the identification number rewrite is performed for the streak defects having continuity. In the process, the same identification number is rewritten.
As a result, it can be immediately understood that the streak-like defects having the same identification number are continuous streak-like defects. For example, the identification number is referred to in the rank determination process, and a series of streak-like defects are identified for the same identification number. This makes it easy to handle.

  The optical panel inspection apparatus of the present invention is an optical panel inspection apparatus that detects a streak-like defect present in an image displayed by the optical panel and performs rank determination according to the state of the streak-like defect. A streak defect detection means for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing an image, and each streak shape detected by the streak defect detection means Continuity determination means for determining whether or not the streak defects have continuity with respect to the defect, and the streak defects determined to have continuity by the continuity determination means are regarded as a series of streak defects and detected. And rank determining means for determining the rank of the optical panel based on a quantitative evaluation value of the streak-like defect.

According to such a configuration, the same effects as those of the above-described invention can be achieved.
That is, since a continuity determination means is provided, it is determined whether there is a set having continuity among the detected streak defects, and the continuity streak defect is handled as one defect to determine the rank. It can be carried out. As a result, a more appropriate rank determination result can be obtained by performing rank determination that matches the impression seen by a person.

  The optical panel inspection program of the present invention incorporates a computer into an optical panel inspection apparatus that detects streak defects existing in an image displayed by the optical panel and performs rank determination according to the state of the streak defects. A streak defect detecting unit that detects a streak defect by applying a streak defect detection filter that detects a streak defect to a captured image obtained by capturing a display image by the optical panel, and the streak defect detecting unit A series of continuity determining means for determining whether or not the stripe-like defects have continuity for each of the stripe-like defects detected in Steps, and a series of the line-like defects determined to have continuity by the continuity determining means. Rank determining means for determining the rank of the optical panel based on a quantitative evaluation value of the detected streak defect while considering it as a streak defect It characterized in that to ability.

  The recording medium of the present invention incorporates a computer into an optical panel inspection apparatus that detects streak defects present in an image displayed by the optical panel and performs rank determination according to the state of the streak defects. A streak defect detecting means for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing a display image by the optical panel, and the streak defect detecting means A continuity determining means for determining whether or not the streak defects have continuity for each detected streak defect, and a series of streak defects determined as having continuity by the continuity determining means. An optical pattern that functions as rank determining means that determines the rank of the optical panel based on a quantitative evaluation value of the streak-like defects detected as defects. And characterized by recording a Le test program readable computer.

According to such a configuration, the same effects as those of the above-described invention can be achieved.
In other words, since the continuity determination means is provided, it is possible to determine whether there is a set having continuity among the detected streak defects, and the streak defects having continuity are treated as one defect and ranked. Judgment can be made. As a result, a more appropriate rank determination result can be obtained by performing rank determination that matches the impression seen by a person.

Note that a computer having a CPU (Central Processing Unit) and a memory (storage device) may be provided in the optical panel inspection device, and an inspection program may be installed so that the computer can execute each process.
In order to install the inspection program, a memory card, CD-ROM, or the like may be directly inserted into the inspection apparatus, or a device that reads these storage media may be externally connected to the inspection apparatus. Further, a LAN cable, a telephone line, or the like may be connected to an electronic device and the program may be supplied and installed by communication, or the program may be supplied and installed wirelessly.
Thus, if the function of each means is realized by the inspection program, various parameters can be easily changed.

DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments of the present invention will be illustrated and described with reference to reference numerals attached to respective elements in the drawings.
(First embodiment)
A first embodiment according to the optical panel inspection apparatus of the present invention will be described.
FIG. 1 is an overall configuration diagram of an optical panel inspection apparatus 100. In the following description, an example in which the liquid crystal light valve 112 is an inspection target will be described.
The optical panel inspection apparatus 100 includes a liquid crystal light valve 112 to be inspected, an image projection unit 110 that projects an image, a pattern generator 120 that drives the liquid crystal light valve 112, and an image projected from the image projection unit 110. A projection unit 130 having a screen 134 for projecting, a CCD camera (imaging means) 140 for capturing an image projected on the screen 134, image processing of image data captured by the CCD camera 140, and the liquid crystal light valve 112. And an arithmetic processing unit 200 that performs rank determination.

The image projection unit 110 is a light source 111, a liquid crystal light valve (optical panel) 112 that is driven by the pattern generator 120 and modulates the light from the light source 111, and a projection lens that collects and projects the light from the liquid crystal light valve 112. 113 and a mirror 114 that reflects light from the projection lens 113 toward the screen 134.
The liquid crystal light valve 112 is a liquid crystal panel provided with a liquid crystal cell for each pixel. When the liquid crystal light valve 112 is defective, a display defect such as a stripe defect occurs in the projected image.
The image projecting unit 110 imitates a projection type image display device such as a projector, for example.

The pattern generator 120 applies a drive signal to each switching element that drives each liquid crystal cell of the liquid crystal light valve 112 to drive the liquid crystal light valve 112 in a predetermined pattern.
Here, in performing the image display inspection, it is inspected whether all the display pixels exhibit the same luminance at the same input level, and the pattern generator 120 drives all the liquid crystal cells at the same input level.

  The projection unit 130 includes a dark box 131 and a screen 134 provided in the dark box 131. The dark box 131 has an opening 132 into which light from the image projection unit 110 is incident, and an opening 133 for capturing an image on the screen 134 with the CCD camera 140.

The CCD camera 140 captures a projected image on the screen 134 and outputs the captured image data to the arithmetic processing unit 200.
Here, since the streak defect is detected based on the image data captured by the CCD camera 140, the CCD camera 140 needs to have a function of capturing an image with the number of gradations exceeding the number of gradations of the liquid crystal light valve 112. There is.
For example, when 256 gradations are expressed by the liquid crystal light valve 112, the image data of the CCD camera 140 has the gradation number of 12-bit data in which “0” is black and “4095” is white.

FIG. 2 is a block diagram illustrating a configuration of the arithmetic processing unit.
The arithmetic processing unit 200 is composed of a CPU (central processing unit) and a predetermined memory (storage device), and by incorporating an optical panel inspection program, as shown in FIG. The functions as the continuity determination unit 400 and the rank determination unit 500 are realized.
Here, the continuity determination unit 400 includes a thinning unit 410, an identification number assigning unit 420, an end point detection unit 430, a separation distance determination unit 440, a deviation degree determination unit 450, and an identification number rewriting unit 460. Prepare. The separation distance determination unit 440 includes an end-to-endpoint distance calculation unit 441 and a distance comparison unit 442. The deviation degree determination unit 450 includes an approximate straight line calculation unit 451, a sandwich angle calculation unit 452, and a sandwich angle comparison. Means 453.
The operation of each means will be described with reference to the flowcharts in FIGS. 3, 6, 14, and 16.

Next, the optical panel inspection method according to the first embodiment will be described with reference to the flowcharts of FIGS. 3, 6, 14, and 16.
FIG. 3 is a flowchart showing the procedure of the optical panel inspection method.
The optical panel inspection method includes a streak defect detection step (ST100) for detecting a streak defect present in a display image by the liquid crystal light valve 112, and streak defects detected in the streak defect detection step (ST100). A continuity determination step (ST200) for determining whether or not continuity is present, and the streak defects determined to have continuity in the continuity determination step (ST200) are regarded as a series of streak defects and the detected streaks A rank determination step (ST300) for determining the rank of the liquid crystal light valve 112 based on the quantitative evaluation value of the defect.

First, the streak defect detection step (ST100) will be briefly described.
Here, the streak defect is an elongated streak defect appearing on the display screen (see FIG. 5). The black streak defect in which the brightness of the streak portion is lower than the surroundings, and the brightness of the streak portion than the surroundings. There are high white streak defects. The optical panel inspection method according to the first embodiment can be applied to black streak defects and white streak defects in the same manner. Therefore, the following description does not distinguish unless particularly necessary.

  In performing the streak defect detection step (ST100), first, an inspection image to be a target for detecting streak defects is acquired. At this time, the liquid crystal light valve 112 to be inspected is set on the image projection means 110 and the light source 111 illuminates the liquid crystal light valve 112. Then, a predetermined drive voltage is applied from the pattern generator 120 to the liquid crystal light valve 112. Then, each liquid crystal cell of the liquid crystal light valve 112 is driven to control the illumination transmittance. When light from the liquid crystal light valve 112 is projected onto the screen 134 via the projection lens 113 and the mirror 114, an image is projected on the screen 134. An image projected on the screen 134 is picked up by the CCD camera 140, and image data to be inspected is acquired. An image from the CCD camera 140 is sent to the arithmetic processing unit 200. With respect to the acquired image data, the streak defect detection means 300 of the arithmetic processing unit 200 detects streak defects.

  For the acquired image data, first, as preprocessing, the luminance value of each pixel is obtained from the output value of the CCD camera 140, and further, background difference and area extraction are performed. Subsequently, a streak defect is detected by the streak defect detection filter 600.

FIG. 4 shows an example of the streak defect detection filter 600.
Here, an example of detecting a streak-like defect extending in the vertical direction in the screen will be described.
The streak defect detection filter 600 has a region of several pixels × several pixels (for example, 7 × 7) around the pixel of interest, and is further divided into three parts on the left side, the central part, and the right side, and then two columns on the left side. Only the left emphasis filter 610 weighted only, the center emphasis filter 620 weighted only in the central three columns, and the right emphasis filter 630 weighted only in the right two columns. The center enhancement filter 620 is weighted twice as much as the left enhancement filter 610 and the right enhancement filter 630.

  Such a streak defect detection filter 600 (left enhancement filter 610, center enhancement filter 620, right enhancement filter 630) is subjected to a convolution operation in each of the filters 610, 620, 630 by applying the 7 × 7 pixel size of the image data. Then, the convolution calculation results by the respective filters 610, 620, and 630 are collated with the following conditional expressions to detect streak defects.

Here, the result of the convolution by the left emphasis filter 610 and _{F VA,} the result of the convolution by the central emphasis filter 620 and _{F VB,} the result of the convolution by the right emphasis filter 630 and _{F VC.}

  In the above conditional expression, when the result of the filtering process satisfies the condition 1, the luminance value of the central pixel is higher than the luminance values of the left and right pixels, so that there is a white stripe defect in the central portion. Will be. When the filter processing result satisfies the condition 2, the luminance value of the central pixel is lower than the luminance values of the left and right pixels, so that a black streak defect exists in the central portion. .

  Furthermore, if this streak-like defect detection filter is divided into an upper stage, a middle stage, and a lower stage, and weighted filters are used, laterally extending streak-like defects are detected, and the center enhancement filter includes a left enhancement filter and a right enhancement. If a filter obtained by rotating the filter by 45 degrees is prepared, a streak-like defect extending in an oblique direction is detected.

  By performing such a streak defect detection filter process on the entire image, for example, streak defects 710 and 720 shown in FIG. 5 are detected. After calculating the line width and contrast of the defective portion, binarization processing is performed in which each pixel in the defective portion is set to “1” and pixels in other portions are set to “0” to obtain a binary image.

  In the detection of the streak defect by the streak defect detection filter 600 as described above, the streak defects that are interrupted on the way like the defect shown in FIG. 5 are detected as different defects 710 and 720, respectively.

When the streak defects 710 and 720 are detected in this manner, it is subsequently determined whether the detected streak defects 710 and 720 have continuity (continuity determination step ST200).
The procedure of the continuity determination step (ST200) will be described with reference to the flowchart of FIG.

First, in ST210, the thinning unit 410 performs a thinning step for thinning each detected streak defect 710, 720.
The thinning process includes a pattern matching process that detects a point that matches a reference pattern that detects points on the contour line excluding the end points by applying a streak defect as a contour point, and a point on the contour line that is detected in the pattern matching process. And a contour deleting step of sequentially deleting.
Since the image data has already been binarized and a value of “1” is assigned to the pixel of the streak defect portion, each reference pattern shown in FIG. ) To detect points on the outline of the streak-like defects 710 and 720 (pattern matching step).
By deleting the detected points of the contour line (contour deletion step), the streak defects 710 and 720 are thinned. At this time, thinning is executed until the line width becomes one pixel.
An example of the thin line-shaped defects 710 and 720 is shown in FIG.

  In each reference pattern in FIG. 7, “*” means 1 or 0.

  When the streak-like defects 710 and 720 are thinned (ST210), next, in ST220, an identification number assigning step for assigning an identification number to each of the streak-like defects 710 and 720 is performed by the identification number assigning means 420. That is, in the streak defect detection step (ST100), when a plurality of streak defects 710 and 720 are detected, different identification numbers are assigned to the detected streak defects 710 and 720, respectively.

FIG. 9 is a diagram for explaining the process of assigning the identification number to the stripe defect.
In the identification number assigning step (ST220), since an identification number cannot be directly recorded in the image data (binarized image), an array space having the same size as the image size is prepared for assigning the identification number (see FIG. 10). .
The array space prepared in this way is referred to as an identification number assignment space.

A 3 × 3 filter 800 centered on the target pixel is sequentially applied to the binarized image (FIG. 9) from the upper left (see FIG. 9).
When the filter 800 is scanned in order from the upper left, the pixel of interest corresponds to a pixel having a streak defect (A in FIG. 9).
Here, looking at the eight pixels surrounding the pixel of interest, if there is no pixel assigned with an identification number, it means that it has hit a streak-like defect without an identification number.
In this case, a new number is assigned to the pixel where the pixel of interest is currently located. That is, in the identification number assignment space (FIG. 10), the identification number “1” is assigned to the array coordinates corresponding to the pixel of interest.

Subsequently, when scanning by the filter 800 is continued, the pixel of interest corresponds to a pixel different from the previous pixel in the streak defect (B in FIG. 9).
Here, looking at the eight pixels including the pixel of interest, it is confirmed whether there is a pixel already assigned with an identification number. That is, the corresponding coordinates of the identification number assignment space are examined.
When a pixel to which an identification number is assigned is in the vicinity, since it is a streak-like defect that has already been assigned an identification number, the same number is also assigned to the pixel in which the pixel of interest is currently located. That is, “1” is assigned as the same identification number to the corresponding array coordinates in the identification number assignment space (FIG. 10).

In this way, the filter 800 is sequentially applied and the pixels around the target pixel are referred to, and the same number is assigned to the pixels included in one streak-like defect.
When the scanning of the filter 800 is continued and the pixel of interest hits a pixel with a streak-like defect, if the surrounding pixels do not have an identification number, a new streak-like defect is hit (C in FIG. 9). . Therefore, a new identification number is assigned to the pixel where the pixel of interest is currently located. At this time, the identification number “1” is assigned to another streak-like defect, so that it is distinguished. For example, an identification number “2” is assigned as another number.
In this way, filter scanning is performed in order, and identification numbers are assigned to pixels with streak-like defects.
Then, as shown in FIG. 10, an identification number is assigned to the coordinates corresponding to the streak defect in the identification number assignment space.

Next, in ST230, an end point detection unit 430 performs an end point detection step of detecting end points of the line-shaped defects 710 and 720 thinned in the thinning step (ST210).
In the end point detection step (ST230), the reference pattern shown in FIG. 11 is sequentially applied to the thinned binarized image.
Then, since the end points of the line are in any of the patterns shown in FIG. 11, the points that match the reference pattern are all the end points of the streak defects 710 and 720.
All the detected end points (P _A , P _B , P _C , P _D ) are picked up (see FIG. 12), the coordinates of the end points (P _A , P _B , P _C , P _D ) and the streak having the end points are collected. The identification number (“1” or “2”) of the defect is stored (see FIG. 13).

Next, in ST240, the separation distance determination unit 440 performs a separation distance determination step of determining the continuity between the stripe defects based on the separation distance between the stripe defects.
The separation distance determination step (ST240) will be described with reference to the flowchart of FIG.
The separation distance determination step (ST240) is a step of determining the continuity of the stripe defects 710 and 720 based on the separation distance between the stripe defects 710 and 720, and the stripe defects 710 and 720 for which continuity is determined. A distance calculation step ST241 for calculating the distance between the two end points, and a distance comparison step ST242 for comparing the calculated distance between the end points with a distance threshold to determine continuity.

First, in ST241, the end point distance calculation step is performed by the end point distance calculation means 441.
For example, as shown in FIG. 15, the distance between the end points is calculated for the stripe defects 710 and 720 having different identification numbers “1” and “2”. In endpoint detection step (ST230), end point _(e.g., P A, _{P C)} Since the coordinate values of being detected, for example, the coordinates of the end point _{P A} streaky defect identification number "1" _(x A, in y _a), when the coordinates of the end point _{P C} streaky defect identification number "2" is a _{(x C, y} C), the end point distance _P a _{P C} is calculated by the following equation.

In the calculation process, the distance between the end points is calculated for all combinations of the end points (P _A , P _B , P _C , P _D ).

Next, in ST242, a distance comparison step is performed by the distance comparison unit 442.
A distance threshold for determining continuity is set in advance in the distance comparison unit 442, and the distance between end points (P _A P _C ) calculated in the end point distance calculation step (ST241) is set as this distance threshold. Compare.
Here, between the streak-like defects having continuity, the separation distance is considered to be narrow even if they are interrupted in the middle, and as the distance threshold, a combination of streak-like defects considered to have continuity can be selected. So that combinations of streak-like defects that are too far apart can be excluded. For example, the distance threshold is set to a distance corresponding to a length of 2 pixels to 400 pixels.

Then, the calculated end point distance _P A _{P C} as compared to the distance threshold value (ST242), when there is a terminal point distance to be within the distance threshold value (ST243: YES), is the endpoint-to-endpoint distance _P A _{P C} The set of streak defects (710, 720) is likely to have continuity because the separation distance is narrow. Therefore, the process further proceeds to the next continuity determination step (deviation degree determination step ST250).
On the other hand, when all the calculated end-point distances exceed the distance threshold (ST243: NO), it is determined that there is no combination of streak defects having continuity, and the next continuity determination process (deviation) The continuity determination process (ST200) is terminated without proceeding to the degree determination process ST250), and the process proceeds to the rank determination process of ST300.

Distance determination step (ST240) if the end point distance _P A _{P C} there was a set of within the distance threshold value in (ST243: YES), subsequently to the set (710, 720), deviation determination step (ST250 ) Is performed by the deviation degree determination means 450.
The deviation degree determination step will be described with reference to the flowchart of FIG.
Deviation determination step (ST250), the continuous stripe-like defect together based on the deviation of the other streaky defects for the approximate straight line _{L 2} formed by extending one of the streak defect (720) (710) (710) This is a step of determining the characteristics, and includes an approximate straight line calculating step (ST251), a sandwiching angle calculating step (ST252), and a sandwiching angle comparing step (ST253).

First, in ST251, the approximate straight line calculating step of calculating an approximate straight line L ₂ which reflects the tendency of several pixels from the end point P _C streaky defect 720 is performed by the approximate straight line calculating means 451.
In this step (ST251), the distance determination step (ST240) endpoint-to-endpoint distance _P A _{P C} was within distance threshold (ST243: YES) end point _(P A, P _C) continuity has is determined that the streak-like defects, and calculates an approximate straight line L ₂ which reflects the tendency of the end point P _C near, seek extension that extends toward the approximate straight line L ₂ to the other side of continuity is determined (Fig. 17 reference).

For example, in FIG. 17, when calculating the approximate straight line which reflects the tendency of the end points near the identification number "2", and the end point P _C detected by the endpoint detection process (ST230), the number of pixels from the end point P _C (e.g. 5-10 pixels) only in that back along the streak-like defects 720 and (starting point) _{P S,} calculates a line connecting to the straight line approximation line _{L 2.}
The approximate straight line L ₂ is considered to indicate the extension direction of assuming extends without streaky defect 720 is interrupted.

Next, in ST252, end points _(P A, P _C) sandwiching angle calculating step of the second to calculate a clamping angle theta ₂ which is sandwiched between the end point connecting line _{L AC} an approximate straight line _{L 2} connecting the the clamping angle calculating means 452 Is done.
The calculation of the second clamping angle θ ₂ is performed as follows.
First, down a perpendicular direction from the end point P _A streaky defect identification number "1" to the approximate line L ₂ (see FIG. 18). _Let this perpendicular foot be a point PH. _Then, it calculates the distance P A _{P H.} For example, the approximate straight line L ₂ is assumed to be expressed by the following.

  ax + by + c = 0

At this time, the length of the perpendicular line _P A _{P H} grated end point _{P A (x A, y A} ) of the streak-like defects from the approximate line _{L 2} of the identification number "1" is determined as follows.

At this time, the second clamping angle θ ₂ is obtained by the following equation.

Similarly, for the streak-like defect 710 having the identification number “1”, the approximate straight line L ₁ is calculated, and then the first sandwich angle that is sandwiched between the end point connection line L _AC that connects the end points and the approximate straight line L _1. θ ₁ is calculated (see FIG. 19).

And then, in ST253, nipping angle theta _1, nipping angle comparing step determines continuity of theta ₂ as compared to the threshold angle is performed by nipping angle comparison unit 453.
An angle threshold for determining continuity is set in advance in the sandwich angle comparing means 453, and the sandwich angles θ ₁ and θ ₂ calculated in the sandwich angle calculating step (ST252) are compared with this angle threshold.
Here, it is considered that between the streak-like defects having continuity is small in opening between the end point connection line _LAC and the approximate straight lines L ₁ and L ₂ , the angle threshold is a combination of the streak-like defects having continuity. is set to an angle that can determine, clamping angle theta _1, theta ₂ is large approximation lines _L 1, _{L 2} and the end point _P a, the distance between _{P C} (eg _P a _{P C} distance) streaky which is too distant Defect combinations can be excluded. For example, the angle threshold is set to +30 degrees to −30 degrees.

Then, when the obtained sandwiching angles θ ₁ and θ ₂ are compared with the angle threshold value and both the sandwiching angles θ ₁ and θ ₂ are both within the angle threshold value (ST254: YES), the sandwiching angle is determined for the set. Since the opening of θ ₁ and θ ₂ are both small, it is determined to have continuity.
On the other hand, when the obtained sandwiching angles θ ₁ and θ ₂ exceed the angle threshold, the two streak defects 710 and 720 are not continuous (ST255), and separate independent streak defects 710 and 720 are present. Therefore, the continuity determination step (ST200) is terminated and the process proceeds to rank determination in ST300.

In the deviation degree determination step (ST250), when the sandwiching angles θ ₁ and θ ₂ are within the angle threshold value (ST254: YES), it is determined that the two stripe defects 710 and 720 have continuity. In step ST260, the identification number rewriting unit 460 performs the identification number rewriting process for the set.

In the identification number rewriting step (ST250), the identification numbers of the stripe defects 710 and 720 determined to have continuity are rewritten to the same number (see FIG. 20).
That is, in the identification number assigning step (ST210), different identification numbers ("1" or "2") are assigned to the streak defects 710 and 720 detected by the streak defect detection filter 600, respectively. The streak defect determined to have continuity by the distance determination step ST240 and the deviation degree determination step (ST250) is rewritten to the same identification number in the identification number rewriting step (ST250), and the streak defect having the same identification number is rewritten. Considered as one continuous streak defect.

  In this way, in the continuity determination step (ST200), for the set of stripe-like defects 710 and 720 having continuity, an identification number is assigned to each stripe-like defect 710 and 720 as the same identification number ("1"). Thereafter, in ST300, a rank determination step is performed by rank determination means 500.

In the rank determination step (ST300), the line-shaped defects 710 and 720 having the same identification number are recognized as a series of line-shaped defects based on the identification number (“1” or “2”), and then the line-shaped defect detection filter For the streak defects 710 and 720 detected at 600 (see FIG. 5), the contrast, defect width, and length of the streak defects 710 and 720 are calculated for each identification number. That is, for the stripe-like defect having continuity, the total number of pixels obtained by adding the number of pixels on the thinned image (see FIG. 8 or FIG. 20) is set as the length of the series of stripe-like defects.
Based on the contrast of the streak defects 710 and 720, the defect width, and the length of the streak defect, for example, the threshold curve shown in FIG. The rank determination of rank D from rank A is performed.

  The inspection of the liquid crystal light valve 112 is completed by the end of the rank determination step (ST300), and the liquid crystal light valve 112 is shipped as a product according to the rank determination.

According to such 1st Embodiment, there exists the following effect.
(1) The streak defect detection filter 600 in the streak defect detection step (ST100) can detect only the streak defects 710 and 720 that are interrupted in the middle, but in this embodiment, the streak defect detection step (ST100). ) And the rank determination step (ST300), the continuity determination step (ST200) is provided, so that it is possible to determine whether there is a set having continuity among the detected streak defects 710 and 720. . Therefore, in the rank determination step (ST300), the streak-like defects 710 and 720 having continuity can be handled as one defect to perform rank determination, and by performing rank determination that matches the impression seen by a person, An appropriate rank determination result can be obtained.

(2) In the continuity determination step (ST200), the presence / absence of continuity is determined based on the separation distance between the streak defects (separation distance determination step ST240). It can be determined that there is no continuity. That is, where neither able to find continuity in the human eye have spaced between streaky defect is large, and calculates the distance P _A P _C between the line defect (end point distance calculating step ST241) (distance comparison step) by comparing the distance P _a P _C the distance threshold, the streak-like defects each other over a distance threshold are greatly separated can be determined that there is no continuity. That is, in determining the continuity between streak-like defects, it is possible to make a determination equivalent to the appearance of a person by the arithmetic processing of the end point distance calculation step ST241 and the distance comparison step ST242.

(3) In performing the continuity determination step ST200, first, the stripe defects 710 and 720 are thinned by the thinning process (ST210). Therefore, the end points (P _A , P _B , and P _{C of the} stripe defects 710 and 720 are displayed. , P _D ) can be detected by simple pattern matching (FIG. 11) in the end point detection step (ST230). In thinning, the points on the contour lines of the stripe defects 710 and 720 are detected by pattern matching (FIG. 7), and further, the points on the detected contour lines are deleted. The stripe-shaped defects 710 and 720 can be thinned while the information on the length and direction of 720 remains unchanged.

(4) deviation determination in step (ST250), one streak approximate the defect 720 is extended linearly _{L 2} and other streaky defects 710 streak defects 710 on the basis of the clamping angle theta ₂ of the deviation degree between, By determining the continuity of 720 (ST253), it is possible to make an equivalent determination on the impression of human appearance. That is, when it is assumed that the streak-like defects 710 and 720 extend without interruption, the streak-like defects 720 appear to the human eye as if the streak-like defects are continuous with each other as long as they are connected to each other. it is determined that the approximation line L ₂ was extended there is continuity in the streak defect together when the deviation degree is small with respect to the other streaky defects 710 (ST253, ST254), equivalent the impression of human appearance Can be determined.

(5) Since the two-stage determination of the separation distance determination step ST240 and the deviation degree determination step ST250 is performed, more appropriate continuity determination can be performed. First, by eliminating the combination of streak defects that are determined not to have continuity in the separation distance determination step ST240 (ST242, ST243), the processing time of the deviation degree determination step ST250 is shortened. The inspection efficiency of the optical panel can be improved.

(6) In the misalignment degree determination step ST250, the sandwiching angles θ ₁ and θ ₂ are calculated for not only one but both (ST252), and both sandwiching angles θ ₁ and θ ₂ are compared with the angle threshold (ST253). for example can be eliminated if that is pinched angle theta ₂ was less chance when from one streak defects 720 viewed other streaky defects 710. That is, since it is determined that there is continuity only when both the sandwiching angles θ ₁ and θ ₂ are within the angle threshold (ST254), it is possible to determine continuity appropriately.

(7) In the identification number assigning step ST220, different identification numbers (“1”, “2”) are individually assigned to the streak defects 710 and 720 detected in the streak defect detection step (ST100), The streak-like defects 710 and 720 determined to have continuity in the separation distance determination step (ST240) and the deviation degree determination step (ST250) are rewritten to the same identification number (“1”). With reference to ("1"), it can be readily seen that the line-shaped defects 710 and 720 having the same identification number are line-shaped defects having continuity. As a result, in the rank determination step (ST300), it is easy to handle the same identification number (“1”) as a series of streak defects with reference to the identification number.

It should be noted that the present invention is not limited to the above-described embodiment, and modifications, improvements, and the like within the scope that can achieve the object of the present invention are included in the present invention.
The present invention is applied to an optical panel as an inspection target as long as it is an optical panel that displays an image such as a plasma display panel, an organic EL panel, and a direct mirror device (DMD) as well as a liquid crystal light valve (liquid crystal panel). Of course, it is possible to inspect for streak-like defects.

The calculation of the holding angles θ ₁ and θ ₂ is not limited to the method described above (use of a sine function), and the calculation method is not limited as long as the holding angle can be calculated.
In deviation determination step, nipping angle theta _1, it has been described as an example when comparing to a threshold magnitude of theta _2, for example, to determine the deviation degree on the basis of a length of P _A P _C 18 May be.

  The present invention can be used for inspection of optical panels.

The whole block diagram of an optical panel inspection apparatus. It is a block diagram which shows the structure of an arithmetic processing part. The flowchart which shows the procedure of an optical panel inspection method. The figure which shows an example of a stripe-shaped defect detection filter. The figure which shows the example of the detected streak-like defect. The flowchart which shows the procedure of a continuity determination process. A reference pattern for detecting points on the outline of a streak-like defect. The figure which shows the example of the stripe-like defect thinned. The figure for demonstrating the process in which an identification number is provided to a stripe defect. The figure which shows the state which provided the identification number to the stripe defect in the identification number provision space. The figure which shows the reference pattern for detecting an end point. The figure which shows the state which detected the end point of the stripe defect. The figure which shows the table which recorded the coordinate of the end point of a stripe-like defect. The flowchart which shows the procedure of a separation distance determination process. The figure which shows a mode that the distance between the endpoints of a stripe defect is calculated. The flowchart which shows the procedure of a deviation | shift degree determination process. The figure which shows an approximate line and a clamping angle. The figure which shows a mode that a clamping angle is calculated. The figure which shows a mode that the clamping angle was computed about both the stripe-like defects. The figure which shows the state which rewritten the identification number to the same number about the stripe-shaped defect which has continuity. The figure which shows the example of a stripe-like defect. The figure which shows an example of the threshold curve set from the relationship between the length of a stripe-shaped defect, and contrast as a judgment threshold value of a rank determination.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Optical panel inspection apparatus, 110 ... Image projection means, 111 ... Light source, 112 ... Liquid crystal light valve, 113 ... Projection lens, 114 ... Mirror, 120 ... Pattern generator, 130 ... Projection part, 131 ... Dark box, 132 ... Opening , 133 ... opening, 134 ... screen, 140 ... CCD camera, 200 ... arithmetic processing part, 300 ... streak defect detection means, 400 ... continuity determination means, 410 ... thinning means, 420 ... identification number assigning means, 430 ... Endpoint detection means, 440 ... Separation distance determination means, 441 ... Distance between end points calculation means, 442 ... Distance comparison means, 450 ... Degree determination means, 451 ... Approximate straight line calculation means, 452 ... Holding angle calculation means, 453 ... Holding Corner comparison means, 460 ... identification number rewriting means, 500 ... rank determination means, 600 ... streak defect detection filter, 610 ... Side enhancement filter, 620 ... central emphasis filter, 630 ... right emphasis filter, 710 ... streaky defect, 720 ... streaky defect, 800 ... filter.

Claims (11)

In an optical panel inspection method for detecting a streak defect present in an image displayed by an optical panel and performing rank determination according to the state of the streak defect,
A streak defect detecting step of detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing a display image by the optical panel;
A continuity determination step for determining whether or not the streak defects have continuity for each streak defect detected in the streak defect detection step;
A rank in which the streak-like defects determined to have continuity in the continuity determination step are regarded as a series of streak-like defects and the rank of the optical panel is determined based on a quantitative evaluation value of the detected streak-like defects. An optical panel inspection method comprising: a determination step. The optical panel inspection method according to claim 1,
The continuity determination step includes
A separation distance determination step for determining the continuity between the stripe defects based on the separation distance between the stripe defects detected by the stripe defect detection step;
For one streak defect and the other streak defect to be judged for continuity, the one and other streak defects are based on the degree of deviation of the other streak defect with respect to an approximate straight line extending one streak defect. An optical panel inspection method comprising: a deviation degree determination step for determining continuity between each other. The optical panel inspection method according to claim 2,
The separation distance determination step includes
A distance calculating step of calculating a separation distance between the streak defects;
A distance comparison step of comparing the calculated separation distance with a predetermined threshold to determine continuity. An optical panel inspection method, comprising: In the optical panel inspection method according to claim 3,
The continuity determination step includes
A thinning process for thinning each detected streak defect;
An end point detection step for detecting an end point of the line-shaped defect thinned in the thinning step,
The distance calculating step includes
An optical panel inspection method comprising: an end-point distance calculation step of calculating a distance between end points of a streak-like defect whose continuity is determined. The optical panel inspection method according to claim 4,
The thinning step is a pattern matching step of detecting a point matched when a reference pattern for detecting a point on a contour line excluding an end point is applied to a streak defect as a point on the contour line;
An optical panel inspection method comprising: a contour deletion step of sequentially deleting points on the contour line detected in a pattern matching step. In the optical panel inspection method according to any one of claims 2 to 5,
The continuity determination step includes
A thinning process for thinning each detected streak defect;
An end point detection step for detecting an end point of the line-shaped defect thinned in the thinning step,
The deviation degree determination step includes
An approximate straight line calculating step for calculating an approximate straight line reflecting a tendency within several pixels from the end point of the streak defect;
A sandwiching angle calculating step of calculating a sandwiching angle between the end point connecting line connecting the end points of the streak-like defect whose continuity is determined and the approximate straight line;
And a sandwiching angle comparing step of determining continuity by comparing the sandwiching angle with a predetermined threshold. An optical panel inspection method comprising: The optical panel inspection method according to claim 6,
The approximate straight line calculating step calculates the approximate straight line for each of one streak-like defect and the other streak-like defect whose continuity is determined,
The sandwiching angle calculation step includes a first sandwiching angle formed by the end point connection line and the approximate straight line of the one streak defect, and a first sandwich angle formed by the end point connection line and the approximate straight line of the other streak defect. 2 respectively calculate the sandwich angle and
In the sandwiching angle comparison step, the first sandwiching angle and the second sandwiching angle are compared with a predetermined threshold value. In the optical panel inspection method according to any one of claims 1 to 7,
The continuity determination step includes
An identification number assigning step for assigning different identification numbers to the respective streak defects detected by the streak defect detection step;
An optical panel inspection method comprising: an identification number rewriting step of rewriting the identification number of a set of streak-like defects determined to have continuity to the same number. In an optical panel inspection apparatus that detects a streak-like defect present in an image displayed by an optical panel and performs rank determination according to the state of the streak-like defect,
A streak defect detecting means for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing a display image by the optical panel;
Continuity determining means for determining whether or not the streak defects have continuity for each streak defect detected by the streak defect detecting means;
Rank in which the streak-like defects determined to have continuity by the continuity determining means are regarded as a series of streak-like defects and the rank of the optical panel is determined based on a quantitative evaluation value of the detected streak-like defects An optical panel inspection apparatus comprising: a determination unit. A computer is incorporated in an optical panel inspection apparatus that detects streak defects present in an image displayed by the optical panel and performs rank determination according to the state of the streak defects.
A streak defect detecting means for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing a display image by the optical panel;
Continuity determining means for determining whether or not the streak defects have continuity for each streak defect detected by the streak defect detecting means;
Rank in which the streak-like defects determined to have continuity by the continuity determining means are regarded as a series of streak-like defects and the rank of the optical panel is determined based on a quantitative evaluation value of the detected streak-like defects A computer-readable optical panel inspection program that functions as a determination means. A computer is incorporated in an optical panel inspection apparatus that detects streak defects present in an image displayed by the optical panel and performs rank determination according to the state of the streak defects.
A streak defect detecting means for detecting a streak defect by applying a streak defect detection filter for detecting a streak defect to a captured image obtained by capturing a display image by the optical panel;
Continuity determining means for determining whether or not the streak defects have continuity for each streak defect detected by the streak defect detecting means;
Rank in which the streak-like defects determined to have continuity by the continuity determining means are regarded as a series of streak-like defects and the rank of the optical panel is determined based on a quantitative evaluation value of the detected streak-like defects An optical panel inspection program that functions as a determination unit is recorded in a computer-readable manner.

Images