JP2008064686A - Inspection image acquisition method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inspection image acquisition method capable of inspecting accurately the ununiformity of an optical characteristic, even using a plurality of cameras. <P>SOLUTION: A duplicated overlapped area 6 is imaged to be picked up when imaging fixed areas 4, 5 different respectively of one inspection object, using the plurality of cameras 2, 3, a correction range to be corrected is determined based on respective brightnesses in a position separated by a fixed distance from a junction part that is the center axis of the overlapped area out of one of picked-up images, and in a position separated by a fixed distance from the junction part out of the other of the picked-up images, the correction range is corrected to make the brightness get near to the brightness of the one picked-up image stepwisely along a direction distant from the junction part, assuming that the correction range is chosen from the junction part, in the one picked-up image, and the correction range is corrected to make the brightness get near to the brightness of the other picked-up image stepwisely along a direction distant from the junction part, assuming that the correction range is chosen from the junction part, in the other picked-up image. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an inspection image acquisition method for an inspection object such as a filter used in a liquid crystal display (LCD) or the like.

  In recent years, liquid crystal displays (LCDs) have attracted attention as flat displays, and due to their thinness, light weight, small power consumption, and flickerless characteristics, the market is rapidly growing mainly for notebook personal computers (PCs) and PC monitors. Expanded. Recently, large LCDs have also been used for TVs for which CRT has been the mainstream. Here, in order to inspect whether or not the optical characteristics of an inspection object such as a filter conventionally used in an LCD or the like are uniform (non-uniformity), an image of the inspection object is obtained by one camera. And inspecting the non-uniformity of the inspection object using the image. However, recently used LCDs are large in size, so that a single camera cannot capture the entire region of an inspection object such as a filter used in a large LCD as in the prior art. Therefore, when such an inspection object is imaged, a plurality of cameras are arranged side by side, each of the cameras captures a different area of the inspection object, and the images captured by each camera are combined later Thus, a method of acquiring an image of the entire region of the inspection object is taken. After that, inspections such as non-uniformity of optical characteristics were performed using the images synthesized in this way.

JP-A-8-254499

  However, each camera used in the above method has different optical characteristics, so that the optical characteristics (brightness) of images obtained by each camera are different. In this state, when each image is synthesized and one inspection image is created, the seam of the image appears clearly. When such inspection image is used to inspect the optical characteristics for nonuniformity, There has been a problem in that the joints are misrecognized as non-uniform optical characteristics. However, if the individual inspection is performed on each image without connecting the images, there has been a problem that nonuniformity in optical characteristics existing between the images cannot be inspected.

  The present invention has been made in view of such a problem, and is an inspection object capable of accurately inspecting non-uniformity in optical characteristics even when images captured using a plurality of cameras are used. It is an object to provide a method for acquiring an inspection image of an object.

  The invention according to claim 1 for solving the above-described problem is an inspection image acquisition method for acquiring an inspection image of one inspection object using a plurality of cameras, and the inspection object is different in each camera. When imaging a certain area, each of the images captured by the adjacent cameras is imaged by the imaging process of imaging the inspection object so that an overlapping overlap area is imaged, and the adjacent camera. Each of a position apart from the joint which is the central axis of the overlap region in one image and a position away from the joint in the other image captured by the adjacent camera. A correction range determination step for determining a correction range to be corrected based on the luminance of the image, and the correction range from the joint portion in the one image The correction range is corrected so as to gradually approach the luminance of the one image in a direction away from the joint, and the correction range is taken from the joint in the other image, The correction range includes a correction step in which correction is performed so as to gradually approach the luminance of the other image in a direction away from the joint.

  According to this, by correcting the image A and the image B in this way, the image A and the image B captured using two cameras are combined to obtain an inspection image of one inspection object. When doing so, the seam of the image can be synthesized smoothly. At that time, in order to determine the correction range according to the difference in luminance at a certain distance from the joint between the image A and the image B, in other words, according to the difference in luminance of the images captured by the respective cameras. For example, the respective images can be smoothly synthesized in accordance with the difference in optical characteristics of the cameras from which the images are captured. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

  The invention according to claim 2 for solving the above-described problem is the inspection image capturing method for an inspection object according to claim 1, wherein the average brightness value of the overlap region of the one image and the other The method further includes a luminance correction step of correcting the luminance of the one image and the luminance of the other image so that the average value of the luminance of the overlap region of the images becomes the same.

  According to this, it is possible to perform the correction without difficulty by performing a rough correction for the overlap region, and further performing a two-stage correction for performing a fine correction, and an image captured using two cameras. When A and the image B are combined and an inspection image of one inspection object is acquired, the seam of the image can be combined smoothly. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

  Invention of Claim 3 for solving the said subject has the movement process which moves the said test target object in the inspection image acquisition method of Claim 1 or Claim 2, The said imaging process, The inspection object is imaged using a line sensor camera.

  According to this, even when the inspection object moves, the image A and the image B captured using the two line sensor cameras are corrected in this way, so that both images are combined and When acquiring the inspection image of the inspection object, the seam of the image can be smoothly synthesized. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

  According to the present invention, by correcting the images A and B in this way, the images A and B captured using two cameras are combined, and an inspection image of one inspection object is obtained. When acquiring, the seam of the image can be smoothly synthesized. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

(I) First Embodiment A first embodiment of the inspection image acquisition method of the present invention will be described with reference to the drawings.

  First, an inspection image acquisition apparatus used in the inspection image acquisition method according to the present embodiment will be conceptually described with reference to FIG. In the present embodiment, a case where an inspection object is imaged using two cameras and an inspection image is acquired will be described.

  FIG. 1 is a conceptual diagram of an inspection image acquisition apparatus, and FIG. 2 is a conceptual diagram illustrating images captured by the camera 2 and the camera 3.

  As shown in FIG. 1, the inspection image acquisition apparatus includes an inspection object 1, a camera 2, a camera 3, an imaging range 4 by the camera 2, an imaging range 5 by the camera 3, and an overlap region 6.

  The inspection object 1 is a filter or the like used for a liquid crystal display or the like, and has a colored layer, a black matrix, or the like formed on a glass substrate (not shown).

  The camera 2 and the camera 3 are for imaging the inspection object 1 and are not particularly limited, and examples include an area sensor camera.

  The imaging range 4 by the camera 2 is an area imaged by the camera 2 and is, for example, a left area of the inspection object 1 as shown in FIG.

  An imaging range 5 by the camera 3 is an area imaged by the camera 3, and is, for example, a right area of the inspection object 1 as shown in FIG. 1.

  The overlap region 6 is a portion that is imaged by the camera 2 and the camera 3 in an overlapping manner.

  In the inspection image acquisition apparatus having such a configuration, the camera 2 and the camera 3 respectively capture different areas of the inspection object 1. At that time, an overlap region 6 which is a portion that is overlapped and captured by the camera 2 and the camera 3 is captured by both cameras. Next, as correction 1, the brightness of image A and the brightness of image B are set so that the average value of the brightness of overlap area 6 of image A and the average value of the brightness of overlap area 6 of image B are the same. Make corrections. Thereafter, correction 2 is performed. As will be described in detail later, first, the luminance a1 at a position away from the joint 7 in the image A and the position at a distance (m = 1) away from the joint 7 in the image B. A correction range to be corrected is determined based on each luminance b2. Thereafter, the image A and the image B are corrected using the correction range, and the image A and the image B are synthesized.

  By correcting the images A and B in this way, the image A and the image B captured using two cameras are combined to obtain an inspection image of one inspection object. Can be synthesized smoothly. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

  Hereinafter, the inspection image acquisition method will be described in the order of steps.

  First, the camera 2 and the camera 3 respectively image different fixed areas of the inspection object 1.

  Specifically, as shown in FIG. 1, the camera 2 is used to image the left side area of the inspection object 1 (imaging range 4 by the camera 2), and the camera 3 is used to capture the right side area (camera). 3 is imaged. At that time, imaging is performed by both cameras so as to provide an overlap region 6 that is a portion that is overlapped by the camera 2 and the camera 3. 2A is an image (image A) obtained by imaging the left area of the inspection object 1 with the camera 2 as described above, and FIG. 2B is an image of the right area of the inspection object 1 with the camera 3. Each of the images (image B) is shown.

  Next, correction 1 is performed as the first stage correction.

  Since the camera 2 and the camera 3 have different camera optical characteristics, the brightness of the images A and B taken by the camera A and the camera B are different. Therefore, the first-stage correction (correction 1) is performed in order to synthesize the image A and the image B to obtain one smooth inspection image.

  Specifically, the luminance of the image A and the luminance of the image B are corrected so that the average value of the luminance of the overlap region 6 of the image A and the average value of the luminance of the overlap region 6 of the image B are the same. Do.

  FIG. 3 is a diagram illustrating the image A and the image B after the correction 1 is performed.

  Correction 1 corrects the luminance of image A and the luminance of image B so that the average value of the luminance of overlap region 6 of image A and the average value of the luminance of overlap region 6 of image B are the same. After being broken, as shown in FIG. 3, the average value of the luminance of the overlap region 6 of the image A is matched with the average value of the luminance of the overlap region 6 of the image B. Hereinafter, the left region from the joint 7 is referred to as an image A, and the right region from the joint 7 is referred to as an image B.

  Next, correction 2 is performed as a second-stage correction.

  As described above, the luminance of image A and the luminance of image B are adjusted so that the average value of the luminance of overlap region 6 of image A and the average value of the luminance of overlap region 6 of image B are the same by correction 1. Correction is performed, and then correction 2 is further performed as a finer second-stage correction.

[About correction range]
First, as shown in FIG. 3, in the image A, the brightness a1 at a position away from the joint 7 which is the axis of the overlap region 6 by a certain distance (m = 1), and in the image B, constant from the joint 7 A correction range to be corrected is determined based on the respective luminances b2 at positions separated by a distance (m = 1).

  Here, the correction range means that when the images A and B captured by the camera 2 and the camera 3 are combined into one inspection image, correction is performed to smoothly combine the joints of the images. A range.

(parameter settings)
As shown in FIG. 2, the image sizes of the images A and B captured by the camera 2 and the camera 3 are a horizontal length X and a vertical length Y, respectively. Here, a correction range coefficient RK and a correction range limit R _MAX necessary for obtaining the correction range R are set in advance. The correction range coefficient RK is an index for determining the correction range R, and the correction range limit R _MAX is a value indicating the maximum area in which the correction range should be performed. Specifically, the following description will be made assuming that RK = 3 and R _MAX = 20.

Here, the brightness of the image A is
(Expression 1) F1 {m, n}, m = 1 to X, n = 1 to Y,
On the other hand, the brightness of the image B is
(Expression 2) F2 {m, n}, m = 1 to X, and n = 1 to Y.

(Determination of correction range)
The correction range R is obtained using the above parameters. The correction range R is expressed by Equation 3 below.

(Formula 3) Correction range R = | a1-b2 | × RK

The luminance a1 at a position away from the joint 7 in the image A by a certain distance (m = 1)
a1 = F1 {1, n},
In the image B, the luminance b2 at a position away from the joint 7 by a certain distance (m = 1) is expressed by Equation 2.
It is represented by b2 = F2 {1, n}.

  Specifically, when the luminance at a position away from the joint 7 by a certain distance (m = 1) is a1 = 3 and b2 = 2, a1 = 3, b2 = 2, and RK in the above formula 1. By substituting = 3, the correction range R = 3 can be determined.

Here, when the value of the correction range R exceeds the value of the correction range limit R _MAX , the value of the correction range limit R _MAX is corrected as the correction range R below. Specifically, since the correction range obtained by the above equation 3 is R = 3 and the preset correction range limit is R _MAX = 20, the value of the correction range R obtained by the above equation 3 is Since the value of the correction range limit R _MAX is not exceeded, the correction range is determined as R = 3.

  Next, as shown in FIG. 4, the correction range R determined as described above takes the range from the joint portion 7 in the image A, and takes the range from the joint portion 7 in the image B. .

  Specifically, in the above method, since the correction range R = 3 is determined, the correction range (R = 3) is taken from the joint portion 7 in the image A, and the correction range (R = 3) is taken from the joint portion 7 in the image B. 3) shall be taken. In addition, as shown in FIG. 4, let the distance from the junction part 7 be m = 1-R.

  Here, the correction range R is represented by the numerical formula 3, the luminance a1 at a position away from the joint portion 7 in the image A by a certain distance (m = 1), and the position at a distance from the joint portion 7 in the image B by a certain distance (m = 1). Therefore, when the difference between the luminance a1 of the image A and the luminance b2 of the image B is large, the value of the correction range R is also large. Therefore, in the correction range R, when the difference between the luminance a1 of the image A and the luminance b2 of the image B is large, the luminance is corrected over a wide range. On the other hand, when the difference between the luminance a1 of the image A and the luminance b2 of the image B is small, the value of the correction range R is also small. Therefore, when the difference between the luminance a1 of the image A and the luminance b2 of the image B is small, the luminance is corrected within a narrow range. In this way, the correction range R is determined according to the difference in luminance at a position away from the joint portion 7 between the image A and the image B by a certain distance (m = 1), so that the luminance of the image captured by each camera is determined. Depending on the difference, in other words, the respective images can be smoothly synthesized in accordance with the difference in the optical characteristics of the cameras from which the images were taken.

  Next, after the correction range is determined as R = 3 in this way, in the image A and the image B, correction is performed on the areas at the distances of m = 1, 2, and 3 from the joint portion 7, respectively. Specifically, in the image A, correction is performed so as to gradually approach the luminance of the image A in a direction away from the joint 7 such that m = 1, m = 2, and m = 3, and in the image B, Correction is performed so as to gradually approach the brightness of the image B in a direction away from the joint 7 such as m = 1, m = 2, and m = 3.

[Image A correction]
Here, the brightness of the image A within the correction range is obtained by the following equation.
(Formula 4) F1 (x, y) = W1 (m) × F1 {m, n} + W2 (m) × F2 {−m, n}
Here, the luminance weight coefficient W1 for the image A is
W1 (m) = 0.5 + Δ × (m−1), distance from the joint: m = 1 to R
And the luminance weighting factor W2 for the image B is
W2 (m) = 0.5−Δ × (m−1), distance from the joint: m = 1 to R
It is represented by

  Furthermore, Δ is defined as Δ = 0.4 / (R−1), and W1 (m) + W2 (m) = 1.0. Here, when the correction range R is determined to be 1 in Equation 3, W1 = W2 = 0.5.

  In this way, according to the above formula, the luminance at a position m = 1 away from the joint portion 7 in the image A is based on the luminance a1 of the image A and the luminance a2 of the image B at a position m = 1 away from the joint portion 7. Can be sought. Similarly, the luminance at a position m = 2 away from the junction 7 in the image A can be obtained based on the luminance c1 of the image A and the luminance c2 of the image B at a position m = 2 away from the junction 7. The luminance at a position m = 3 away from the joint 7 in the image A can be obtained based on the luminance e1 of the image A and the luminance e2 of the image B at a position m = 3 away from the joint 7 (FIG. 5). reference). Therefore, it is possible to perform correction that gradually approaches the luminance of the image A from the joint portion 7.

  When the correction range is 0 (a1 = b2), the brightness of the image A captured by the camera 2 and the brightness of the image B captured by the camera 3 are the same. Luminance correction is not performed when combining.

[Image B correction]
Next, the image B is corrected in the same manner as the image A.

Here, the brightness of the image B within the correction range is obtained by the following equation.
(Formula 5) F2 (x, y) = W1 (m) × F1 {−m, n} + W2 (m) × F2 {m, n}
Here, the luminance weight coefficient W1 for the image A is
W1 (m) = 0.5−Δ × (m−1), distance from the joint: m = 1 to R
And the luminance weighting factor W2 for the image B is
W2 (m) = 0.5 + Δ × (m−1), distance from joint: m = 1 to R
It is represented by Other values are the same as those described in the correction of the image A.

  Thus, based on the above formula, the luminance at a position m = 1 away from the joint 7 in the image B is based on the luminance b1 of the image A and the luminance b2 of the image B at a position m = 1 away from the joint 7. Can be sought. Similarly, the brightness at a position m = 2 away from the joint 7 in the image B can be obtained based on the brightness d1 of the image A and the brightness d2 of the image B at a position m = 2 away from the joint 7. The luminance at a position m = 3 away from the joint portion 7 in the image B can be obtained based on the luminance f1 of the image A and the luminance f2 of the image B at a position m = 3 away from the joint portion 7 (FIG. 5). reference). Therefore, it is possible to perform correction that gradually approaches the luminance of the image B from the joint portion 7.

  As described above, when the image A and the image B are corrected as described above, the image A and the image B captured using the two cameras are combined to obtain an inspection image of one inspection object. In addition, the seam of the image can be synthesized smoothly (see FIG. 6). Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

(II) Second Embodiment Hereinafter, a second embodiment of the inspection image acquisition method of the present invention will be described with reference to the drawings. In the first embodiment described above, images A and B are captured by the camera 2 and the camera 3 while the inspection object is stationary. However, in the present embodiment, the inspection object is moved by a moving process (on the paper surface). A case where the image A and the image B are captured while being moved upward will be described. In addition, the common name and code | symbol are used for the member etc. which are common in 1st Embodiment and FIG. 1 thru | or FIG.

  The inspection image acquisition apparatus shown in FIG. 1 is the same as that described in the first embodiment, but the cameras 2 and 3 are for capturing images A and B while moving the inspection object. And a line sensor camera or the like.

  The inspection image acquisition method is the same as that described in the first embodiment, and correction 1 and correction 2 are performed to correct image A and image B and combine them into one inspection image.

  Here, unlike the first embodiment, when the image is captured by the line sensor camera, the brightness of the image varies depending on the traveling direction of the inspection object 1 (vertical direction on the paper surface). In the correction 2 described in the above, when obtaining the luminance within the correction range of the image A, y (n) is n = 1 to Y, and when obtaining the luminance within the correction range of the image B, y (n) is n. = 1 to Y is corrected. Accordingly, it is possible to acquire a smooth inspection image of one inspection object in consideration of the luminance difference in the Y-axis direction between the image A and the image B captured by the line sensor camera.

  By correcting the images A and B in this way, the image A and the image B captured using two cameras are combined to obtain an inspection image of one inspection object. Can be synthesized smoothly. Therefore, it is possible to eliminate the unevenness of the seam imaged by the two cameras, and when inspecting the nonuniformity of the optical characteristics, erroneous recognition such as the optical characteristics being considered nonuniform due to the unevenness. It is possible to eliminate this, and it is possible to reliably inspect the non-uniformity of the optical characteristics using the inspection image.

It is a conceptual diagram of a test | inspection image imaging device. It is a conceptual diagram which shows each image imaged with the camera 2 and the camera 3. FIG. It is a figure which shows the image A and the image B after correction 1 was performed. It is a figure for demonstrating each correction range of the image A and the image B. FIG. It is a figure for demonstrating the correction 2 in the correction range of the image A and the image B. FIG. It is a figure after the image A and the image B were synthesize | combined.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Inspection object 2 ... Camera 3 ... Camera 4 ... Imaging range imaged with camera 2 5 ... Imaging range imaged with camera 3 6 ... Overlapping area 7. ..Junction X: Horizontal length of image Y: Vertical length of image

Claims (3)

In an inspection image acquisition method for acquiring an inspection image of one inspection object using a plurality of cameras,
When each of the cameras captures different areas of the inspection object, the inspection object is imaged so that overlapping images are captured in the images captured by the adjacent cameras. Imaging process;
Of the one image captured by the adjacent camera, a position that is a fixed distance away from the joint that is the central axis of the overlap region, and the other image captured by the adjacent camera is constant from the joint A correction range determination step for determining a correction range to be corrected based on the respective luminances at positions separated by
In the one image, the correction range is taken from the joint, and the correction range is corrected so as to gradually approach the luminance of the one image in a direction away from the joint, and in the other image The correction step is to take the correction range from the joint, and the correction range is corrected so as to gradually approach the brightness of the other image in a direction away from the joint,
An inspection image acquisition method characterized by comprising: The inspection image capturing method for an inspection object according to claim 1,
Correction of the luminance of the one image and the luminance of the other image so that the average value of the luminance of the overlap region of the one image and the average value of the luminance of the overlap region of the other image are the same. A method for acquiring an inspection image, further comprising: a luminance correction step for performing the step. In the inspection image acquisition method according to claim 1 or 2,
A moving step of moving the inspection object;
The said imaging process images the said test target object using a line sensor camera, The inspection image acquisition method characterized by the above-mentioned.