TWI536003B - Apparatus and method of detecting defect for patterned retardation film and method of manufacturing patterned retardation film - Google Patents

Description

Defect detecting device and method for patterning retardation film and manufacturing method

The present invention relates to a defect detecting apparatus and method and a manufacturing method relating to a patterned retardation film.

In recent years, there has been a spatial segmentation method as one of the three-dimensional (3D) television display methods that the market has gradually expanded. In the spatial division method, only the image light for the right eye such as an odd line of the display screen enters the right eye, and only the image light for the left eye of the even line enters the left eye. As a display method using such a spatial division method, a method of arranging a patterned retardation film on the surface of a display panel is known. As a typical patterned retardation film, a patterned retardation film having a function of a quarter-wave plate is known.

The patterned retardation film is alternately arranged with a stripe shape of two phase difference regions at a fixed pitch. The two phase difference regions have a function of a λ/4 wavelength plate that orthogonalizes the optical axes (the phase axis or the slow phase axis). The linearly polarized image light of the odd-numbered rows is incident on one phase difference region, and the even-line linearly polarized image light is incident on the other phase difference region. Thereby, for example, the odd-numbered lines of image light are converted into right-handed circularly polarized light, the right-handed circularly polarized light is emitted, and the even-numbered lines of image light are converted into left-handed circularly polarized light, and the left-handed circularly polarized light is emitted. Further, the observer observes the circularly polarized glasses having different rotational directions of the left and right polarized lights, thereby allowing only the image for the right eye of one line to enter the right eye, and only the image for the left eye is entered to the left. Eyes, the result will be observed 3D images.

The manufacturing steps of the patterned retardation film are carried out under strict management. Due to various factors in this manufacturing step, it is difficult to completely eliminate defects. As a defect, there is a mixture of foreign matter, uneven light distribution, damage, and the like. Therefore, a so-called on line inspection which performs inspection on the manufacturing line is carried out. Next, the position of the patterned retardation film on which the defect has occurred is grasped, and the defective portion is removed or the product is not used. Further, depending on the situation, the treatment for eliminating the cause of the defect is performed.

Defect detection techniques targeting patterned retardation films are not known. In the case of the transparent film or the retardation film, for example, a defect such as Japanese Patent Laid-Open No. Hei 6-148095 (Patent Document 1), Japanese Patent Laid-Open Publication No. 2008-267991 (Patent Document 2) Inspection devices are known. In the inspection apparatus of the patent document 1, the transparent film is placed between the first polarizing plate and the second polarizing plate, and the transparent film is placed between the first polarizing plate and the second polarizing plate, and the light is emitted through the first polarizing plate. The second polarizing plate receives the light emitted from the transparent film.

In the inspection apparatus of the patent document 2, the retardation film is inspected, and in the same manner as in the case of the patent document 1, the phase difference film is placed between the two polarizing plates, and the light is received by the light receiver. In the inspection apparatus, a cross nicol is arranged for each polarizing plate, and the vertical light receiving angle θ1 of the light receiver is set to 0°<θ1<90°, and the parallel light receiving angle θ2 is set to 0°< Θ2 < 90°, thereby detecting a phase difference defect caused by retardation in the thickness direction of the retardation film.

However, as described above, the patterned retardation film is alternately arranged with a phase difference region in which the directions of the plurality of optical axes are different from each other. Therefore, if the above special As in the case of Patent Document 1 and Patent Document 2, when only the polarizing plate is disposed, the brightness observed in each phase difference region is different. Therefore, there is a problem that the defect cannot be detected with high precision.

The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a defect detecting apparatus, method, and manufacturing method of a patterned retardation film which can accurately detect defects of a patterned retardation film.

In order to achieve the above-described problem, the defect detecting device for a patterned retardation film of the present invention includes: a first polarizing plate and a second polarizing plate, which are arranged to be orthogonally polarized with a patterning retardation film interposed therebetween, and patterned in the above-described patterning In the retardation film, a stripe-shaped first phase difference region and a second phase difference region in which a plurality of optical axes are substantially perpendicular to each other are alternately arranged side by side, and the region functions as a λ/4 wavelength plate; The inspection light is irradiated onto the patterned retardation film via the first polarizing plate; the imaging device captures the patterned retardation film via the second polarizing plate to obtain a luminance image; and the defect detecting unit detects the luminance image defect. The direction of one polarization transmission axis of the first polarizing plate and the second polarizing plate is adjusted so that any one of the polarization transmission axes is substantially parallel to any one of the first phase difference region and the second phase difference region. Each of the luminances when the normal first phase difference region and the second phase difference region are imaged by the imaging device reaches the same level in the vicinity of the extinction state. In this case, for the first phase difference region and the normal portion of the second phase difference region of the patterned retardation film taken at a low luminance level, the defective portion can be photographed with high luminance, and since the brightness of the normal portion is 1 phase difference region and phase 2 in the second phase difference region In the same way, defects can be detected with high precision.

It is preferable that when the patterned retardation film is laminated on the support including the transparent film to form the first retardation region and the second retardation region, the phase difference characteristic of the support is eliminated. The phase difference compensation plate is disposed between the first polarizing plate and the patterned retardation film or between the second polarizing plate and the patterned retardation film. Conveniently, the phase difference compensation plate is set to be the same transparent film as the support.

Preferably, the image processing unit includes a binarization circuit that binarizes the captured luminance image with a predetermined threshold value, and sets each pixel to a bright pixel or lower than a threshold value. a dark pixel of the threshold value; and a candidate defect extraction circuit, wherein a region in which a plurality of bright pixels are connected is a candidate defect region, and the defect detecting portion is adjacent to the candidate defect region and adjacent to each other in the direction in which the boundary line is arranged. The number of pixels between the candidate defect regions is counted, and when the number of pixels counted is equal to or less than a predetermined value, the relevant candidate defect region is set as a defect.

Preferably, the image processing unit erases a boundary line in the luminance image corresponding to the boundary between the first phase difference region and the second phase difference region, and the defect detecting unit has been erased based on the boundary line. The brightness image is used to detect defects.

Further, preferably, the image processing unit includes a binarization circuit that binarizes the captured luminance image with a predetermined threshold value, and sets each pixel to a bright pixel of a threshold or more. a dark pixel below the threshold; and a shrinkage processing circuit that shrinks the binarized luminance image by the number of times corresponding to the width of the boundary line, and eliminates the boundary line, and the shrinking process is A process of shrinking a region of a plain pixel in the direction in which the boundary lines are arranged.

Preferably, the image processing unit includes an expansion processing circuit that expands the contracted luminance image to restore the area of the bright pixel remaining in the contraction process to the size before the contraction process. The expansion process is a process of expanding a region of the plain pixel in the same direction as the shrinkage process, and the defect detection unit detects the defect based on the brightness image subjected to the expansion process.

Further, preferably, the image processing unit includes a difference image generation circuit that performs difference processing on each pixel on the luminance image, thereby generating a difference image in which the boundary line is substantially eliminated, and the difference processing is For the pixels on the luminance image, at least in the direction of the boundary line, the pixels of the pixel number corresponding to the boundary of the correlation pixel are separated from the boundary line as the opposite pixel, and the luminance of the autocorrelation pixel is reduced. The value obtained by removing the luminance of the pixel of the opposite party is set to a new value of the relevant pixel; and the binarization circuit binarizes the difference image with a predetermined threshold value, and sets each pixel as a threshold value The above bright pixels and dark pixels below the threshold.

Preferably, the difference image generating circuit performs differential processing between the opposite pixels in mutually different directions on the luminance image, thereby generating differential images in respective directions, and the binarization circuit has a predetermined threshold The values are binarized to generate a plurality of binarized images, and the defect detecting unit performs defect detection on the plurality of binarized images.

Further, the present invention includes: the first polarizing plate and the second polarizing plate are disposed so as to sandwich the retardation film; and the illumination light source irradiates the illumination light to the patterned retardation film via the first polarizing plate; /4 wavelength board, configuration Between the patterned retardation film and the second polarizing plate, the imaging condition changing unit is selectively set to a first state and a second state, and the first state is to pattern the transmission axis of the first polarizing plate When the optical axis of the retardation film is inclined by 45 degrees, the transmission axis of the second polarizing plate is brought to +45 degrees with respect to the optical axis of the λ/4 wavelength plate, and the second state is such that the transmission axis of the second polarizing plate is relatively The optical axis of the λ/4 wavelength plate is -45 degrees, and the imaging device captures the patterned phase via the λ/4 wavelength plate and the second polarizing plate in the first state and the second state generated by the imaging condition changing unit. The first film and the second image are obtained by the difference film, and the image combining unit matches the first phase difference region and the second phase difference region of the first image with the second image to make the first image and the first image The two images are superimposed to combine the first image and the second image, and the defect detecting unit detects the defect based on the composite image signal from the image combining unit. In this case, the defect of the patterned retardation film can be detected similarly to the general retardation film, and since the imaging is performed under the optical conditions following the principle of use, the accuracy can be accurately and accurately Test for defects that become problems in practical use.

Preferably, the image synthesizing unit changes the relative positions of the first image and the second image in a direction orthogonal to the boundary line between the first phase difference region and the second phase difference region during image synthesis. The first image is superimposed on the second image, and a composite image is produced at a relative position where the average luminance of the composite image after the overlap is minimized. Further, it is preferable that the brightness of the bright portion of the first phase difference region and the second phase difference region in the first image and the second image is equal to the dynamic range of the image combining unit and the imaging device. The dynamic range of the range is less than half the brightness of the narrow dynamic range, Shooting. Preferably, the image synthesizing unit generates a binarized signal for the first image and the second image, inverts the brightness of one binarized signal, and sets the inverted binary signal to obtain the inverted binary signal. Exclusive OR with another binarized signal.

Preferably, the pattern boundary erasing unit erases the pattern boundary portion based on the boundary between the first phase difference region and the second phase difference region in the binarized composite image generated by the image combining unit. Preferably, the pattern boundary erasing unit performs a reduction process on the pixel region including the boundary width of the first phase difference region and the second phase difference region as noise. Preferably, the patterned retardation film is a web that is transported in the column direction of the first phase difference region and the second phase difference region, and the imaging device includes a first imaging device body that acquires the first image, The mesh is photographed in synchronization with the conveyance of the mesh with the second image pickup apparatus main body that has acquired the second image.

In the method of detecting a defect of the patterned retardation film of the present invention, the adjustment step is such that the direction of the polarization transmission axis of any of the first polarizing plate and the second polarizing plate is different from the first phase difference region and the second phase difference region. The optical axis is adjusted in a substantially parallel range so that the patterned retardation film is disposed between the first polarizing plate and the second polarizing plate disposed orthogonally polarized, and is passed through the first polarizing plate. When the inspection light is irradiated onto the patterned retardation film, the luminances of the normal first phase difference region and the second phase difference region imaged by the second polarizing plate reach the same level in the vicinity of the extinction state, and the patterning phase difference is obtained. In the film, a stripe-shaped first phase difference region and a second phase difference region in which a plurality of optical axes are substantially orthogonal to each other are alternately arranged side by side, and the region functions as a λ/4 wavelength plate; and a photographing step The patterned retardation film is disposed between the first polarizing plate and the second polarizing plate which are adjusted in the direction of the polarization transmission axis, and irradiates the inspection light to the patterned retardation film via the first polarizing plate, and passes through the second polarizing plate. The phase difference film is photographed to obtain a brightness image; and the detecting step detects the defect based on the acquired brightness image.

Preferably, the patterned retardation film is laminated on the support including the transparent film to form the first retardation region and the second retardation region, and the support is removed in the adjustment step and the photographing step. The phase difference compensation plate having the phase difference characteristics is disposed between the first polarizing plate and the patterned retardation film or between the second polarizing plate and the patterned retardation film. It is simple to set the phase difference compensation plate to be the same transparent film as the support.

Preferably, the image processing unit includes a binarization step of binarizing the captured luminance image with a predetermined threshold value, and setting each pixel to a clear pixel or more below a threshold value. a dark pixel of the threshold value; and a candidate defect extraction step, wherein a region in which a plurality of bright pixels are connected is a candidate defect region, and the detection step is in the direction in which the boundary lines are arranged, and the candidate defect region and the adjacent other The number of pixels between the candidate defect regions is counted, and when the number of pixels counted is equal to or less than a predetermined value, the relevant candidate defect region is set as a defect.

Preferably, the erasing step includes erasing a boundary line in the luminance image corresponding to the boundary between the first phase difference region and the second phase difference region, and the detecting step is based on the luminance image in which the boundary line has been erased. Detect defects.

Preferably, the erasing step comprises: a binarization step to a prescribed threshold To binarize the captured luminance image, set each pixel to a bright pixel above the threshold and a dark pixel below the threshold; and a shrinking step to correspond to the width of the boundary line The number of times is to shrink the binarized luminance image and to remove the boundary line, which is a process of shrinking the region of the bright pixel in the direction in which the boundary lines are arranged.

Preferably, the erasing step includes an expansion processing step of expanding the contracted brightness image to restore the area of the bright pixel remaining in the shrinking process to the size before the shrinking process, the expansion process being The process of expanding the area of the bright pixel in the same direction as the shrinking process, the detecting step detects the defect based on the brightness image of the expanded process.

Preferably, the erasing step includes: a difference image generating step of performing differential processing on each pixel on the luminance image, thereby generating a difference image that substantially eliminates the boundary line, wherein the difference processing is for the luminance image The upper pixel will be at least the pixel of the pixel corresponding to the boundary line spacing in the direction of the boundary line, and the pixel of the autocorrelation pixel is subtracted from the luminance of the autocorrelation pixel. The value obtained by the brightness is set to a new value of the correlation pixel; and the binarization step is to binarize the difference image with a predetermined threshold value, and each pixel is set as a threshold above the limit value Prime and dark pixels below the threshold.

Preferably, the difference image generating step performs differential processing between the opposite pixels in mutually different directions on the luminance image, thereby generating difference images in respective directions, and the binarization step is performed with a predetermined threshold a value, respectively binarizing a plurality of difference images to generate a plurality of binarized images, and detecting steps Defect detection is performed on a plurality of binarized images.

Further, in the present invention, the first polarizing plate and the second polarizing plate are disposed so as to sandwich the retardation film, and the illumination light is applied to the patterned retardation film via the first polarizing plate by the illumination light source. The λ/4 wavelength plate is disposed between the patterned retardation film and the second polarizing plate, and is set to the first state and the second state, and is first by the imaging device via the λ/4 wavelength plate and the second polarizing plate. The first image is acquired by photographing in the state, and the second image is obtained by photographing in the second state, and the first state is such that the transmission axis of the first polarizing plate is inclined with respect to the optical axis of the patterned retardation film. In the state of 45°, the transmission axis of the second polarizing plate is brought to +45 degrees with respect to the optical axis of the λ/4 wavelength plate, and the second state is such that the transmission axis of the second polarizing plate is opposite to the λ/4 wavelength plate. When the optical axis reaches -45 degrees, the image combining unit overlaps the first phase difference region and the second phase difference region of the first image and the second image to form a composite image based on the composite image. To detect the above defects.

A method for producing a patterned retardation film according to the present invention includes a manufacturing step of producing a patterned retardation film in which a plurality of stripe first phase difference regions having a plurality of optical axes substantially orthogonal to each other and The second phase difference regions are alternately arranged side by side, and the region functions as a λ/4 wavelength plate; and the obtaining step is to arrange the patterned retardation film in the direction of the polarization transmission axis to adjust the first polarizing plate and the second Between the polarizing plates, the inspection light is irradiated onto the patterned retardation film via the first polarizing plate, the retardation film is imaged through the second polarizing plate, and a luminance image is obtained. The detection step is based on the acquired luminance image. , detecting defects; and adjusting steps, The pre-step is performed in a range in which the direction of the polarization transmission axis of any of the first polarizing plate and the second polarizing plate is substantially parallel to any one of the first phase difference region and the second phase difference region. In the state where the patterned retardation film is disposed between the first polarizing plate and the second polarizing plate, when the inspection light is irradiated to the patterned retardation film via the first polarizing plate, the second polarized light is transmitted. The luminances of the normal first phase difference region and the second phase difference region captured by the panel reach the same level in the vicinity of the extinction state.

Moreover, the present invention includes a manufacturing step of manufacturing a patterned retardation film in which a plurality of strip-shaped first phase difference regions and second phase difference regions having a plurality of optical axes orthogonal to each other are alternately arranged In the illumination step, the first polarizing plate and the second polarizing plate are disposed so as to sandwich the retardation film, and the illumination light is irradiated to the patterned phase difference via the first polarizing plate by the illumination light source. In the filming step, the λ/4 wavelength plate is disposed between the patterned retardation film and the second polarizing plate, and is set to the first state and the second state, and is passed through the λ/4 wavelength plate and the second by the imaging device. The polarizing plate captures the first image by capturing the image in the first state, and acquires the second image by capturing the image in the second state. The first state is to make the transmission axis of the first polarizing plate relative to the patterned phase. When the optical axis of the differential film is inclined by 45 degrees, the transmission axis of the second polarizing plate is brought to +45 degrees with respect to the optical axis of the λ/4 wavelength plate, and the second state is such that the transmission axis of the second polarizing plate is opposite to the transmission axis. The optical axis of the λ/4 wavelength plate reaches -45 degrees; the image synthesis step makes A first retardation region image and the second image and the second retardation region is consistent overlap, a synthetic image; and a step of detecting, based on the synthesis of Image to detect defects.

According to the present invention, it is possible to accurately detect a defect of the patterned retardation film, and in the patterned retardation film, a stripe first phase difference region and a second phase difference in which a plurality of optical axes are substantially orthogonal to each other The regions are alternately arranged side by side, and the above regions function as λ/4 wavelength plates.

[First Embodiment]

In FIG. 1, the defect detecting device 10 is disposed on a manufacturing line of a patterned retardation film. The stripe patterned retardation film (hereinafter simply referred to as retardation film) 12 from the retardation film forming step is continuously supplied to the defect detecting device 10, and the retardation film 12 is subjected to defect detection. In the following description, the longitudinal direction of the retardation film 12 is defined as the X direction, and the width direction of the retardation film 12 orthogonal to the X direction is referred to as the Y direction.

As shown in FIG. 2, in the retardation film 12, the first phase difference region 14 and the second phase difference region 15 having different phase difference characteristics of the phase difference layer 12b are alternately arranged in the Y direction. In the retardation film forming step, the retardation film 12 is formed with a retardation layer 12b on the surface of the support 12a including a transparent film such as Triacetyl Cellulose (TAC), and the retardation layer 12b exhibits the first phase. The phase difference between the difference region 14 and the second phase difference region 15 is obtained.

As shown in FIG. 1, the first phase difference region 14 and the second phase difference region 15 have a strip shape extending in the longitudinal direction of the retardation film 12, and are alternately arranged at a fixed pitch in the Y direction. The width (length in the Y direction) of each of the phase difference regions 14 and 15 is, for example, 270 μm, and each phase difference region 14, 15 are arranged in the Y direction at the same pitch as the above width. Further, in Fig. 1, the widths of the phase difference regions 14, 15 are exaggeratedly drawn.

With reference to the X direction (0°), as shown in FIG. 3, the first phase difference region 14 is formed in such a manner that the optical axis (late phase axis or phase advance axis), for example, the phase axis A1 is only tilted. The λ/4 wavelength plate of "-45° (=-θ)" functions, and one second phase difference region 15 (see Fig. 1) is formed in such a manner that the retardation axis As2 is only tilted. "+45° (= +θ)" λ/4 wavelength plate functions. Therefore, the retardation film 12 is a patterned retardation film formed such that the slow axis A1 of the first phase difference region 14 and the slow axis A2 of the second phase difference region 15 are orthogonal to each other. Such a retardation film 12 is attached to, for example, a liquid crystal screen and is used to observe a stereoscopic image.

In addition, as described below, the retardation axes As1 and As2 of the retardation film 12 need not be strictly inclined by ±45° with respect to the X direction, and the direction as a reference is also arbitrary, as long as the phase difference regions 14 and 15 are delayed. When the patterned retardation film having substantially the same axis axis is orthogonal to each other, the patterned retardation film serves as an inspection target. Moreover, the patterned retardation film which forms each phase difference area so that the advancing axes are orthogonal to each other, and the slow axis is orthogonal to each other, is also inspected.

As shown in FIG. 1, the defect detecting apparatus 10 includes a light source unit 16, an imaging device 17, a first polarizing plate 18, a second polarizing plate 19, and a defect detecting unit 20. Assuming that the 3D television is the final product using the retardation film 12, when the viewer observes the 3D television normally, for example, as long as a defect having a diameter of about 100 μm can be detected, the method can be determined in this way. Each specification of each part in the defect detecting device 10 is determined. The reason is that the TV viewer does not recognize a defect having a size of 100 μm or less, and therefore, there is no problem in practical use. Further, as a defect, foreign matter, dirt, damage, or uneven light distribution (local disturbance of phase difference characteristics) may be mentioned.

In the defect detecting device 10, the retardation film 12 supplied from the retardation film forming step is continuously conveyed in the X direction by a transport mechanism (not shown). In the transport path of the retardation film 12, two guide rollers (not shown) are arranged at predetermined intervals. The retardation film 12 is hung on the guide roller, whereby the retardation film 12 is maintained in a planar shape on an inspection stage between the respective guide rollers.

The light source unit 16, the imaging device 17, the first polarizing plate 18, and the second polarizing plate 19 are disposed in the inspection platform. The light source unit 16 and the imaging device 17 are disposed on the opposite side of each other with a transport path interposed therebetween. In the present embodiment, the light source unit 16 includes a light source such as a halogen lamp. In the present embodiment, the inspection light is uniformly applied to the lower surface of the retardation film 12 in the width direction via the first polarizing plate 18 from the lower side of the transport path. Further, the light source of the light source unit 16 may be another lamp such as a metal halide lamp, or may be a light emitting diode (LED) or the like.

The imaging device 17 captures the luminance image of the phase difference film 12 from the normal direction of the phase difference film 12 via the second polarizing plate 19 on the upper side of the transport path. The photographing device 17 is a linear array camera including a photographing lens 17a, a charge coupled device (CCD) line sensor 17b, and the like, and uses one shot at a time. a pair of parallel lines in the width direction of the retardation film 12 The line is photographed, and the CCD line sensor 17b includes a plurality of light receiving elements arranged in a line. The value of each pixel of the luminance image is the amount of light that is emitted from the portion of the corresponding phase difference film 12 and transmitted through the second polarizing plate 19, that is, the brightness observed through the second polarizing plate 19. Each time the phase difference film 12 of one line is conveyed, the photographing device 17 performs one shot, whereby the brightness image of the phase difference film 12 passing through the inspection stage is sequentially photographed. The pixel resolution of the photographing device 17, that is, the length in the Y direction on the retardation film 12 corresponding to one pixel of the luminance image is, for example, 10 μm/pixel. Further, instead of using a line camera, an area camera that simultaneously photographs a plurality of lines may be used as the photographing device 17.

The first polarizing plate 18 is disposed between the light source unit 16 and the retardation film 12 in a posture parallel to the retardation film 12, and the second polarizing plate 19 is disposed in a posture parallel to the retardation film 12. Between the retardation film 12 and the photographing device 17. Each of the polarizing plates 18 and 19 is a linearly polarizing polarizing plate. Further, in the so-called orthogonal polarization arrangement, the polarization transmission axis P1 of the first polarizing plate 18 and the polarization transmission axis P2 of the second polarizing plate 19 are orthogonal to each other.

The direction of the first polarizing plate 18 and the second polarizing plate 19 with respect to the polarization transmission axes P1 and P2 of the retardation film 12 is adjusted so that the normal first phase difference region 14 without defects is imaged by the imaging device 17. When the second phase difference region 15 is formed, the phase difference regions 14 and 15 reach the same brightness level. The above adjustment is performed in a range in which the polarization transmission axis P1 is substantially parallel to the slow axis A1, and the polarization transmission axis P2 is substantially parallel to the slow axis A2. Thereby, close to the matting In the state of the state (near the extinction state), the luminances of the first phase difference region 14 and the second phase difference region 15 reach the same level, and the extinction state is transmitted through the normal first phase difference region 14 and the second phase difference region. The inspection light of 15 does not pass through the state of the second polarizing plate 19.

The retardation film 12 forms the first phase difference region 14 and the second phase difference region 15 such that the slow phase axes As1 and As2 are orthogonal to each other. Actually, the retardation film A1 may deviate from the slow phase axis As1 for various reasons. The state in which As2 is orthogonal to each other. Therefore, for example, when the polarization transmission axis P1 is adjusted to be parallel to the slow axis A1, the first phase difference region 14 may be in an extinction state in which the inspection light is not passed, but the second phase difference region 15 is not in the In the matte state, the luminance levels of the first phase difference region 14 and the second phase difference region 15 are different. As described above, when the difference in luminance is generated in each of the phase difference regions 14 and 15, the defect cannot be accurately detected based on the luminance image of the imaging device 17. Therefore, the adjustment is performed such that, in the state of being close to the extinction state, the luminances of the first phase difference region 14 and the second phase difference region 15 reach the same level as described above, and the accuracy can be adjusted. Defect detection is performed well.

In this example, the normal first phase difference region 14 and the second phase are adjusted so that the slow axis A1 of the first phase difference region 14 and the polarization transmission axis P1 of the first polarizing plate 18 are substantially parallel. The respective luminances of the phase difference regions 15 are at the same level. According to the above, it is understood that the reason for the substantially parallel state in the above-described manner is that the state is close to the extinction state. Here, in the state close to the extinction state, the luminances of the normal first phase difference region 14 and the second phase difference region 15 are the same. The level means that the angle between the angle formed by the retardation axes As1 and As2 and 90° is Δθ, and the angle formed by the polarization transmission axes P1 and P2 is maintained at 90. In the state of (orthogonal polarization), the polarization transmission axis P1 is shifted by Δθ/2 with respect to the slow axis A1, and the polarization transmission axis P2 is shifted by Δθ/2 with respect to the slow axis A2.

In addition, by adjusting the polarization transmission axis of any of the first polarizing plate 18 and the second polarizing plate 19, it is possible to achieve either of the optical axes of the first phase difference region 14 and the second phase difference region 15 (late phase axis or The phase advancement axis is substantially parallel, and the luminances of the normal first phase difference region 14 and the second phase difference region 15 may be at the same level. Therefore, adjustment may be made such that the normal first phase difference region 14 and the second phase are in a state where the polarization transmission axis P1 of the first polarizing plate 18 and the slow axis A1 of the first phase difference region 14 are substantially orthogonal to each other. The brightness of the difference region 15 reaches the same level. Further, adjustment may be made such that the normal first phase difference region 14 and the second phase difference are in a state in which the polarization transmission axis P1 and the first phase difference region 14 are substantially parallel to each other or in an orthogonal state. The brightness of the area 15 reaches the same level.

In addition, the first polarizing plate 18 and the second polarizing plate 19 are arranged in an orthogonal polarization, and the optical axes of the first phase difference region 14 and the second phase difference region 15 (for example, the slow phase axes As1 and As2) are substantially Orthogonal. Therefore, the polarization transmission axis of one polarizing plate and the optical axis of one phase difference region are substantially parallel, which is equivalent to the optical axis of one polarizing plate and the optical axis of the other phase difference region. Orthogonal state; the optical axis of the polarizing transmission axis of the other polarizing plate and the optical phase of the other phase difference region The substantially parallel state is set; and the polarization transmission axis of the other polarizing plate and the optical axis of one phase difference region are substantially orthogonal to each other.

The luminance image from the photographing device 17 is sent to the defect detecting unit 20. The defect detecting unit 20 detects a defect of the phase difference film 12 based on the luminance image obtained by the photographing device 17. As shown in FIG. 4, the defect detecting unit 20 includes an image processing unit 21 and a defect detecting unit 22, and the image processing unit 21 includes a D/A converter (digital-to-analog converter) 23, a memory (memory) 24, The binarization circuit 25, the contraction processing circuit 26, the expansion processing circuit 27, and the noise removing circuit 28.

The D/A converter 23 makes the luminance level of each pixel of the luminance image into digital data. The memory 24 is capable of memorizing the luminance images of a plurality of lines, and sequentially stores the luminance images input line by line, thereby generating a two-dimensional luminance image. Further, the memory 24 is also used as a work memory for processing the respective portions of the defect detecting unit 20.

A two-dimensional luminance image including a predetermined number of lines, for example, 400 lines, is generated in the memory 24, and the two-dimensional luminance image is subjected to subsequent image processing as one processing unit. A two-dimensional luminance image including 400 lines is generated in such a manner as to overlap between the front and rear luminance images, so that defects near the boundary between the two-dimensional luminance images can be detected without fail. Each time a luminance image having a predetermined number of lines is generated in the memory 24, each portion of the image processing unit 21 reads the luminance image, and then performs predetermined processing on the luminance image, and then the luminance is again performed. The image is saved to the memory 24.

The identification data is given to the brightness image. This identification data indicates the history of the processing performed by the image processing unit 21, and each time the image processing unit 21 is used. When the part is processed, the content of the above identification data is updated. Each part of the image processing unit 21 recognizes whether or not the above-described processing is performed on the luminance image stored in the memory 24 based on the identification data, and reads the luminance image from the memory 24 if there is a luminance image to be subjected to the above processing. For example, the above processing is carried out. Thereby, the image processing unit 21 sequentially performs predetermined image processing on the two-dimensional luminance image.

The binarization circuit 25 binarizes the luminance level with respect to each pixel of the luminance image stored in the memory 24 by comparing the luminance level with a predetermined threshold value. By this binarization, a pixel whose luminance level is lower than the threshold value is set to a dark pixel whose value is "0", and a pixel whose luminance level is equal to or greater than the threshold value is set to a value of "1". Picture. In the above binarization process, the pixels corresponding to the normal portions of the phase difference regions 14 and 15 are dark pixels, and the luminance is higher than the normal portion, and the pixel that becomes the candidate defect is a bright pixel. Therefore, after adjusting the direction of the polarization transmission axis P1 of the first polarizing plate 18, the threshold value is set based on the luminance obtained from the normal portion.

In a state where the first polarizing plate 18 and the second polarizing plate 19 are adjusted, the luminances of the normal portions of the first phase difference region 14 and the second phase difference region 15 are close to the extinction state, and both of them are the same. Brightness level. The threshold value for the above binarization may be a brightness level higher than the brightness level of the normal portion described above. The threshold value is determined based on the brightness portion above which the brightness level is equal to the candidate defect. In general, a sample of a defect generated in advance or an actually generated defect sample may be placed in an experimental machine of the adjusted defect detecting device 10 or an equivalent optical system, and then the brightness of the defect is measured to determine Above brightness level.

The shrinkage processing circuit 26 erases the boundary line or noise of the phase difference regions 14, 15 included in the luminance image. Although the width of the boundary between the first phase difference region 14 and the second phase difference region 15 is narrow, the phase difference characteristics are not good, and the brightness is increased similarly to the defect. Therefore, in the two-dimensional luminance image, the boundary line is a predetermined number of pixels in the Y direction, for example, two pixels to three pixels, and the luminance on the straight line along the X direction becomes high.

In order to erase the boundary line, the contraction processing circuit 26 performs a contraction process on the binarized luminance image. Further, by the shrinking process, the bright pixels having a small length in the Y direction and being noise are eliminated. Each pixel of the boundary line is converted into a bright pixel by binarization. The shrinkage process is performed so that the width of the bright pixel is reduced in the Y direction of the arrangement direction of the phase difference regions 14 and 15. The shrinking process is performed in the number of times the boundary line can be erased only, that is, the number of times corresponding to the width of the boundary line (the number of pixels in the Y direction). By eliminating the boundary line in the above manner, the influence of the boundary of each of the phase difference regions 14, 15 is eliminated, and the defect can be accurately detected.

In the primary contraction process, for the pixel on the gaze luminance image, when the other side of the pixel adjacent to the pixel in the same line, for example, the right side is a dark pixel (the value is "0"), For each pixel, a process of setting the attention pixel to a dark pixel (the value is "0") is performed. That is, on the same line, the pixel at the right end of the bright pixel region is removed as a dark pixel. Therefore, for example, when the width of the boundary line is 2 pixels, the contraction process is performed twice, and when the width of the boundary line is 3 pixels, the contraction process is performed three times, thereby eliminating the boundary line. According to the resolution of the photographing device 17 and the phase difference film 12 The width of the boundary line determines the number of shrinkage processes to be performed. It is preferable to consider the deviation in the normal range of the width of the boundary line, and determine the number of times of the above-described contraction processing by experiment and based on the width (number of pixels) of the boundary line on the luminance image.

Further, in the above-described contraction processing, the pixel at the right end of the bright pixel region is removed as a dark pixel every time the processing is performed, but the pixel at the left end of the bright pixel region may be set as a dark image. Removed.

The expansion processing circuit 27 performs expansion processing on the luminance image after the contraction processing. By this expansion processing, the length in the Y direction of the region of the bright pixels remaining in the luminance image is restored. Thereby, when the defect determination is performed, the defect size can be accurately determined. In the Y direction, the expansion process is performed by the inverse logic of the contraction process, and for the attention pixel on the luminance image, the pixel adjacent to the attention pixel on the left side of the same line is a bright pixel (value is In the case of "1"), the process of setting the fixation pixel to the bright pixel (the value is "1") is performed for each pixel. The above expansion process is performed in the same number of processes as the shrinkage process.

Further, in the present example, when the defect detection is performed, in order to easily extract the defect or determine the defect based on the number of pixels, the expansion process is performed to restore the size of the area of the remaining bright pixels. Therefore, if it is determined that the defect is not based on the size of the region of the bright pixel, the expansion process can be omitted. Moreover, in this example, since the shrinking process and the expansion process are performed only in the Y direction, the shrinkage process and the expansion process may be performed line by line without making the brightness image two-dimensional. Further, in this example, the shrinkage and expansion treatment are performed only for the Y direction, but the shrinkage may be performed in the X direction. The expansion process is performed by simultaneously removing the bright pixels which are small in the X direction and become noise.

The noise removing circuit 28 removes the area of the bright pixel having a predetermined number of pixels or less as noise from the luminance image. In this example, after the shrinking process and before the expansion process, the clear pixels (isolated points) of one of the surrounding pixels which are all dark pixels are eliminated as noise. The above-described erasing is performed by converting the bright pixels into dark pixels. In addition, the noise removal by the noise removing circuit 28 is to eliminate the bright pixels of the noise which are caused by various factors such as electrical noise, or the area (the number of pixels) is as small as the following. In the region of the pixel, the above-mentioned degree means that the defect is not a problem even if the defect is not detected, and is not limited to the region of the bright pixel of one pixel, and the bright painting in which a plurality of bright pixels are connected. The area of the prime is removed as a noise. Further, the noise of the luminance image before the shrinking process may be removed, or the noise may be removed after the expansion process and immediately before the defect detection.

The defect detecting unit 22 detects a defect based on the brightness image after the expansion process. The defect detecting unit 22 extracts the region of the bright pixel on the luminance image during the detection of the defect, and then determines the extracted region as the defective portion. Further, the length, the size, and the shape of the X-direction and the Y-direction of the defect can be determined by checking the number of pixels of the region of the bright pixel. That is, it is also possible to determine whether or not the region of the bright pixel is a defective portion based on the length or size of the region of the bright pixel. When a defect is detected, the defect detecting portion 22 generates and outputs defect information including the position of the defect on the retardation film 12 or the size of the detected defect. The The defect information is used as control information, and the purpose of the control information is not to use the defective portion as an article.

Next, the action of the above configuration will be described. Before the defect inspection by the defect detecting device 10, as shown in Fig. 5, the direction of the polarization transmission axis is first adjusted. First, a predetermined normal portion of the retardation film 12 to be inspected is placed between the first polarizing plate 18 and the second polarizing plate 19 of the inspection platform. Then, the polarization-transmission axes P1 and P2 of the first polarizing plate 18 and the second polarizing plate 19 are orthogonal to each other to be in a state of orthogonal polarization.

While maintaining the orthogonal polarization state and the parallel state parallel to the retardation film 12, coarse adjustment is performed, and the polarization adjustment axis P1 of the first polarizing plate 18 and the first phase difference region 14 are delayed by the coarse adjustment. The axis As1 is parallel. At this time, it is not necessary to make the polarization transmission axis P1 and the slow axis A1 strictly parallel, and it is possible to be substantially parallel. In the state in which the light source unit 16 is turned on, when the first retardation region 14 or the second retardation region 15 is observed via the second polarizing plate 19, the matte state is achieved. The adjustment may be performed, and the above-described extinction state is a state in which the amount of transmitted light is minimized.

After the coarse adjustment, the light source unit 16 and the imaging device 17 are activated so that the respective brightness levels of the first phase difference region 14 and the second phase difference region 15 can be referred to based on the luminance image captured by the imaging device 17. Then, the fine adjustment is performed such that the luminance levels of the phase difference regions 14 and 15 are the same.

In the micro adjustment, one side maintains the orthogonal polarization state, and the phase difference film 12 is in a parallel parallel state, and one side is set so that the brightness levels are the same. In this fine adjustment, the direction in which the polarization transmission axis P1 and the slow axis A1 are substantially parallel and close to the extinction state is maintained without changing the direction of the polarization transmission axis P1, and the direction of the polarization transmission axis P1 is faced. Make adjustments.

The inspection light from the light source unit 16 is linearly polarized via the first polarizing plate 18, and then incident on the retardation film 12. When the polarization transmission axis P1 is accurately parallel to the slow axis A1 of the first phase difference region 14, the inspection light of the linearly polarized light emitted from the first polarizing plate 18 does not pass through the first phase difference region 14, A phase difference will result. Therefore, the inspection light is incident on the second polarizing plate 9 in a state of being linearly polarized. Then, the polarization direction of the inspection light incident on the second polarizing plate 19 is orthogonal to the polarization transmission axis P2 of the second polarizing plate 19, and therefore, it is not emitted from the second polarizing plate 19.

Further, when the polarization transmission axis P1 is substantially parallel to the slow axis A1 of the first phase difference region 14, and is not exactly parallel, the inspection light is phase-diffused when it passes through the first phase difference region 14, and becomes elliptically polarized. Then, it is incident on the second polarizing plate 19. Therefore, in this case, in the inspection light that has become elliptically polarized, the polarization component orthogonal to the polarization transmission axis P2 is not emitted from the second polarizing plate 19, but the polarization component parallel to the polarization transmission axis P2 is from the first. 2 The polarizing plate 19 is emitted. Further, the more the direction relationship between the polarization transmission axis P1 and the slow axis A1 is largely deviated from the parallel state, the larger the polarization component parallel to the polarization transmission axis P2, and therefore the amount of light emitted from the second polarizing plate 19 When the size is increased, the brightness of the first phase difference region is increased.

On the other hand, the same is true for the second phase difference region 15, but The relationship between the polarization transmission axes P1 and P2 with respect to the slow axis A2 is opposite to that of the first phase difference region 14. In other words, when the polarization transmission axis P1 is accurately orthogonal to the slow axis A2 of the second phase difference region 15, the polarization direction of the inspection light of the linearly polarized light from the first polarizing plate 18 is accurately orthogonal to the slow axis A2. Therefore, even if it passes through the second phase difference region 15, it will enter the second polarizing plate 19 without causing a phase difference. Further, since the polarization direction of the inspection light is orthogonal to the polarization transmission axis P2 of the second polarizing plate 19, it is not emitted from the second polarizing plate 19.

Further, when the polarization transmission axis P1 is substantially orthogonal to the slow axis A1 of the second phase difference region 15, and is not orthogonally orthogonal, the inspection light is phase-diffused when it passes through the second phase difference region 15, and becomes elliptically polarized. . Among the inspection lights that have become elliptically polarized, only the polarization component parallel to the polarization transmission axis P2 is emitted from the second polarizing plate 19. When the direction relationship between the polarization transmission axis P1 and the slow axis A2 is largely shifted from the orthogonal state, the amount of light emitted from the second polarizing plate 19 increases as the polarization component parallel to the polarization transmission axis P2 increases. The brightness becomes higher.

When the retardation film 12 as follows is an inspection object, the increase or decrease of the retardation axis As1 and the polarization transmission axis P1 with respect to the parallel shift is an offset with respect to the orthogonal phase axis A2 and the polarization transmission axis P1 with respect to the orthogonal direction. The increase and decrease are the same, and the magnitude of the offset is the same. The retardation film 12 is formed with the first phase difference region 14 and the second phase difference region 15 such that the slow axis axes As1 and As2 are accurately orthogonal to each other. Therefore, the polarization transmission axis P1 is in a state parallel to the slow axis A1, and the polarization transmission axis P1 is in a state orthogonal to the slow axis A2. Therefore, the first phase difference region 14 and the second phase difference are obtained. The inspection light of the area 15 is not emitted from the second polarizing plate 19, and the brightness of any of the areas can be adjusted to "0" (extinction state).

As described above, in the retardation film 12, the ideal state is a state in which the slow phase axes As1 and As2 are exactly orthogonal to each other, but in many cases, the orientation of the slow phase axis is formed due to the following phase difference region formation. The error phase and the like, the retardation axis As1 and the slow phase axis As2 are not exactly orthogonal, but slightly deviate from the orthogonal state, and the phase difference region exhibits birefringence even if the support 12a is slightly smaller. In this case, the polarization transmission axis P1 cannot be brought into a state parallel to the slow axis A1, and the polarization transmission axis P1 cannot be made orthogonal to the slow axis A2. In other words, the fine adjustment cannot be performed in such a manner that the luminance levels of the first phase difference region 14 and the second phase difference region 15 are simultaneously set to "0", and the two phase difference regions are captured.

However, the offset of the slow axis A1 and the polarization transmission axis P1 with respect to the parallel direction and the offset of the slow axis A2 and the polarization transmission axis P1 with respect to the orthogonal direction can be made the same, so that the first phase difference region 14 is transmitted. The polarization components of the inspection light emitted from the second polarizing plate 19 in the second phase difference region 15 have the same magnitude. Therefore, the luminances of the first phase difference region 14 and the second phase difference region 15 can be finely adjusted to the same brightness level in a state close to the extinction state, which is not "0" but the luminance level is low.

As described above, after the polarization transmission axis P1 is finely adjusted, the brightness level obtained by the imaging device 17 for capturing the normal portion of the first phase difference region 14 or the second phase difference region 15 is slightly higher than the brightness level. The threshold value is set to the binarization circuit 25. Then, start the phase difference film 12 Perform defect inspection.

After the start of the defect inspection, the retardation film 12 is continuously supplied to the defect detecting device 10 from the retardation film forming step, and the retardation film 12 is further transported downstream via the inspection platform. In the transport process, the inspection light from the light source unit 16 is irradiated to the retardation film 12 via the first polarizing plate 18, and each time a phase difference film 12 of one line is transported, a line is taken by the photographing device 17. The amount of brightness image.

Whenever shooting is performed, the luminance image of the amount of one line is sent from the photographing device 17 to the defect detecting unit 20, and after the D/A converter 23 makes the luminance level of each pixel into digital data, the digital data is written to Memory 24. Thereby, a two-dimensional luminance image is generated. When the luminance image of the predetermined number of lines is written to the memory 24, the luminance image is read from the memory 24 by the binarization circuit 25, and then the pixels are set to the previously set threshold. Compare to achieve binarization.

As shown in the example of FIG. 6A, in the binarized luminance image I, the dark pixel region D1 corresponding to the normal portion of the first phase difference region 14 and the second phase difference region 15 is included. In addition to the line D2, the bright pixel area D3 to the bright pixel area D7 is also included. Since the luminance of the normal portion of the first phase difference region 14 and the second phase difference region 15 is lower than the threshold value, the normal portion becomes the strip-shaped dark pixel region D1 of the dark pixel. When the width of each of the phase difference regions 14 and 15 is set to about 270 μm and the pixel resolution of the imaging device 17 is set to 10 μm/pixel, each of the dark pixel regions D1 is arranged in the Y direction by about 27 dark. A pixel with a long strip in the X direction. Similarly to the defective portion, the boundary line D2 is adjacent to each of the dark pixel regions D1. The phase difference characteristic is disordered and appears as a bright pixel. Each of the boundary lines D2 has, for example, about two to three bright pixels arranged in the Y direction, and has a linear shape extending in the X direction.

In the first phase difference region 14 and the second phase difference region 15, the phase difference characteristics of the defective portion such as the portion in which the foreign matter is mixed, the portion where the damage is present, and the portion where the light distribution is uneven are disturbed. Therefore, when When the inspection light of the linearly polarized light is transmitted, the direction of the polarization is disturbed, and the component emitted from the second polarizing plate 19 is made larger than the normal portion. Therefore, the brightness of the defective portion becomes high and becomes a bright pixel, and the area size of the bright pixel corresponds to the size of the defective portion. Further, for example, when electrical noise is superimposed on a pixel of a luminance image, the pixel becomes a bright pixel. The bright pixel region D3 to the bright pixel region D7 corresponds to a defective portion or noise as described above.

By the shrinkage processing circuit 26, the binarized luminance image is read from the memory 24, and the contraction processing related to the Y direction is performed a predetermined number of times, for example, three times. Thereby, as shown in FIG. 6B, the boundary line D2 is erased as a dark pixel. Further, the bright pixel area D4 to the bright pixel area D7 corresponding to the noise in the Y direction and corresponding to the noise or the minute defective portion are simultaneously erased as the dark pixels. The area of the bright pixel which is removed by the shrinkage processing is small, and it does not become a problem in practical use. Further, when the boundary line is thicker than the desired width, even if the shrinkage process is performed a predetermined number of times, the moiré of the portion remains, and therefore, it can be detected as a defect.

The noise removal circuit 28 removes the noise of the luminance image after the predetermined number of shrinking processes. Thereby, the area of the small bright pixel which becomes the noise which is not removed by the shrinking process is erased from the luminance image.

Then, the expansion processing circuit 27 performs the expansion processing for a predetermined number of times, for example, three times, and the width of the bright pixel region of the contracted luminance image is restored. Thereby, as shown in FIG. 6C, the width of the bright pixel region D3 remaining in the shrinking process is restored to the width before the shrinking process.

After the expansion processing, the defect detecting unit 22 extracts the bright pixel region remaining in the luminance image, and determines the bright pixel region as the defective portion. Thereby, in FIG. 6C, the bright pixel region D3 is detected as a defective portion. Next, defect information including position information, defect size, and the like is generated and output, and the position information indicates the position of the detected defective portion.

[Second Embodiment]

In the second embodiment, the luminance of the attention pixel on the luminance image is compared with the luminance of the pixel separated by only a predetermined number of images, and the difference is obtained as a new value of the attention pixel. So, the boundary line is eliminated. In the same manner as in the first embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and the detailed description of the components will be omitted.

As shown in FIG. 7, the image processing unit 21 in this example includes a D/A converter 23, a memory 24, a difference image generating circuit 31, a binarization circuit 25, and a noise removing circuit 28. The difference image generation circuit 31 performs differential processing on the attention pixel on the luminance image of one line that has been digitized by the D/A converter 23, which is a pixel that separates only a predetermined number of images as The opponent's pixel acquires the difference between the brightness of the above-mentioned attention pixel and the brightness of the opponent's pixel, and sets the difference as a new value of the attention pixel. The above difference processing is performed on each pixel of the luminance image, thereby generating a difference image. When the difference is obtained, the value obtained by subtracting the value of the opponent's pixel from the value of the gaze pixel is set as a new value of the gaze pixel.

The difference image is generated by erasing the boundary line between the first phase difference region 14 and the second phase difference region 15 in the luminance image, and when the pixel is the pixel on the boundary line, the pixel of the opposite direction is used. The way to be a pixel on the boundary line determines the distance (pixel number) of the pixel at the time of the difference. Since each pixel on the boundary line has substantially the same brightness level, the new value of the pixel on the boundary line is set to substantially “0”, and the boundary line is substantially eliminated.

The direction of the self-care pixels toward the opponent's pixels may be at least separated in the Y direction, and may be the two directions of the X direction and the Y direction. As shown in FIG. 8, the number of pixels in the X direction up to the opponent pixel Px2 is set to "x1" pixels, and the number of pixels in the Y direction is set to "y1" with respect to the attention pixel Px1. In the case of a single pixel, "y1" is an interval of the boundary line, that is, n times (n = 1, 2, 3, ...) of the number of pixels corresponding to the width of the phase difference region (the length in the Y direction). "x1" can be set to any value. For example, when "x1" is set to "0", and when the pixel Px1 is the pixel on the boundary line, the pixel Px2 is a pixel on the other boundary line on the same line.

In addition, "y1" may be set to n times or more of the number of pixels corresponding to the width of the phase difference region (interval of the boundary line), and x1 may be set to "1" or more. In this case, since the difference between the pixels on the line separated by the "x1" line is obtained, when the pixel Px1 is the pixel on the boundary line, the pixel Px2 is on a different line. The pixels on other boundary lines.

Further, when it is possible to produce a bright spot (hereinafter referred to as a stripe defect) which is generated in a stripe shape in a fixed angular direction with respect to the X direction, The values of x1 and y1 in which the pixels of the stripe defect are not compared with each other are set so as not to erase the bright spot.

For example, when the stripe defect is likely to occur in a direction of 45° with respect to the X direction, in order not to eliminate the stripe defect, the relative pixel Px2 is not in the direction of 45° of the gaze pixel Px1. The combination of x1" and "y1". In this case, it is necessary to determine the combination of "x1" and "y1" in consideration of the pixel resolutions in the X direction and the Y direction. As in the case of the present embodiment, when the photographing device 17 is a line camera, the pixel resolution in the Y direction is determined based on the imaging resolution of the photographing device 17, but the photographing period and the retardation film 12 in accordance with one line of the photographing device 17 are used. The transfer speed determines the pixel resolution in the X direction. For example, when the pixel resolution in the X direction is 6 μm and the pixel resolution in the Y direction is 3 μm, “x1: y1=1: 2” is not satisfied in order to eliminate the stripe defect in the 45° direction.

The binarization circuit 25 binarizes the difference image with a predetermined threshold value, and sets each pixel as a bright pixel and a dark pixel. The noise removing circuit 28 erases the area of the minute bright pixels from the binarized difference image as noise. The defect detecting unit 22 determines the region of the bright pixel remaining after the processing of the noise removing circuit 28 as the defective portion.

Further, when the difference image is formed as described above, unlike the contraction processing of the first embodiment, the small bright pixel region is not erased at the generation stage of the difference image, and is derived from the photographing device 17 The error of the resolution and the deviation of the edge portion of the boundary line itself may cause a large amount of small bright pixel regions to remain. Therefore, when a difference image is generated for defect detection, it is preferable to erase the noise.

According to this example, the difference between the gaze pixel and the opposite pixel is obtained for each pixel of the luminance image, and a difference image is generated, and each pixel on the boundary line of the luminance image and the other party on the same or different boundary line are obtained. The difference of pixels. Further, since the luminances of the pixels on the respective boundary lines are at substantially the same level, the new value of the attention pixel is substantially "0". Therefore, if the pixel is binarized, the pixel will become a dark pixel, and as a result, the boundary line is erased.

On the other hand, when the pixel is the pixel of the defective portion, in most cases, the pixel of the normal portion whose luminance level is relatively low is the pixel of the opposite party, and therefore, the new value of the pixel of the defective portion is less likely. When a change occurs, the pixel of the defect portion becomes a bright pixel by binarization. Therefore, the defect portion can be detected as a defect without being erased.

Furthermore, in order to detect stripe-shaped defects in all directions such as streaky defects generated in a random direction, it is also possible to look at the pixels as a base point and obtain the plurality of opponents in the different directions. The difference of the pixels generates a difference image associated with each direction based on the same luminance image, and then detects a defect from each of the difference images. Thereby, a pixel at a position different in the ratio of x1 and y1 (x1/y1) can be generated as a difference image as each of the respective pixels. For example, x1 and y1 are set so that the direction of the opponent's pixel from the gaze pixel is 45°, and a difference image in the 45° direction is created, and the opponent is calculated from the gaze pixel. The direction of the pixel is 30°, and x1 and y1 are set to form a difference image in the direction of 30°. Next, defect detection is performed on each difference image, and the defective portion determined by the difference image in the 45° direction and the difference image in the 30° direction are confirmed. The predetermined defect portion is collectively set as the defective portion of the phase difference film 12. Furthermore, in this example, the two directions are 45° and 30°, but the direction (angle) and the number of directions may be appropriately set.

[Third embodiment]

In the third embodiment, a region of a pixel (bright pixel) having a high luminance in a two-dimensional luminance image is extracted as a candidate defect, and candidates adjacent to each other in the Y direction are selected for each candidate defect. The number of pixels up to the defect is counted as the adjacent distance pixel number, and when the counted adjacent distance pixel number is equal to or less than a predetermined value, the candidate defect is determined as a defect. In the same manner as in the first embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and the detailed description of the components will be omitted.

In the present example, as shown in FIG. 9, the image processing unit 21 includes a D/A converter 23, a memory 24, a binarization circuit 25, and a counter circuit 41. The pixels of the luminance image of one line are digitized by the D/A converter 23 and written to the memory 24, and then become a bright pixel or a dark pixel by the binarization circuit 25.

The counting circuit 41 sets the bright pixel regions in which a plurality of bright pixels on the luminance image are connected as candidate defects. The counting circuit 41 counts the number of pixels up to the candidate defect adjacent to the Y direction in one line for each candidate defect as the adjacent distance pixel number Cp. When the adjacent distance pixel number Cp counted by the counter circuit 41 is equal to or less than the predetermined pixel number Cpth, the defect detecting unit 22 determines the candidate defect as the defective portion. The specified pixel number Cpth is set to be larger than the boundary line D2 on the luminance image. A smaller value than Cy.

Furthermore, the bright pixel region may be set as a candidate defect based on a bright pixel, but in this example, there is no problem even if the bright pixel region is not detected as a noise or a practical defect. Therefore, such a bright pixel area is not set as a candidate defect.

As shown in the example of FIG. 10, the bright pixel region which is a candidate defect in the luminance image I includes the boundary line D2 in addition to the bright pixel region D8 which is actually a defect. In this way, the boundary line D2 is set as the candidate defect, but in the normal portion, the boundary line D2 is separated from each other by a predetermined number of pixels Cpth in the Y direction (Cp>Cpth), and therefore, the boundary line D2 is not determined as the defective portion. . On the other hand, when there is a defect in the first phase difference region 14 or the second phase difference region 15, or when the width of the boundary line is abnormally thick, the bright pixel region is larger than the boundary line with respect to other bright pixel regions. The intervals Cy are closer to each other. Therefore, the adjacent distance pixel number Cp is equal to or less than the predetermined pixel number Cpth, and the bright pixel area as described above is determined as the defective portion. Further, when the candidate defect other than the boundary line D2 approaches in the vicinity of the boundary line D2, the boundary line D2 is also determined to be defective, but there is no actual damage in the operation of the candidate defect inspection.

This example takes advantage of the fact that the boundary lines are separated by a fixed interval (the number of pixels), and the extremely simple logic can be used to extract the defects. When the direction in which the adjacent distance pixel number Cp is counted is the Y direction, the count can be counted in either direction from the candidate defect, and the count can be performed in both directions. Further, similarly to the other embodiments, a removal circuit that sets the number of pixels to a predetermined value (for example, 2) may be provided. The bright pixel area of the following pixels is removed as noise.

[Fourth embodiment]

In the fourth embodiment, the phase difference compensation plate for eliminating the phase difference characteristic of the support of the retardation film is provided, and the phase difference characteristic of the support is canceled. In addition, the first embodiment to the third embodiment can be used in addition to the phase difference compensation plate.

In the fourth embodiment, as shown in FIG. 11, the phase difference compensation plate 37 is disposed between the retardation film 12 and the second polarizing plate 19. The phase difference compensation plate 37 eliminates the phase difference characteristics of the support 12a of the retardation film 12, and is disposed in a posture parallel to the retardation film 12. The phase difference compensation plate 37 is not limited as long as it is a phase difference compensation plate that eliminates the phase difference characteristic of the support 12a. In the simplest manner, the phase difference compensation plate 37 can be transparently used in the same manner as the support 12a of the retardation film 12 to be inspected. The film is also the same in this example. In this case, the phase difference compensation plate 37 is disposed such that the optical axis is orthogonal to the optical axis (for example, the slow phase axis) of the support 12a.

According to this example, the luminance levels of the first phase difference region 14 and the second phase difference region 15 can be made closer to the extinction state, and the contrast of the luminance image can be improved. Thereby, the defect of the phase difference film 12 can be detected with higher precision. In the present embodiment, the phase difference compensation plate 37 is disposed between the retardation film 12 and the second polarizing plate 19, but the phase difference compensation plate 37 may be disposed on the first polarizing plate 18 and the retardation film 12. between.

[Fifth Embodiment]

As shown in FIG. 12, the defect detecting device 50 of the patterned retardation film includes a light source unit 51, a photographing device 52, a first polarizing plate 53, and a λ/4 wavelength plate. 54. The second polarizing plate 55 and a controller 56. Between the light source unit 51 and the imaging device 52, the first polarizing plate 53, the λ/4 wavelength plate 54, and the second polarizing plate 55 are disposed in this order from the light source unit 51 side. Between the first polarizing plate 53 and the λ/4 wavelength plate 54, the patterned retardation film 60 as the inspection target film is located close to the first polarizing plate 53. The image pickup signal from the photographing device 52 is sent to the image processing unit 57 in the control unit 56, and the image processing unit 57 detects various defects.

The patterned retardation film 60 is cut into a sheet shape by a size corresponding to a display used after passing through the retardation film forming step, and the patterned phase is formed by a film supply device (not shown). The poor film 60 is positioned at the inspection position. The patterned retardation film 60 includes patterns 60A, 60B which are alternately arranged on a transparent TAC film as a support by phase difference layers having different optical axes and having λ/4 wavelength plate characteristics. Made. Each of the patterns 60A and 60B is formed in a strip shape, for example, at a width of 270 μm. For one adjacent stripe pattern 60A, when the longitudinal direction (column direction) of the pattern, that is, the Y direction is set to the reference 0°, the optical axis which is the characteristic of the λ/4 wavelength plate is + with respect to the reference 0°. 45°, for another pattern 60B, the optical axis is -45°. Moreover, the direction orthogonal to the longitudinal direction (Y direction) of the patterns 60A and 60B is called the pattern arrangement direction (X direction) when the patterns 60A and 60B are arranged.

When a 3D television is considered as a final product using the patterned retardation film 60, it is assumed that a television viewer performs normal observation, for example, detecting light leakage (bright spots) having a diameter of about 100 μm, foreign matter, or dirt, so that Defects do not flow downstream. TV viewers will not recognize The size is less than 100 μm, so there is no problem in practical use.

The light source unit 51 includes an illumination surface 51A that is larger than the size of the cut piece to be inspected, and incorporates a light source such as a halogen lamp, and irradiates uniform scattered light to the inspection target film. In addition, although the illustration is omitted, the light source unit 51 includes a light amount adjustment unit that controls the halogen lamp so that the amount of light is fixed based on the light amount detection signal detected by the sensor provided near the light source. Thereby, light having a uniform amount of light can be irradiated onto the patterned retardation film 60, so that defect detection can always be performed with the same sensitivity. Further, other lamps such as a metal halide lamp or an LED may be used as the light source.

The photographing device 52 includes a CCD area sensor, and for the patterned retardation film 60, a pixel resolution of 10 μm/pixel is used. The photographing device 52 focuses on the surface of the patterned phase difference layer of the film 60 to perform photographing.

The control unit 56 includes, for example, a personal computer. Further, by installing an application, the image processing unit 57, the imaging condition changing unit 61, and the defect data storage unit 63 are constructed in the control unit 56, and the image processing unit 57 performs the defect signal based on the imaging signal. In the detection, the imaging condition changing unit 61 changes the imaging condition, and the defect data storage unit 63 identifies the defect detected by the detection signal and the portion of the defect.

The transmission axis angle of the first polarizing plate 53 is set to the reference 0°. Thereby, the light that has been patterned by the retardation film 60 is circularly polarized light having different rotation directions by the respective patterns 60A and 60B. The transmitted light that becomes the above-described circularly polarized light is imaged through the λ/4 wavelength plate 54 and the second polarizing plate 55. λ/4 wavelength plate 54 and 2nd The polarizing plate 55 functions as a circularly polarized spectacles at the time of 3D observation. Therefore, similarly to the circularly polarized glasses, the light receiving optical system acquires the two images of the first image 65 and the second image 66 under the following conditions, and the above conditions refer to the optical with respect to the λ/4 wavelength plate 54. The axis is such that the transmission axis of the second polarizing plate is ±45° (if one transmission axis is +45°, the other transmission axis is -45°). For example, when visualizing a 3D image, the first image 65 is a right-eye image, and the second image 66 is a left-eye image.

As shown in FIG. 12, the specific light receiving optical system has two types, that is, a light receiving optical system in which the transmission axis of the second polarizing plate 55 is set to +45° with respect to the λ/4 wavelength plate 54 whose mounting position is fixed, A light receiving optical system in which the transmission axis of the second polarizing plate 55 is set to -45°. Therefore, the second polarizing plate 55 is rotated by the rotation mechanism 64 so as to have a transmission axis of +45° and -45° in an angular range of 90°. The state in which the transmission axis is set to +45° is the first state, and the state in which the transmission axis is set to -45° is the second state. The rotation mechanism 64 is controlled by the imaging condition changing unit 61 in the control unit 56. When the second polarizing plate 55 is in the first state and the first image 65 is obtained, the control unit 56 drives the rotating mechanism 64 to rotate the second polarizing plate 55 by 90° to reach the second state, thereby obtaining the first state. 2 image 66. After acquiring the second image 66, the control unit 56 returns the second polarizing plate 55 to the first state.

Further, the second polarizing plate 55 may be fixed instead of fixing the λ/4 wavelength plate 54. In this case, the optical axis of the λ/4 wavelength plate 54 (in the phase axis or the slow phase axis) The λ/4 wavelength plate 54 is rotated within an angular range of 90° by a rotation mechanism (not shown) such that any one of the axes is set to +45° and -45°. Thus, the λ/4 wavelength plate 54 or the polarizing plate 55 is made Any one of them is rotationally shifted in the range of 90° to obtain the first image 65 and the second image 66 under two different conditions.

As shown in FIG. 12, the first image 65 and the second image 66 are controlled by the rotation mechanism 64, and are set to different optical conditions again, and are imaged by one imaging device 52, as shown in FIG. The first image 65 and the second image 66 are obtained by using the two imaging units 82 and 83 and performing imaging using respective dedicated optical systems. Further, the λ/4 wavelength plate 54 or the second polarizing plate 55 is rotationally shifted to change the imaging conditions, and then the first image 65 and the second image 66 are obtained, but they may be adjacently provided. The λ/4 wavelength plate 54 having the optical axes of +45° and -45° is set, and any optical axis is set as the photographic optical axis, whereby the imaging conditions are changed. Similarly, instead of the rotational displacement of the second polarizing plate 55, the second polarizing plate 55 having the transmission axes of +45° and -45° may be adjacently provided, and any one of the transmission axes may be set as the imaging optical axis. Thereby the photography conditions are changed.

In the fifth embodiment, first, the linearly polarized light orthogonal to each other is received by the imaging device 52 via the second polarizing plate 55. At this time, the rotation mechanism 64 is driven, and the transmission axis angle of the second polarizing plate 55 is set to +45°, and imaging is performed to obtain the first image 65. Moreover, the rotation mechanism 64 is driven, and the transmission axis angle of the second polarizing plate 55 is set to -45°, and imaging is performed to obtain the second image 66.

As shown in FIG. 13A and FIG. 13B, in the first image 65 and the second image 66, the patterns 60A and 60B (see FIG. 12) are obtained, and the brightness and darkness of the portions corresponding to the respective patterns are reversed. image. At this time, the two images are used by the light amount control of the light source unit 51 or the aperture of the imaging device 52. The bright portions of the patterns of the 65 and 66 normal portions are adjusted such that the brightness of the bright portion reaches the intermediate range of the dynamic range of the brightness of the image processing apparatus and the dynamic range of the narrow range of the dynamic range of the photographing apparatus. . Specifically, if the brightness level is 256 gray levels, the intermediate brightness level is 128 gray levels. On the other hand, since the amount of light in the dark portion generated by the patterns 60A and 60B is extremely low, the luminance level is substantially "0". Here, if there is a bright spot (light leakage) in the dark portion image portion of the pattern, as shown in FIGS. 13A and 13B, high-brightness bright spots (bright defects) BP1, BP2 appear. Further, if there is a light-shielding defect such as foreign matter or dirt in the bright image portion of the pattern, low-brightness bright spots (dark defects) BP3 and BP4 appear.

As shown in FIG. 12, the signals of the first image 65 and the second image 66 from the imaging device 52 are sent to the image processing unit 57. In the image processing unit 57, the defects are extracted by the processing sequence shown in Figs. 14 and 15 . First, after the first image and the second image (ST1 to ST4) are imaged, the image combining unit 70 causes the brightness of the bright portion of each of the patterns 60A and 60B of the patterned retardation film 60 to reach the image. The dynamic range of the processing unit 57 and the substantially intermediate luminance of the dynamic range having a narrow range in the dynamic range of the imaging device are added to the first image 65 and the second image 66 (ST5). When image synthesis is performed by the above-described addition, as shown in ST6 to ST9 of FIG. 15, the relative position is gradually changed in the X direction orthogonal to the boundary portion 68 of each pattern, and the first position is made for each relative position. The average luminance when the image 65 coincides with the second image 66 is calculated. The above calculation is performed within a fixed range, and the first image 65 and the second image 66 are superimposed on the relative position at which the minimum average luminance is reached. Thereby, by a simple configuration, as shown in FIG. 13C, The composite image 67 is obtained, and the composite image 67 is an image obtained by superimposing the first image 65 and the second image 66 at positions where the first phase difference region and the second phase difference region match each other. Therefore, the first image 65 and the second image 66 can be simply placed in a state where the positions of the respective patterns of the patterned retardation film 60 are accurately aligned without using a high image processing method. synthesis. Further, the boundary portion is actively used, and the two are superimposed on the basis of the boundary portion, whereby the alignment in the X direction can be easily performed without positional displacement.

In the state where the average luminance level is set to 128 gray scales (the intermediate gray scale when the gray scale level is 256 gray scales), image synthesis is performed on the first image 65 and the second image 66. The composite image 67 is a substantially uniform density image in each of the patterns. Further, the brightness of the first image 65 and the second image 66 is reversed at the same position as each pattern of the patterned retardation film 60, and a composite image 67 in a state in which the state is referred to as the first image is obtained. The bright defects BP1, the bright defects BP2, the dark defects BP3, and the dark defects BP4 of the 65 and the second image 66 maintain the luminance difference between the brightness levels of the periphery and the brightness of the periphery.

Then, the synthesized image 67 is binarized using a predetermined threshold (ST10). Based on the binarized composite images 71 and 72, candidate defects are extracted. Regarding the threshold value of the binarization processing, the two threshold values are used for the high luminance side threshold and the low luminance side threshold with respect to the intermediate luminance (128 gray level). Thereby, for example, as shown in FIG. 13D, the optical leakage defects (bright defects) BP1, BP2 are extracted from the binarized composite image 71 using the high luminance side threshold value higher than the intermediate luminance, as shown in the figure. As shown in 13E, since The binarized composite image 32 having a low luminance side threshold value lower than the intermediate luminance is used, and the light-shielding defects (dark defects) BP3 and BP4 caused by foreign matter and dirt are extracted.

As shown in FIG. 13D and FIG. 13E, even if binarization of the bright defect extraction or the dark defect extraction is performed, the pattern boundary portion remains as the abnormal portion 69 in the binarized composite images 71 and 72. Therefore, the amount of the pixel slightly larger than the predetermined boundary width in the 90° direction (the width direction (X direction) in the image) (about 3 pixels in the condition of the present embodiment) The valued composite images 71 and 72 perform reduction processing for noise (ST11). When the pattern boundary portion remains as the abnormal portion 69, the contraction processing is performed in the number of times the boundary portion can be erased and the number of times corresponding to the width of the boundary portion (the number of pixels in the Y direction).

For example, the phrase "one pixel shrinks" means "reducing one pixel to the top, bottom, left, and right". When shrinking three pixels in the X direction, "shrinking three pixels to the left or right" = "to the left" Or right, shrink one pixel three times." In Fig. 13D, since the bright portion is a defect, the contraction is performed on the basis of the bright portion, and the boundary portion is eliminated as shown in Fig. 13F. Further, in Fig. 13E, since the dark portion is a defect, the contraction is performed on the basis of the dark portion, and the boundary portion is eliminated as shown in Fig. 13G. The boundary portion is eliminated in the above manner, whereby the influence of the boundary can be eliminated, so that the defect can be detected with high precision.

By the reduction processing, since the width of the abnormal portion 69 due to the boundary portion of the pattern is smaller than the reduced pixel, the abnormal portion 69 is completely erased as shown by the reduction processing images 73 and 74 of FIGS. 13F and 13G, and the other abnormal portion 69 is completely eliminated. aspect, Since the size of the bright spot or the foreign matter to be detected is about 100 μm, the size of the bright spot or the foreign matter in the width direction is at least close to 10 pixels, and is not eliminated by the above-described reduction processing. Next, only the bright defects BP1 and BP2 and the dark defects BP3 and BP4 larger than the predetermined size are extracted. In the present embodiment, a defect of less than 100 μm is regarded as a defect that is not problematic in practical use, and a defect of 100 μm or more is extracted as a bright defect and a dark defect (ST12). According to a well-known method, for example, the extracted bright defects BP1, BP2 and dark defects BP3, BP4 are determined in the screen position, and the position data and the defect information are collectively stored in the defect data storage unit 63 (ST13).

Furthermore, instead of removing the boundary line by the noise removal method by pixel reduction, the stripe in the longitudinal direction of the image can be removed, thereby eliminating the boundary line portion. Further, the streak defect appearing at a fixed pitch in the lateral direction of the image may be determined as a boundary portion and eliminated. As described above, by removing the noise caused by the boundary portion, only the bright defects BP1 and BP2 and the dark defects BP3 and BP4 can be extracted, and the luminance defect or the foreign matter defect can be accurately determined.

[Sixth embodiment]

Next, instead of the patterned retardation film 60 cut into a sheet shape, referring to FIG. 16 and FIGS. 17(A) to 17(F), the defect inspection device 81 of the patterned retardation film 80 connected in a mesh shape is subjected to the defect inspection device 81. Description. The sixth embodiment is different from the fifth embodiment in that a CCD line image sensor is used as a photographing device to continuously acquire a web in order to continuously acquire an image for the meshed patterned retardation film 80 that has migrated. A photographic image of the object. therefore, The imaging unit 82 dedicated to the first image, the imaging unit 83 dedicated to the second image, and the control unit 84 are provided to continuously acquire an image. The imaging units 82 and 83 are basically the same configuration as the imaging unit of the fifth embodiment, and the imaging conditions for the first image and the imaging conditions for the second image are also the same as those of the fifth embodiment. However, in the sixth embodiment, since the first image-only imaging unit 82 and the second image-only imaging unit 83 are provided, the rotation mechanism 64 or the imaging condition changing unit 61 of the fifth embodiment is not required. Wait.

The first image capturing unit 82 includes a light source unit 91, an imaging device 92, a first polarizing plate 93, a λ/4 wavelength plate 94, and a second polarizing plate 95. Between the light source unit 91 and the imaging device 92, the first polarizing plate 93, the λ/4 wavelength plate 94, and the second polarizing plate 95 are sequentially disposed from the light source unit 91 side. The patterned retardation film 80 is located between the first polarizing plate 93 and the λ/4 wavelength plate 94. Further, the second image capturing unit 83 is configured similarly to the first image capturing unit 82, and includes a light source unit 101, an imaging device 102, a first polarizing plate 103, a λ/4 wavelength plate 104, and a second Polarizing plate 105.

The control unit 84 includes an image processing unit 96 and a defect data storage unit 97. The imaging signals from the imaging devices 92 and 102 are transmitted to the image processing unit 96 of the control unit 84, and the image processing unit 96 detects various defects.

The image processing unit 96 includes a binarization circuit 106, an image inversion circuit 107, a mutually exclusive OR operation circuit 108, and a noise removal circuit 109. First, as shown in FIGS. 17(A) and 17(B), image signals of respective lines obtained by the line sensor are obtained. Since each of the photographing units 82 and 83 is disposed in the moving direction of the web, the amount of separation is used to determine the moving direction of the web. The pixels on the same line are on each other. Then, as shown in FIG. 17(C) and FIG. 17(D), in the same manner as in the fifth embodiment, the binarization circuit 106 is used, and the high luminance side threshold and the low luminance side threshold are used. The first image signal and the second image signal of one line are binarized.

Next, as shown in FIG. 17(E), the image inversion circuit 107 inverts the binarized data of the second image. Next, as shown in FIG. 17(F), the inverted signal and the binarized signal of the first image are calculated by the exclusive OR operation circuit 108, and the outputs are mutually exclusive. Thereby, an output signal in which a portion having a bright defect and a dark defect is determined to be "1" is obtained. Further, when there is a defect at the same position of the first image and the second image, the exclusive OR operation circuit 108 outputs "0", but in terms of the nature of the defective image, since it is impossible to simultaneously A dark defect in a bright image and a bright defect in a dark image are generated, and therefore, even if "0" is specifically output, there is no problem. For example, in the bright portion of the pattern of the normal portion of the first image, a portion such as a foreign matter that blocks light and becomes a dark defect exists in the dark portion in the second image, and since the portion originally shields the light, the Will become a defect. There are cases where, for example, a foreign matter is embedded in a film to impair the polarization characteristics around the foreign matter, but even in this case, the foreign matter that becomes the core of the defect is only a dark defect, and the polarization characteristic around the foreign object is only abnormal. It is recognized as a bright defect (bright spot), and the positions of the above-mentioned dark defects and bright defects (bright spots) are close, but they are necessarily different. In this way, the inverted output signal of the other image signal is mutually exclusive or calculated with respect to one image signal, whereby the bright defect and the dark defect can be clearly displayed on the same screen by a simple configuration. Furthermore, when making Figure 17 (A) ~ Figure 17 (F) When the first image data of FIG. 17(A) and the second image data of FIG. 17(B) are aligned in the X direction, as shown in FIG. 15, the first image data can be obtained. The average brightness at the time of synthesizing the second image data is the smallest relative position, and then the two image data are aligned based on the relative position.

Further, when noise occurs due to leakage of light at the boundary portion of each of the patterns 60A and 60B, the noise is removed by the noise removing circuit 109 as in the fifth embodiment. In this manner, even for the patterned retardation film that continuously moves, the bright defects and the dark defects can be accurately detected.

In the sixth embodiment, since the image signal is binarized at the initial stage of image processing, the processed image data can be greatly reduced, and the defect inspection can be performed at high speed. Therefore, even if a high-speed and expensive image processing apparatus is not used, it is possible to comprehensively inspect the defects of the moving mesh in real time. Further, as shown in FIG. 16, in the sixth embodiment, the light source units 91 and 101 are individually provided in the respective imaging units 82 and 83. However, the illumination area of one light source unit may be enlarged to provide a light source. To carry out lighting.

Further, in the sixth embodiment, the defect is detected line by line based on the image signal of the line sensor, but the image area sensor can be used to acquire the image signals of the plurality of lines at one time. Defects are detected based on image signals for each line. Further, in the sixth embodiment, instead of the sheet-shaped patterned retardation film 60, the mesh-shaped patterned retardation film 80 is placed at the inspection position, and a web reservoir is provided and the mesh film is provided. Advancing intermittently, defect inspection is performed for each fixed section, and the mesh reservoir temporarily accumulates a mesh having a length exceeding the photographing range, whereby the entire surface of the mesh film can be continuously detected for defects.

[Example 1]

In the example 1, the defect detecting device 10 configured as described in the first embodiment is used to actually perform defect inspection on the phase difference film 12. In the defect inspection, the retardation film 12 is set as an inspection object, and the retardation film 12 has TAC as the support 12a.

The width of the first phase difference region 14 and the second phase difference region 15 formed in the retardation film 12 is 270 μm. Further, the pixel resolution in the Y direction in the luminance image captured by the imaging device 17 is 10 μm/pixel. The boundary between the first phase difference region 14 and the second phase difference region 15 is a boundary line (bright pixel) having two to three pixel numbers in the Y direction on the luminance image, and therefore is performed in the Y direction. Only shrinking the three-contraction process of three pixels, and then eliminating the bright pixel region of one pixel as noise, performing a three-expansion process of expanding only three pixels in the Y direction, and then the bright pixel region Determined as a defective part. As a result, a defect having a diameter of about 100 μm can be detected.

[Example 2]

In the example 2, the phase difference compensation plate 37 is provided as in the fourth embodiment, and the phase difference film 12 is subjected to defect detection. The defect detecting device 10 used for the defect detection has the same configuration as that of the second embodiment except that the phase difference compensating plate 37 is provided, and a difference image is prepared to detect the defect. The phase difference compensation plate 37 is the same as the support 12a of the retardation film 12. The normal portion of the transparent film is disposed such that the optical axis of the phase difference compensation plate 37 is orthogonal to the optical axis of the support 12a. In the second example, the width of the first phase difference region 14 and the second phase difference region 15 is 270 μm, and the pixel resolution is set to 3 μm/pixel in order to detect the defect size with higher precision. Therefore, on the luminance image, the boundary line is the width of 5 to 7 pixels in the Y direction.

By arranging the phase difference compensation plate 37, as described above, the first phase difference region 14 and the second phase difference region 15 of the phase difference film 12 are all closer to the extinction state, and the brightness level of the normal surface of the image is higher than that of the example 1. Low (darker), the contrast of the bright spots is improved. The contrast is improved, and accordingly, the threshold value at the time of binarization can be set to a low threshold value, and the defective portion having a small amount of light transmitted through the second polarizing plate 19 can be converted into a bright pixel, thereby achieving high precision. Detect defects.

With respect to each pixel of the luminance image, "y1=90" is generated, and a difference image is generated between the pixels of the opposite pixel that are separated by 90 pixels in the Y direction. As a result, for each pixel on the boundary line, the new value obtained by taking the pixel on the other boundary line adjacent to the boundary line as the opponent pixel is approximately "0". Thereby, the boundary line can be eliminated from the luminance image by binarization. After the binarization, a defect can be detected from an image with sufficient precision as an image obtained by canceling the bright pixel region of two pixels or less as noise. It can be detected even if it is a minute defect that is only the same as the width of the boundary line.

10, 50‧‧‧ Defect detection device

12‧‧‧Phase retardation film / retardation film

12a‧‧‧Support

12b‧‧‧ phase difference layer

14, 15‧‧‧ phase difference area

16, 51, 91, 101‧‧‧ Light source department

17, 52, 92, 102‧ ‧ photographic equipment

17a‧‧‧Photographic lens

17b‧‧‧CCD line sensor

18, 19, 53, 55, 93, 95, 103, 105‧‧‧ polarizing plates

20‧‧‧ Defect detection unit

21, 57, 96‧‧‧Image Processing Department

22‧‧‧Defect Detection Department

23‧‧‧D/A converter

24‧‧‧ memory

25, 106‧‧‧ Binarization Circuit

26‧‧‧Shrinkage processing circuit

27‧‧‧Expansion processing circuit

28, 109‧‧‧ noise removal circuit

31‧‧‧Differential image generation circuit

37‧‧‧ phase difference compensation board

41‧‧‧Counting circuit

51A‧‧‧Lighting surface

54, 94, 104‧‧‧λ/4 Wavelength Board

56, 84‧‧‧Control Department

60‧‧‧ patterned retardation film

60A, 60B‧‧‧ pattern

61‧‧‧Photographic Condition Change Department

63, 97‧‧‧Defects in the Memory of Defects

64‧‧‧Rotating mechanism

65, 66‧‧‧ images

67‧‧‧Composite images

69‧‧‧Exception Department

70‧‧‧Image Synthesis Department

71, 72‧‧‧ Binarized composite image

73, 74‧‧‧ Reduced processing image

80‧‧‧ patterned retardation film

81‧‧‧ Defect inspection device

82, 83‧‧‧Photographic unit

107‧‧‧Image reversal circuit

108‧‧‧Exclusive or arithmetic circuit

As1, As2‧‧‧ late phase axis

BP1, BP2‧‧‧ Defects

BP3, BP4‧‧‧Dark defects

Cp‧‧‧adjacent distance

Cy‧‧‧ interval

D1‧‧‧Dark pixel area

D2‧‧‧ boundary line

D3~D8‧‧‧ bright pixels area

I‧‧‧Brightness image

P1, P2‧‧‧ polarized transmission axis

Px1‧‧‧ gaze

Px2‧‧‧Participating pixels

ST1~ST13‧‧‧Steps

X, Y‧‧ direction

X1, y1‧‧ ‧ prime numbers

Θ‧‧‧ angle

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a perspective view showing the configuration of a defect detecting device embodying the present invention. Figure.

2 is a cross-sectional view showing a layer structure of a patterned retardation film.

3 is an explanatory view showing a slope of a slow phase axis of each phase difference region.

4 is a block diagram showing the configuration of a defect detecting unit.

Fig. 5 is a flow chart showing the processing procedure for inspection.

FIG. 6A is an explanatory diagram showing an example of a luminance image that has been binarized.

FIG. 6B is an explanatory diagram showing an example of a luminance image in which boundary lines, noise, and minute defective portions are erased by the contraction process.

FIG. 6C is an explanatory diagram showing an example of a luminance image in which the width of the bright pixel region of the luminance image is restored by the expansion processing.

Fig. 7 is a block diagram showing a defect detecting unit in an example in which a boundary line is erased by generating a difference image.

FIG. 8 is an explanatory diagram showing a distance of a counterpart pixel when a difference image is generated.

FIG. 9 is a block diagram showing a defect detecting unit in an example of determining whether or not there is a defect based on the number of pixels up to the adjacent candidate defect.

FIG. 10 is an explanatory diagram showing an example in which a defect is determined based on the number of pixels up to the candidate defect adjacent in the Y direction.

Fig. 11 is a side view showing an example in which a phase difference compensation plate is provided.

FIG. 12 is a perspective view showing a defect detecting device of a sheet-shaped patterned retardation film.

FIG. 13A is an explanatory diagram showing an example of a first image.

FIG. 13B is an explanatory diagram showing an example of a second image.

13C is a diagram showing an example of a composite image of the first image and the second image. Illustration of the diagram.

FIG. 13D is an explanatory diagram showing an image when a bright defect is extracted based on a composite image. FIG.

FIG. 13E is an explanatory diagram showing an image when a dark defect is extracted based on a composite image. FIG.

Fig. 13F is an explanatory diagram showing an image obtained by erasing the boundary portion from the image shown in Fig. 13D.

Fig. 13G is an explanatory diagram showing an image obtained by erasing the boundary portion from the image shown in Fig. 13F.

Fig. 14 is a flow chart showing the processing procedure of defect detection.

Fig. 15 is a flowchart showing a processing procedure of the X-direction positional alignment at the time of image composition.

Fig. 16 is a perspective view showing a defect detecting device of a mesh-shaped patterned retardation film.

17(A) to 17(F) are timing charts showing a method of extracting a defect based on the first image signal and the second image signal.

As1, As2‧‧‧ late phase axis

X‧‧‧ direction

Θ‧‧‧ angle

Claims (34)

A defect detecting device for patterning a retardation film, which is a defect detecting device for detecting a defect of a patterned retardation film, wherein the plurality of optical axes are substantially orthogonal to each other in the patterned retardation film The phase difference region and the second phase difference region are alternately arranged side by side, and the first phase difference region and the second phase difference region function as a λ/4 wavelength plate, and the patterned retardation film defect detecting device The first polarizing plate and the second polarizing plate are arranged to be orthogonally polarized so as to sandwich the patterned retardation film, and the light source unit irradiates the inspection light to the pattern via the first polarizing plate. a retardation film; an imaging device that images the patterned retardation film through the second polarizing plate to obtain a luminance image; and a defect detecting unit that detects the defect from the luminance image to the first polarizing plate And adjusting a direction of one polarization transmission axis of the second polarizing plate so that any one of the polarization transmission axis and the first phase difference region and the second phase difference region are In a state in which one optical axis is substantially parallel, each of the luminances when the normal first phase difference region and the second phase difference region are imaged by the imaging device reaches the same level in the vicinity of the extinction state. The defect detecting device of the patterned retardation film according to claim 1, wherein the patterned retardation film laminates the phase difference layer to include a transparent film The support body is configured to form the first phase difference region and the second phase difference region, and the defect detecting device for the patterned retardation film includes a phase difference compensation plate, and the phase difference compensation plate is disposed on the first polarization beam The plate is interposed between the patterned retardation film or the second polarizing plate and the patterned retardation film, and the phase difference characteristic of the support is eliminated. The defect detecting device for a patterned retardation film according to claim 2, wherein the phase difference compensation plate is the same transparent film as the support. The defect detecting device for a patterned retardation film according to claim 3, further comprising an image processing unit, wherein the image processing unit corresponds to a boundary between the first phase difference region and the second phase difference region The boundary line in the luminance image is erased, and the defect detecting unit detects the defect based on the luminance image in which the boundary line has been erased. The defect detecting device of the patterned retardation film according to claim 4, wherein the image processing unit includes: a binarization circuit that performs the above-described luminance image with a predetermined threshold value; For each value, each pixel is a bright pixel having a threshold value or more and a dark pixel lower than the threshold value; and a candidate defect extraction circuit is provided with a region in which a plurality of the bright pixels are connected to each other. Alternate defect area, The defect detecting unit counts a pixel number between the candidate defect region and another adjacent candidate defect region in the direction in which the boundary line is arranged, and when the counted pixel number is equal to or smaller than a predetermined value, The related candidate defect region described above is set as the above defect. The defect detecting device of the patterned retardation film according to claim 4, wherein the image processing unit includes: a binarization circuit that performs the above-described luminance image with a predetermined threshold value; For each value, each pixel is set to a bright pixel above the threshold and a dark pixel lower than the threshold; and the shrink processing circuit has a number of times corresponding to the width of the boundary line, The value-enhanced luminance image is subjected to a contraction process to erase the boundary line, and the contraction process is a process of shrinking the region of the bright pixel in the direction in which the boundary lines are arranged. The defect detecting device of the patterned retardation film according to claim 6, wherein the image processing unit includes an expansion processing circuit that expands the brightness image subjected to the shrinking process to cause expansion processing The area of the bright pixel remaining in the shrinkage process is restored to the size before the shrinkage process, and the expansion process is a process of expanding the region of the bright pixel in the same direction as the shrinkage process, and the defect detecting unit The above defects are detected based on the above-described luminance image subjected to the above expansion processing. Lack of patterned retardation film as described in claim 4 a trapping detection device, wherein the image processing unit includes a difference image generating circuit that performs differential processing on each of the pixels on the luminance image, thereby generating a difference image that substantially eliminates the boundary line, The difference processing is a pixel on the luminance image, and the pixel of the pixel number corresponding to the boundary of the boundary is separated from the correlation pixel by at least the arrangement direction of the boundary line as a pixel of the opposite party. a value obtained by subtracting the luminance of the opposite pixel from the luminance of the relevant pixel as a new value of the correlation pixel; and a binarization circuit binarizing the difference image with a predetermined threshold value Each pixel is set to a bright pixel above the threshold and a dark pixel below the threshold. The defect detecting device of the patterned retardation film according to claim 8, wherein the differential image generating circuit performs the differential processing between the opposite pixels in mutually different directions on the luminance image. Thereby, each of the difference images in each direction is generated, and the binarization circuit binarizes the plurality of difference images by a predetermined threshold value to generate a plurality of binarized images. The defect detecting unit performs defect detection on the plurality of binarized images. A defect inspection device for patterning a retardation film, which is a defect detecting device for detecting a defect of a patterned retardation film, wherein the plurality of optical axes are substantially orthogonal to each other in the patterned retardation film 1 The phase difference region and the second phase difference region are alternately arranged side by side, and the first phase difference region and the second phase difference region function as a λ/4 wavelength plate, and the defect inspection device of the patterned retardation film The first polarizing plate and the second polarizing plate are disposed so as to be interposed between the patterned retardation film, and the illumination light source irradiates the illumination light to the patterned retardation film via the first polarizing plate; The λ/4 wavelength plate is disposed between the patterned retardation film and the second polarizing plate, and the imaging condition changing unit is selectively set to a first state and a second state, wherein the first state is The transmission axis of the first polarizing plate is inclined by 45° with respect to the optical axis of the patterned retardation film, and the transmission axis of the second polarizing plate is +45 degrees with respect to the optical axis of the λ/4 wavelength plate. In the second state, the transmission axis of the second polarizing plate is set to -45 degrees with respect to the optical axis of the λ/4 wavelength plate, and the imaging device is in the first state and the second state generated by the imaging condition changing unit. State The λ/4 wavelength plate and the second polarizing plate capture the patterned retardation film to obtain a first image and a second image, and the image combining unit makes the first image and the second image The first phase difference region and the second phase difference region are identical, and the first image and the second image are superimposed to synthesize the first image and the second image; The defect detecting unit detects the defect based on the composite image signal from the image combining unit. The defect inspection apparatus of the patterned retardation film according to claim 10, wherein the image synthesizing unit changes a boundary line between the first phase difference region and the second phase difference region during image synthesis. The relative position between the first image and the second image in the orthogonal direction is such that the first image and the second image are superimposed, and the average luminance of the composite image after the overlap is minimized. The above composite image was produced in relative position. The defect inspection apparatus of the patterned retardation film according to claim 11, wherein the first phase difference region and the second phase difference region in the first image and the second image are provided The brightness of the bright portion reaches the brightness of the dynamic range of the image synthesizing unit and approximately half of the narrow dynamic range of the dynamic range of the imaging device, and then the imaging is performed. The defect inspection apparatus of the patterned retardation film according to claim 12, wherein the image synthesizing unit generates a binarized signal for the first image and the second image to make a binarization The light and dark of the signal is inverted to be an inverted binary signal, and an exclusive OR of the inverted binary signal and another binary signal is obtained. The defect inspection device for a patterned retardation film according to claim 13, comprising a pattern boundary erasing portion, wherein the pattern boundary erasing portion is to be imaged by the image In the binarized composite image generated by the combining unit, the pattern boundary portion based on the boundary between the first phase difference region and the second phase difference region is erased. The defect inspection device for a patterned retardation film according to claim 14, wherein the pattern boundary erasing unit uses a pixel region including a boundary width of the first phase difference region and the second phase difference region as a impurity The news is reduced. The defect inspection apparatus of the patterned retardation film according to any one of claims 10 to 15, wherein the patterned retardation film is in the first phase difference region and the second phase difference region. In the mesh device that is transported in the column direction, the imaging device includes: a first imaging device main body that acquires the first image, and a second imaging device main body that acquires the second image, and the mesh device is transported The above mesh was taken synchronously. A method for detecting a defect of a patterned retardation film, which is a defect detecting method for detecting a defect of a patterned retardation film, wherein the plurality of optical axes are substantially orthogonal to each other in the patterned retardation film The phase difference region and the second phase difference region are alternately arranged side by side, and the first phase difference region and the second phase difference region function as a λ/4 wavelength plate, and the patterning retardation film defect detecting method The adjustment step includes an adjustment step of substantially paralleling one of the first retardation region and the second retardation region with respect to one of the first retardation region and the second retardation region of the first polarizing plate and the second polarizing plate. Adjust within the range so that In a state in which the patterned retardation film is disposed between the first polarizing plate and the second polarizing plate which are disposed orthogonally polarized, the inspection light is irradiated to the patterned phase via the first polarizing plate. In the case of a poor film, the luminances of the normal first phase difference region and the second phase difference region which are imaged by the second polarizing plate are at the same level in the vicinity of the extinction state, and the imaging step is to arrange the patterned retardation film. Between the first polarizing plate and the second polarizing plate adjusted in the direction of the polarization transmission axis, the inspection light is irradiated onto the patterned retardation film via the first polarizing plate, and the second polarizing plate is passed through the second polarizing plate. The patterned retardation film is captured to obtain a luminance image, and a detecting step is performed to detect the defect based on the acquired luminance image. The method for detecting a defect of a patterned retardation film according to claim 17, wherein the patterned retardation film is formed by laminating a phase difference layer on a support including a transparent film, thereby forming the first phase difference. In the adjustment step and the imaging step, the phase difference compensation plate that eliminates the phase difference characteristic of the support is disposed in the region and the second phase difference region, and is disposed in the first polarizing plate and the patterned retardation film. Between or between the second polarizing plate and the patterned retardation film. The method for detecting a defect of a patterned retardation film according to claim 18, wherein the phase difference compensation plate is the same transparent film as the support. The patterned retardation film according to claim 19 of the patent application scope The defect detecting method includes an erasing step of erasing a boundary line in the luminance image corresponding to a boundary between the first phase difference region and the second phase difference region, and the detecting step is based on the boundary line The above-mentioned luminance image is erased to detect the above defect. The method for detecting a defect of a patterned retardation film according to claim 20, comprising: a binarization step of binarizing the captured luminance image with a predetermined threshold value; And each pixel is a bright pixel having a threshold value or more and a dark pixel lower than the threshold value; and a candidate defect extraction step, wherein a region in which the plurality of bright pixels are connected is a candidate defect In the region, the detecting step is to count the number of pixels between the candidate defect region and the adjacent candidate defect regions in the direction in which the boundary lines are arranged, and when the number of pixels counted is equal to or less than a predetermined value, The related candidate defect region is set as the above defect. The method for detecting a defect of a patterned retardation film according to claim 20, wherein the eliminating step includes a binarization step of binarizing the captured luminance image with a predetermined threshold value. And setting each pixel to a bright pixel above the threshold value and a dark pixel lower than the threshold value; and a shrinking step, the number of times corresponding to the width of the boundary line, The binarized luminance image is subjected to a contraction process to erase the boundary line, and the contraction process is a process of shrinking the region of the bright pixel in the direction in which the boundary lines are arranged. The defect detecting method of the patterned retardation film according to claim 22, wherein the eliminating step includes an expansion processing step of expanding the brightness image subjected to the shrinking process to cause the shrinkage The region of the bright pixel remaining in the process is restored to the size before the shrinkage process, and the expansion process is a process of expanding the region of the bright pixel in the same direction as the shrinkage process, and the detecting step is based on the above The above-described luminance image of the expansion process is used to detect the above defect. The method for detecting a defect of a patterned retardation film according to claim 20, wherein the eliminating step includes: a difference image generating step of performing differential processing on each pixel on the luminance image, thereby generating a difference image obtained by substantially eliminating the boundary line, wherein the difference processing is for a pixel on the luminance image, and at least in the arrangement direction of the boundary line is separated from the correlation pixel by only the boundary line interval The pixel of the prime number is used as the counterpart pixel, and the value obtained by subtracting the brightness of the above-mentioned pixel from the brightness of the above-mentioned pixel is set to a new value of the related pixel; and the binarization step is specified The threshold value is used to binarize the difference image, and each pixel is set to be above the threshold value and lower than the upper limit. A dark pixel that describes the threshold. The method of detecting a defect of a patterned retardation film according to claim 24, wherein the difference image generating step performs the differential processing between the opposite pixels in mutually different directions on the luminance image. Thereby, each of the difference images in each direction is generated, and the binarization step binarizes the plurality of difference images by a predetermined threshold value to generate a plurality of binarized images. The detecting step performs defect detection on a plurality of the above-described binarized images. A defect inspection method for patterning a retardation film, which is a defect detection method for detecting a defect of a patterned retardation film, wherein the plurality of optical axes are substantially orthogonal to each other in the patterned retardation film The phase difference region and the second phase difference region are alternately arranged side by side, and the first phase difference region and the second phase difference region function as a λ/4 wavelength plate, and the patterned retardation film defect inspection method The illumination step includes arranging the first polarizing plate and the second polarizing plate so as to interpose the retardation film, and irradiating the illumination light to the pattern via the first polarizing plate by the illumination light source. In the imaging step, the λ/4 wavelength plate is disposed between the patterned retardation film and the second polarizing plate, and is set to the first state and the second state, and is passed through the λ by the imaging device. The fourth wavelength plate and the second polarizing plate are imaged in the first state to obtain a first image, and the second image is captured in the second state, and the first image is obtained. The transmission axis of the first polarizing plate is inclined by 45° with respect to the optical axis of the patterned retardation film, and the transmission axis of the second polarizing plate is increased by +45 degrees with respect to the optical axis of the λ/4 wavelength plate. The second state is such that the transmission axis of the second polarizing plate reaches -45 degrees with respect to the optical axis of the λ/4 wavelength plate; and the image combining step causes the first image and the second image to be The first phase difference region and the second phase difference region are coincident with each other to form a composite image, and a detecting step of detecting the defect based on the composite image. The defect inspection method of the patterned retardation film according to Item 26, wherein the synthetic image step is changed in a direction orthogonal to a boundary line between the first phase difference region and the second phase difference region. The relative position between the first image and the second image is such that the first image and the second image are superimposed, and the relative position of the composite image after the overlap is minimized. The above composite image. The defect inspection method of the patterned retardation film according to claim 27, wherein the photographing step makes the first phase difference region and the second portion in the first image and the second image The brightness of the bright portion of the phase difference region is equal to the brightness of the dynamic range of the image combining unit and approximately half of the dynamic range of the narrow range of the dynamic range of the imaging device, and then the image is captured. The patterned retardation film according to claim 28 of the patent application scope In the defect inspection method, the image synthesizing step generates a binarized signal for the first image and the second image, and inverts the brightness of one binarized signal to obtain an inverted binary signal. Reverse the exclusive OR of the binary signal and another binary signal. The defect inspection method of the patterned retardation film according to claim 29, wherein the image synthesis step is based on the first phase difference region and the second phase difference region in the binarized composite image The boundary portion of the boundary of the border is eliminated. The defect inspection method of the patterned retardation film according to claim 30, wherein the image synthesis step uses a pixel region including a boundary width of the first phase difference region and the second phase difference region as a impurity The reduction process is performed to eliminate the boundary portion of the pattern. The defect inspection method of the patterned retardation film according to any one of claims 26 to 31, wherein the patterned retardation film is different from the first phase difference region and the second phase difference a mesh that is transported in a direction in which the arrangement direction of the regions is orthogonal to the image capturing device includes: a first imaging device main body that acquires the first image; and a second imaging device main body that acquires the second image, and the above The mesh is conveyed in synchronization with the mesh. A method for manufacturing a patterned retardation film, characterized by In the manufacturing step, the patterned retardation film is formed, and the stripe first phase difference region and the second phase difference region in which the plurality of optical axes are substantially orthogonal to each other are alternately arranged side by side in the patterned retardation film. The first phase difference region and the second phase difference region function as a λ/4 wavelength plate, and the obtaining step is to adjust the first polarizing plate and the first polarizing plate in which the patterned retardation film is disposed in the direction of the polarization transmission axis. 2 between the polarizing plates, the inspection light is irradiated onto the patterned retardation film via the first polarizing plate, and the patterned retardation film is imaged through the second polarizing plate to obtain a luminance image; and the detecting step is based on The acquired luminance image detects the defect; and the adjusting step is performed before the obtaining step, and the direction of the polarization transmission axis of the first polarizing plate and the second polarizing plate is made to be the first phase The difference between the difference region and the optical phase of the second phase difference region is substantially parallel, and the patterned retardation film is disposed on the first polarizing plate and In the state between the second polarizing plates, when the inspection light is irradiated onto the patterned retardation film via the first polarizing plate, the normal first phase difference region imaged by the second polarizing plate and Each of the luminances in the second phase difference region reaches the same level in the vicinity of the extinction state. A method of manufacturing a patterned retardation film, comprising: a manufacturing step of fabricating a patterned retardation film at the patterned phase In the differential film, the strip-shaped first phase difference regions and the second phase difference regions in which the plurality of optical axes are orthogonal to each other are alternately arranged side by side, and the illumination step is arranged such that the patterned retardation film is interposed therebetween. a polarizing plate and a second polarizing plate, wherein illumination light is applied to the patterned retardation film via the first polarizing plate by an illumination light source; and a λ/4 wavelength plate is disposed on the patterned retardation film by an imaging step The first state and the second state are set between the first polarizing plate and the second polarizing plate, and the first illuminating plate and the second polarizing plate are imaged by the imaging device in the first state to obtain the first state. The image is captured in the second state to obtain a second image, and the first state is a state in which the transmission axis of the first polarizing plate is inclined by 45° with respect to the optical axis of the patterned retardation film. The transmission axis of the second polarizing plate is +45 degrees with respect to the optical axis of the λ/4 wavelength plate, and the second state is optical for transmitting the transmission axis of the second polarizing plate with respect to the λ/4 wavelength plate. The axis reaches -45 degrees; the image synthesis step makes the above Consistent with an image of the first retardation region in the second image and the second retardation region overlap, a synthetic image; and a detection step to detect the above-mentioned defects based on the synthesized image.