JP2005337857A - Surface unevenness inspection method and inspection apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface unevenness inspection method and an inspection apparatus which can reduce the occurrence of incorrect detection due to brightness irregularity included in an imaged image obtained from an object to be inspected, in optically detecting an uneven defect generated on a luster surface such as a film surface, a steel plate surface, a glass surface, a coating surface, an evaporation surface, or the like, wherein the optical effect emerging in the imaged image due to the uneven state of the defect expected in the object to be inspected is enlarged. <P>SOLUTION: A specular reflection image of a reference surface S1 which has a prescribed periodic bright and dark pattern, in the inspection surface S of the object to be inspected, is imaged by an imaging means. It is constituted so that the positional relationship of respective elements and so forth satisfy a prescribed relationship expression, in detecting the striped unevenness on the inspection surface S, from the distortion in the pattern shape of the obtained specular reflection image. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a surface inspection method and a surface inspection apparatus. In particular, the present invention relates to an inspection method and an inspection apparatus suitable for inspecting uneven defects generated on glossy surfaces such as a plastic film surface, a steel plate surface, a glass surface, a painted surface, and a vapor deposition surface. The preferable form of this invention is related with the manufacturing method of a plastic film.

  In general, when inspecting uneven defects present on an object to be inspected, a method of irradiating the object to be inspected with light and referring to the state of the reflected light to estimate the state of unevenness is often used. ing. Specifically, a surface to be inspected by applying a light cutting method as shown in Patent Document 1 or a surface to be inspected on which bright and dark pattern light is projected as shown in Patent Document 2 Inferring the uneven state from the degree of brightness change of the light and dark edge part in the obtained image, as shown in Patent Document 3, the surface to be inspected on which the light and dark pattern light is projected, The thing which estimates the state of an unevenness | corrugation from the distortion state of the light-and-dark pattern in the obtained image is mentioned.

  However, in the case of the method disclosed in Patent Document 1, basically, it is difficult to perform high-precision inspection on a glossy surface in order to capture the scattered reflected light on the surface to be inspected. Further, in the case of the method disclosed in Patent Document 2, inspection is performed based on unevenness in the luminance of the light source itself, which is a light irradiating unit that emits inspection light, or unevenness in the imaging brightness due to light interference occurring in a laminated portion of a transparent body such as a plastic film roll. Accuracy may be reduced. Further, in the case of the method shown in Patent Document 3, the pattern distortion in the obtained captured image can be sufficiently increased with the configuration of the bright and dark pattern light only from the longitudinal direction of the irregularities expected on the surface to be inspected. Therefore, there is a risk of missing a defect.

Furthermore, in common with all of Patent Document 1, Patent Document 2, and Patent Document 3, the object to be inspected is flat and is arranged so as to always have a fixed positional relationship with respect to the light irradiation means and the imaging means during inspection. It is assumed that That is, for example, in the case of a plastic film roll or the like in which the object to be inspected is wound up during the manufacturing process and gradually becomes thicker, the positional relationship of the surface to be inspected with respect to the light irradiation means and the image pickup means Because of the change, high-precision inspection at all times during product manufacturing cannot be expected.
Japanese Patent Laid-Open No. 11-023485 JP 2002-148195 A JP 2002-202113 A

  In view of the above, an object of the present invention was obtained from an object to be inspected when optically detecting uneven defects generated on a glossy surface such as a film surface, a steel plate surface, a glass surface, a painted surface, and a vapor deposition surface. Provided is a surface irregularity inspection method and inspection apparatus that can reduce the occurrence of false detection due to uneven brightness in a captured image, and that the optical effect that appears in the captured image due to the unevenness state of a defect expected in the inspection target is increased. There is to do. Furthermore, a surface unevenness inspection method and inspection device for inspecting unevenness defects of a test object whose positional relationship between the light irradiation means and the imaging means with respect to the surface to be inspected changes with time is provided. There is.

  In order to achieve the above object, the surface irregularity inspection method and inspection apparatus of the present invention have the following configurations.

That is, in the inspection method of the present invention, the regular reflection image on the surface S to be inspected of the object to be inspected of the reference surface Sl whose surface has the brightness of a predetermined periodic bright and dark pattern is imaged by the imaging unit, and the obtained regular reflection is obtained. A surface unevenness detecting method for detecting streak unevenness on the surface S to be inspected from a pattern shape distortion of an image, wherein an angle formed between a normal Z of the surface S to be inspected and a line-of-sight direction I of the imaging means is φ When the direction of the reference vector q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl including the normal line Z and the line-of-sight direction I is the same direction as the periodic change direction Hv of the periodic pattern An angle ψ formed with respect to the movement vector i satisfies the following equation (1), a direction Hh perpendicular to the direction Hv of the periodic change of the periodic pattern, an infinite plane including the reference surface Sl, and the ratio inspection surface S So that the angle θ formed by the line of intersection with It is characterized in placing the surface and the imaging means.

Here, Ψmin can be a value between 0.5 and 4.0 ° for the slope parameter α, and can be expressed by the following Equation 3 when the slope parameter β can take a value between -25 and 25 °. The defined minimum value of Ψ0, Ψmax is the maximum value of Ψ0.

Further, the inspection method of the present invention uses a method in which the magnitude | i | of the movement vector i satisfies the following expression 4 and the periodic bright / dark pattern has a bright part with a width pa and a dark part with a width pb: It is also preferable to install the reference surface and the imaging means so that the representative width pc satisfies the following expression (5).

Here, | i | min is assumed to be a value when the slope parameter α can take a value between 0.5 and 4.0 °, and when the slope parameter β can take a value between -25 and 25 °. The minimum value of | i | 0 defined by Equation 6 can be taken, and | i | max is the maximum value of | i | 0. The larger one of the widths pa and pb.

In the inspection method of the present invention, the regular reflection image of the reference surface Sl having a predetermined periodic brightness / darkness pattern brightness on the inspection surface S of the object to be inspected is imaged by the imaging means, and the obtained regular reflection is obtained. A surface unevenness detecting method for detecting streak unevenness on the surface S to be inspected from a pattern shape distortion of an image, wherein an angle formed by a normal Z of the surface S to be inspected and a line-of-sight direction I of the imaging means is φ, The inclination of the assumed streak unevenness with respect to the surface S to be inspected α, the assumed streak unevenness with respect to the intersection line CL1 of the plane S0 including the normal Z and the line-of-sight direction I and the surface S to be inspected. The direction of the reference vector q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl is moved in the same direction as the direction Hv of the periodic change of the periodic pattern when the longitudinal inclination of β is β An angle ψ formed with respect to the vector i satisfies the following formula 7, and the period The reference surface so that an angle θ formed by a direction Hh perpendicular to the direction Hv of the periodic change of the pattern and an intersection line between the infinite plane including the reference surface Sl and the specific inspection surface S satisfies the following equation (8): And the imaging means.

Here, Ψmin can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and takes a value between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value Ψmax that can be taken by the value of Ψ0 defined by the following equation 9 when it can be obtained is the maximum value that Ψ0 can take.

The inspection method of the present invention uses a method in which the magnitude | i | of the movement vector i satisfies the following expression 10 and the periodic bright / dark pattern has a bright part with a width pa and a dark part with a width pb: It is also preferable to install the reference surface and the imaging means so that the representative width pc satisfies the following expression (11).

Here, | i | min can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value of | i | 0 defined by the following equation 12 when the value can be taken, | i | max is the maximum value of | i | 0. The larger one of the widths pa and pb.

  Further, the inspection method of the present invention uses a stripe pattern as the periodic light / dark pattern, and uses an image pickup means for picking up a two-dimensional image based on reflected light from the surface S to be inspected. It is also preferable that the imaging resolution is set so as to include five or more pixels for each of the apparent bright part width and dark part width of the stripe pattern on the obtained two-dimensional image.

  In the inspection method of the present invention, it is also preferable to use a stripe pattern as the periodic light and dark pattern, and subject the obtained two-dimensional image to a differentiation process in the stripe longitudinal direction.

  In the inspection method of the present invention, a stripe pattern is used as the periodic light / dark pattern, and a one-dimensional image is picked up based on the reflected light from the surface S to be inspected as the imaging means. The imaging resolution of the two-dimensional image obtained by arranging the obtained one-dimensional images is five for each of the apparent bright part width and dark part width of the stripe pattern on the obtained two-dimensional image. It is also preferable to set so as to include the above pixels.

  In the inspection method of the present invention, it is also preferable to detect the concavo-convex portion signal from the light / dark period disturbance of the stripe pattern of the obtained image.

  In the inspection method of the present invention, it is also preferable to detect the concavo-convex portion signal by performing a comparison operation between the stripe pattern of the obtained image and a predetermined master image.

  The inspection method of the present invention is an inspection method for detecting irregularities on the surface to be inspected from the pattern shape distortion of an image obtained by imaging the light and dark pattern projected on the surface to be inspected S. The light and darkness of each part in the light emitting surface is projected onto the surface S to be inspected by the light / dark pattern light irradiating means that can be arbitrarily set.

  Further, the inspection method of the present invention is an inspection method for detecting irregularities on the surface S to be inspected from the pattern shape distortion of an image obtained by imaging the bright and dark pattern projected on the surface S to be inspected, and the inspection object is The roll is rotated, and the distance between the surface S to be inspected and the rotation center axis of the roll changes with time, and / or the rotation center axis of the roll moves. And blurring of a captured image that accompanies a change in the distance of the rotation center axis of the roll and / or movement of the rotation center axis of the roll.

  In the inspection method of the present invention, it is also preferable to set the exposure time of the imaging means so that the rotational movement of the surface S to be inspected during exposure is 0.25 to 2.5 mm.

Further, the inspection apparatus of the present invention was obtained, comprising a reference surface Sl whose surface has a luminance of a predetermined periodic bright and dark pattern, and an imaging means for capturing a regular reflection image of the inspection object on the inspection surface S. A surface unevenness detecting device for detecting streak unevenness on the surface to be inspected from the pattern shape distortion of the regular reflection image, wherein the angle formed between the normal line Z of the surface to be inspected and the line-of-sight direction I of the imaging means Is the direction of the reference pattern q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl including the normal line Z and the line-of-sight direction I is the direction Hv of the periodic change of the periodic pattern. The angle ψ formed with respect to the movement vector i in the same direction satisfies the following expression 13, and the ratio Hh perpendicular to the direction Hv of the periodic change of the periodic pattern and the infinite plane including the reference surface Sl and the ratio Before the angle θ formed by the line of intersection with the inspection surface S satisfies the following equation (14) It is characterized in that the reference surface and the imaging means are installed.

Here, Ψmin can be a value between 0.5 and 4.0 ° for the slope parameter α, and the following equation 15 when the slope parameter β can take a value between -25 and 25 °. The defined minimum value of Ψ0, Ψmax is the maximum value of Ψ0.

Further, in the inspection apparatus of the present invention, the magnitude | i | of the movement vector i satisfies the following Expression 16, and the periodic bright / dark pattern has a bright part having a width pa and a dark part having a width pb. It is also preferable that the reference surface and the imaging unit are installed so that the representative width pc satisfies the following expression (17).

Here, | i | min is assumed to be a value when the slope parameter α can take a value between 0.5 and 4.0 °, and when the slope parameter β can take a value between -25 and 25 °. The minimum value of | i | 0 defined by Equation 18 can be taken, and | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

Further, the inspection apparatus of the present invention was obtained, comprising a reference surface Sl whose surface has a luminance of a predetermined periodic bright and dark pattern, and an imaging means for capturing a regular reflection image of the inspection object on the inspection surface S. A surface unevenness detecting device for detecting streak unevenness on the surface to be inspected from the pattern shape distortion of the regular reflection image, wherein the angle formed between the normal line Z of the surface to be inspected and the line-of-sight direction I of the imaging means Φ, inclination of the assumed streaky irregularity with respect to the surface to be inspected α, α, the normal line Z and the plane S0 including the line-of-sight direction I and the intersection line CL1 of the surface to be inspected S1 When the inclination in the longitudinal direction of the streaks is β, the direction of the reference vector q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl is the same as the direction Hv of the periodic change of the periodic pattern. The angle ψ formed with respect to the moving vector i in the direction satisfies the following equation 19 and The angle θ formed by the direction Hh perpendicular to the periodic change direction Hv of the periodic pattern and the line of intersection between the infinite plane including the reference surface Sl and the specific inspection surface S satisfies the following Expression 20. A reference surface and the imaging means are installed.

Here, ψmin can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and takes a value between the assumed minimum value αmin and the maximum value αmax of the slope β. The minimum value Ψmax that can be taken by the value of Ψ0 defined by the following equation 21 when it is assumed to be Ψ0 is the maximum value that Ψ0 can take.

Further, in the inspection apparatus of the present invention, the magnitude | i | of the movement vector i satisfies the following expression 22, and the periodic bright / dark pattern has a bright part having a width pa and a dark part having a width pb. It is also preferable that the reference surface and the imaging unit are installed so that the representative width pc satisfies the following equation (23).

Here, | i | min can take a value between the assumed minimum value αmin and the maximum value αmax of the slope parameter α, and the assumed minimum value βmin and the maximum value βmax of the slope parameter β The minimum value of | i | 0 defined by the following equation 24 when the value between can be taken, | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

  Further, in the inspection apparatus of the present invention, the periodic light / dark pattern is a stripe pattern, the imaging means is an area sensor camera in which photoelectric conversion elements are two-dimensionally arranged, and the imaging resolution is the obtained two-dimensional image. It is also preferable to set so as to include five or more pixels for each of the apparent bright part width and dark part width of the stripe pattern above.

  The inspection apparatus of the present invention is a line sensor camera in which the periodic bright and dark pattern is a stripe pattern, the imaging means is one-dimensionally arranged with photoelectric conversion elements, and the imaging resolution is the obtained two-dimensional image. It is also preferable that the stripe-type pattern is set so as to include five or more pixels for each of the apparent bright part width and dark part width.

  Further, the inspection apparatus of the present invention is characterized in that the illumination means and the light / dark pattern light creating means are composed of light / dark pattern light irradiation means capable of arbitrarily setting the light and darkness of each part in the light emitting surface. is there.

  The inspection apparatus of the present invention is an inspection apparatus provided with illumination means, light / dark pattern light creation means, imaging means, and image processing means, and the object to be inspected is in a roll shape and rotated around the roll center axis. The movement is performed, and the distance between the surface S to be inspected and the rotation center axis of the roll changes with time, and / or the rotation center axis of the roll moves from the position before the start of rotation. In addition, a mechanism for compensating for a change in the distance between the surface S to be inspected and the rotation center axis of the roll and / or blurring of the captured image accompanying the movement of the rotation center axis of the roll is provided.

  In the inspection apparatus of the present invention, it is also preferable that the exposure time of the imaging means is set to 1/60 to 1/3000 sec.

  Moreover, the manufacturing method of the present invention includes a step of inspecting a transparent plastic film or a plastic film having a metal thin film deposited thereon using any one of the above-described inspection methods or any of the inspection apparatuses. It is.

  According to the present invention, as will be described below, it is obtained from an object to be inspected when optically detecting uneven defects generated on a glossy surface such as a film surface, a steel plate surface, a glass surface, a painted surface, and a vapor deposition surface. A method and apparatus for inspecting surface irregularities that can reduce false detections due to uneven brightness in captured images and increase the optical effect that appears in the captured images due to the irregularities of defects expected in the inspection object. Can do. Furthermore, it is possible to obtain a surface unevenness inspection method and inspection apparatus for inspecting unevenness defects of an inspection object whose positional relationship between the light irradiation means and the imaging means with respect to the inspection object surface changes with time. Can do.

  Hereinafter, examples of the best embodiment of the present invention will be described in the case of detecting folding wrinkles generated in a transparent plastic moving in one direction on a plane and winding generated on the surface of a transparent plastic film roll wound up in a roll shape. A case where wrinkles are detected will be described as an example with reference to the drawings.

  FIG. 1 is a schematic diagram showing the basic configuration of the present invention. 2 (a) and 2 (b) are taken images of a bright and dark pattern. Of these, (a) shows a case in which the concave / convex defect 31 does not exist on the inspection target surface S, and (b) shows an inspection target surface S. This is an image when the concave-convex defect 31 exists.

  The bright / dark pattern light irradiating means 10 and the imaging means 20 are arranged in a regular reflection relationship in which the angle with respect to the normal Z of the surface 30 to be inspected is φ. Here, the light / dark pattern light irradiating means 10 is provided with light / dark pattern light creating means 12 in which light transmitting parts 12a and light shielding parts 12b are regularly arranged in a stripe shape in the light emitting part of the illuminating means 11 having a two-dimensional spread. It is realized by.

  Here, the stripe type means that the boundary between the bright part and the dark part in the light / dark pattern is straight and parallel, the bright part and the dark part are repeatedly arranged in order, and the width of the bright part (hereinafter sometimes referred to as the bright pitch pa), The width of the dark portion (hereinafter sometimes referred to as dark pitch pb) is substantially the same for each repeating unit, and the sum of one pa and one pb is the stripe pitch p (hereinafter simply referred to as pitch p). The direction of the stripe brightness change is defined as the direction Hv, and the direction of the light-dark boundary line orthogonal to the direction is defined as the direction Hh. For convenience of explanation, in this embodiment, the surface of the light / dark pattern light irradiation means 10 is defined as a reference surface Sl. The imaging means 20 is composed of a photoelectric conversion means 21 typified by a general-purpose CCD camera and a condensing means 22 such as a lens having a condensing function, and the focus is not the surface 30 to be inspected but the reference surface. Fit for Sl. Therefore, when the concave / convex defect 31 does not exist on the surface 30 to be inspected, the imaging means 20 images a light / dark pattern without distortion, and the image is as shown in FIG. On the other hand, when the concave / convex defect 31 is present on a part of the surface 30 to be inspected, the imaging means 20 Since a partially distorted light / dark pattern is imaged, the image has a distorted portion 32 as shown in FIG. In other words, it is possible to detect the concavo-convex defect 31 on the surface 30 to be inspected by detecting the distortion of the light and dark pattern in the captured image.

  Further, according to the knowledge of the present inventors, the configuration of the optical system such as the illumination unit 11, the imaging unit 20, the positional relationship of the surface 30 to be inspected, the shape of the light / dark pattern provided in the light / dark pattern light creating unit 10, and There is a close relationship between the geometrical state of the concave / convex defect 31 and the degree of distortion 32 of the light / dark pattern appearing in the obtained image. This means that the quality of the product as a transparent plastic film can be judged by estimating the degree of the irregularity defect 31 from the degree of the distortion 32 of the light and dark pattern appearing in the obtained image, and on the surface 30 to be inspected. Based on the geometrical characteristics of the concavo-convex defect 31 that can be expected, the optical effect that appears in the captured image due to the concavo-convex defect 31 is maximized, in other words, it is the highest for the state of the concavo-convex defect 31 that is the detection target. This means that the optical system can be configured to achieve sensitivity.

  The relationship between the configuration state of the optical system, the geometric state of the concavo-convex defect 31, and the distortion 32 of the light / dark pattern appearing in the obtained image will be described with reference to FIG. FIGS. 3A and 3B are analysis diagrams showing the geometric state of the optical system portion according to the present invention, FIG. 3A is a diagram showing the entire optical system, and FIG. It is the figure which looked at the light emission surface from the front. Here, the installation angle of the light / dark pattern light irradiation means 10 and the imaging means 20 arranged in a regular reflection relationship with respect to the surface to be inspected φ is φ, and the uneven defect 31 expected on the surface 30 to be inspected is the surface 30 to be inspected. And α is the angle of the rotation direction of the concavo-convex defect with the vertical axis Z of the surface 30 to be inspected (normal line of the surface 30 to be inspected) as the central axis. In addition, an angle formed by an intersection line CL3 between an infinite plane including the light-dark pattern light creating means 12 surface configured in a stripe shape and the surface S to be inspected on the inspection target object 30 with respect to the direction Hh is θ, Let p be the pitch of the light-dark pattern repetition. For easy understanding of the geometric explanation, only the geometric elements are extracted from FIG. 3 and necessary elements are added and shown in FIG. As shown in FIG. 11, a plane including the vertical axis Z and the line-of-sight direction I of the imaging means 20 is defined as a plane S0, and the angle β is based on an intersection line CL1 between the plane S0 and the surface S to be inspected (0 °). Counterclockwise direction is defined as positive. Further, an intersection line CL2 between the plane S0 and the reference surface Sl is defined.

  Returning to FIG. First, let us consider a case where the uneven defect 31 does not exist on the surface 30 to be inspected. Then, one reflected light Lr incident on the imaging means 20 is the reflected light Pr1 incident on the reflection point Pr1 on the surface 30 to be inspected and reflected by the reflection point Pr1. The incident light Li1 is a regular reflection relationship with the reflected light Lr on the surface to be inspected 30, and is light emitted from the light emitting point Pi1 on the reference surface Sl. Therefore, the light emitting point Pi1 is in the position of regular reflection that is separated from the imaging means 20 by the installation angle φ from the vertical axis Z at the reflection point Pr1 when viewed from the reflection point Pr1 on the surface 30 to be inspected due to the arrangement configuration of the apparatus.

  Similarly, various reflected lights having substantially the same direction as the reflected light Lr and different in two-dimensional (planar) positions are incident on the imaging means 20, but all of these are reflected in the same manner as described above. The relationship of light emission is established. That is, the surface 30 to be inspected acts as if it were a mirror surface, and a light / dark pattern image without distortion as shown in FIG.

  On the other hand, consider a case where the uneven defect 31 exists on the surface 30 to be inspected. Then, one reflected light Lr incident on the imaging means 20 is the reflected light Pr2 incident on the reflection point Pr2 on the inclined surface 31a of the concave-convex defect 31 at the reflection point Pr2. This incident light Li2 is a light that is emitted from the light emitting point Pi2 on the reference surface S1 and has a regular reflection relationship with the reflected light Lr not on the surface 30 to be inspected but on the inclined surface 31a. Therefore, this light emitting point Pi2 is not in the position of regular reflection that is separated from the imaging means 20 by the installation angle φ from the vertical axis Z at the reflection point Pr1 as seen from the reflection point Pr1 of the surface to be inspected. .

  Similarly, various reflected light beams having substantially the same direction as the reflected light beam Lr and having different two-dimensional (planar) positions are incident on the image pickup means 20, but the reflection point is on the slope of the concavo-convex defect 31. As for the reflected light, the same reflection / emission relationship as described above is applied, and the reflected light on the surface 30 to be inspected where the reflection point 31 does not exist is reflected and emitted when the uneven defect 31 is not present. A relationship is established. That is, the surface 30 to be inspected and the concavo-convex defect 31 act as if they are mirror surfaces having concavo-convex portions, and a light / dark pattern image having a distortion 32 as shown in FIG.

  Here, referring to FIG. 3B, when a certain reflected light Lr captured by the imaging means 20 is considered, the imaging means 20 has different reference surfaces depending on whether or not the uneven defect 31 is present on the surface 30 to be inspected. The light emitting points Pi1 and Pi2 on the Sl are imaged. It can be said that this phenomenon is geometrically caused by a movement vector i from the light emitting point Pi1 to the light emitting point Pi2.

  The movement vector i is considered for the assumed uneven defect 31, and the relationship between the movement vector i and the light / dark pattern of the reference surface is one factor that determines the sensitivity of the entire system. Hereinafter, in order to examine this, a reference vector q which is on the reference plane Sl and is perpendicular to the plane S0 including the normal Z and the line-of-sight direction I and the intersection line CL2 of the reference surface Sl will be considered.

  If there is no concavo-convex defect 31 on the surface S to be inspected, the light emitting point Pi1 on the reference surface Sl that the camera will capture will move to the light emitting point Pi2 on the reference surface Sl captured by the camera because the concavo-convex defect 31 exists Therefore, in order to increase the optical effect of the concavo-convex defect 31 on the light and dark stripe image captured by the camera in order to facilitate the detection of the presence of the concavo-convex defect 31, in order to increase the distortion of the light and dark stripe, The reference surface SI is preferably configured so that the light and dark borders of the stripes are arranged perpendicular to the angle ψ formed by i with respect to the reference vector q. That is, it is preferable that the light and dark stripes are configured so that the light and dark change direction Hv is the same as the angle ψ. Also, with respect to the magnitude | i | of the movement vector i corresponding to the distance between the light emitting point Pi1 and the light emitting point Pi2, the light and dark pitch p is the larger of the bright part width pa and the dark part width pb constituting p. (Representative width pc) is the same or less (the movement trajectory from the light emission point Pi1 to the light emission point Pi2 always includes at least one light / dark boundary part), and the finest possible as far as light and dark can be distinguished by image resolution Is preferred.

Here, the installation angle φ of the light / dark pattern light irradiation means 10 and the imaging means 20 defined above, the slope angle α formed by the concave / convex defect 31 with the surface S to be inspected, and the vertical axis Z of the surface S to be inspected as the central axis The direction ψ and the magnitude | i | of the movement vector i vary depending on the state of the rotation direction angle β of the uneven defect and the distance l between the surface S to be inspected and the illumination surface Sl. If the range of the shape condition (α, β) of the concavo-convex defect 31 expected to occur on the surface to be inspected with respect to the range of (φ, l), the range of ψ, | i | can also be determined. In the present invention, the following formulas 25 and 26 hold.

  Here, ψmin can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and takes a value between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value Ψmax that can be taken by the value of Ψ0 defined by the following equation 27 when it is assumed to be Ψ0 is the maximum value that Ψ0 can take.

Similarly, | i | min can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value | i | max defined by the following equation 28 and | i | max is the maximum possible value | i | 0.

  For the above, it is assumed that the installation angle φ of the light / dark pattern light irradiating means 10 and the imaging means 20 is 5 °, the distance l between the surface S to be inspected and the illumination surface Sl is 2000 mm, and the slope on which the irregularity defect 31 forms the surface 30 to be inspected. The angle α is a typical case of 0.4 to 4 °, and the rotation direction angle β of the concavo-convex defect around the vertical axis Z of the surface 30 to be inspected is also a typical case of the range of -25 to 25 °. , The direction ψ of the movement vector i takes a range as shown in FIG. 9, and the magnitude | i | takes a range as shown in FIG. That is, the range that Equation 25 can take is 0.035 ° <ψ <25.4 °, and the range of values that Equation 26 can take is 27.82 mm <| i | <280.2 mm. However, the range of values obtained here is one example under a certain assumed condition, and the range of values that can be taken by the expressions 25 and 26 of the present invention is not limited to this. Based on the relationship as described above, the slope angle α formed with the surface to be inspected 30 of the uneven defect 31 assumed and the rotation direction angle β of the uneven defect with the vertical axis Z of the surface 30 to be inspected as the central axis If an input is provided, an optimal inspection can be easily performed by providing a means for outputting an appropriate installation angle of the light / dark pattern light irradiation means 10 and an arrangement of the light / dark pattern on the reference plane SI.

In addition, if the direction ψ and the magnitude | i | of the movement vector i are determined according to the geometrical law described above, the light and dark stripe image captured by the camera to facilitate detection of the presence of the concavo-convex defect 31 is determined. A preferable arrangement of the light and dark stripes on the reference surface Sl for increasing the optical influence of the uneven defect 31, that is, the distortion of the light and dark stripes, is determined. The direction ψ of the movement vector i and the inclination angle θ of the light and dark stripes with respect to the intersection line CL3 on the surface S to be inspected (the light and dark boundary direction Hh perpendicular to the direction Hv of the periodic change of the periodic pattern) It is assumed that an angle θ) formed by an intersection line CL between an infinite plane including the reference surface Sl and the surface to be inspected is arranged vertically. That is, the following relational expression 29 holds.

Also, with respect to the magnitude | i | of the movement vector i, the bright and dark pitch p is equal to or smaller than the larger one (representative width pc) of the bright portion width pa and the dark portion width pb constituting the p (light emitting point). It is preferable that the movement trajectory from Pi1 to the light emission point Pi2 always includes at least one light / dark boundary portion), that is, it is preferably set to a value represented by Expression 30.

  Here, pc in Expression 30 is the larger one of the width pa of the bright portion and the width pb of the dark portion constituting the light / dark pitch p (that is, pc = pb if pa <pb), and pa and pb are substantially equal. As long as they are the same,

Further, the light and dark pitch p is preferably as fine as possible as long as light and dark can be distinguished by the resolution of the two-dimensional image obtained by the imaging means, that is, satisfies the value represented by Expression 31 and is represented by Expression 32. It is preferable to set so as to satisfy a certain value. Here, b in Expression 31 and Expression 32 is the horizontal resolution of the camera.

  As described above, by determining the shape conditions (α, β) of the concavo-convex defect 31 that is expected to occur on the inspection target surface S to be detected, the direction ψ and the magnitude of the movement vector i from Equation 25 and Equation 26. It is assumed that | i | is determined, and further, this direction ψ and size | i | are used to construct a light and dark stripe that satisfies Expression 29, Expression 30, and Expression 31 and reflect it in the light / dark pattern light creating means 12 The effect of the uneven defect 31 on the captured image of the camera can be maximized, that is, the highest sensitivity can be obtained in the inspection of the uneven defect 31.

  Further, the smaller the slope angle α and the rotation direction angle β of the concavo-convex defect 31, the smaller the direction ψ of the movement vector i (= the angle formed by the reference vector q and the movement vector i on the reference surface Sl) and the magnitude | i | As the slope angle α and the rotation direction angle β increase, the direction ψ and the magnitude | i | of the movement vector tend to increase. Therefore, among the concavo-convex defects to be detected that are expected on the surface S to be inspected, with respect to the concavo-convex defects whose slope angle α and rotation direction angle β are relatively small and the direction ψ of the movement vector i and the size | i | It is preferable to construct a light and dark stripe so as to obtain the highest sensitivity.

  As described above, each value of the shape condition (α, β) of the concavo-convex defect to be actually measured is referred to as an inclination parameter. It is preferable that the inclination parameters α and β constitute an optical system based on the above relational expressions using the assumed inclination value range as it is. On the other hand, even if there is no value that can be assumed, if the tilt parameter is calculated backward from the configuration of the optical system, it is possible to detect the concavo-convex defect close to that value with high accuracy.

  More specifically, as shown in FIG. 9 and FIG. 10, the magnitude of the slope angle α has a relatively strong influence on the magnitude | i | of the movement vector i, and the magnitude of the rotation direction angle β is the direction of the movement vector i. It is proved geometrically that it has a relatively strong influence on ψ. In addition, with respect to the stripe inclination angle θ and the stripe pitch p, which are important elements of the configuration of the bright and dark stripes, the pitch p is determined by detecting the width pa of the bright portion constituting the p and the dark portion of the dark portion. If the larger one of the widths pb (representative width pc) is less than or equal to the size of the movement vector i | i |, each of the bright pitch pa and the dark pitch pb is at least 5 times the camera resolution b. The inspection sensitivity increases as it is set to be as small as possible while satisfying the above conditions, whereas the inclination angle θ at which the inspection sensitivity is highest is a value uniquely determined by the direction ψ of the movement vector i. That is, preferentially, the stripe angle θ is set so that the detection sensitivity is the highest for the concave / convex defect 31 to be detected among those having the smallest rotation direction angle β, and the pitch p is the bright part of p The larger width (representative width pc) of the width pa and the dark portion width pb is equal to or smaller than the size | i | of the movement vector i, and each of the bright pitch pa and the dark pitch pb is 5 of the camera resolution b. It is preferable to set it to the smallest value that satisfies the condition of doubling or more. By constructing the light and dark stripe in this way, when a concavo-convex defect 31 ′ having a larger slope angle α and rotational direction angle β than the concavo-convex defect 31 having the smallest slope angle α and rotational direction angle β to be detected occurs. Although the optimum condition for the light and dark stripes is not satisfied, the direction ψ and the magnitude | i | of the movement vector i become large, so that sufficient sensitivity can be obtained to detect this, and there is no problem in actual operation.

  Next, a signal processing method for determining the presence / absence and degree of the concavo-convex defect 31 from the stripe image obtained by the imaging means 20 will be described with reference to FIGS. 4A is an image obtained by distorting the stripes because the unevenness defect 31 exists on the surface S to be inspected, and FIG. 4B is an image obtained by distorting the stripes 31 because the unevenness defect 31 does not exist on the surface S to be inspected. (C), (d) and (e) are images obtained by performing a predetermined operation 1 on (a) and (b), respectively, under different conditions. f) and (g) are images obtained by further applying a predetermined calculation 2 to (c) and (e), respectively.

  First, calculation 1 is performed on FIG. The calculation 1 is a technique generally called differentiation processing, calculates the luminance difference between the target pixel and its neighboring comparison pixels, converts the absolute value of the luminance difference into new luminance, and plots it at the position of the target pixel. From the viewpoint of macroscopic view of (a), this is a process having a function of converting a part with a sharp change in brightness, for example, an edge part of a stripe, in particular, into a new high brightness part. In using this processing as the calculation 1, the present invention is characterized in that the pixel of interest and the comparison pixel are arranged in the same direction as the longitudinal direction of the stripe. That is, for the image (a) obtained in the preferred form of the present invention, for example, when calculation 1 is performed by arranging the target pixel and the comparison pixel in the horizontal direction of the image, as shown in FIG. In addition, an image in which bright and dark edge portions are emphasized is obtained regardless of the non-distorted portion and the distorted portion of the stripe, and it becomes difficult to specify only the distorted portion from (d). On the other hand, according to the preferred embodiment of the present invention, by performing the operation 1 with the target pixel and the comparison pixel arranged in the same direction (oblique) as the longitudinal direction of the stripe, the non-distortion as shown in (c) Since the brightness difference between the target pixel and the comparison pixel is small in the part, the brightness is converted to low, and in the distortion part, the brightness difference between the target pixel and the comparison pixel is large, resulting in a high brightness converted image. The undistorted part and the distorted part in the image can be characterized. In addition, when there is no distorted portion in the stripe image as in the image (b) obtained when the concave / convex defect 31 does not exist on the surface S to be inspected, it is compared with the target pixel in the same direction (oblique) as the longitudinal direction of the stripe. By performing the calculation 1 by arranging the pixels, a uniformly low-brightness processed image as shown in (e) can be obtained.

  Next, calculation 2 is performed on images (c) and (e) obtained by applying calculation 1 of the present invention to (a) and (b). Arithmetic 2 is a process generally called dilation, which designates a pixel of interest luminance and converts surrounding pixels to the same luminance as the attention pixel or designated luminance, and exists in the neighborhood However, this is a process having a function of connecting a target pixel that is not in contact with each other or a target region in which a plurality of target pixels are gathered to convert it into a larger region. First, when the irregularity defect 31 exists on the surface S to be inspected and the calculation 2 is performed on the image (c) obtained by the calculation 1 of the preferred embodiment of the present invention, the high-intensity portions are connected to form one region. An image (f) having a region converted into is obtained. Further, when the calculation 2 is performed on the image (b) obtained by the calculation 1 of the present invention when the unevenness defect 31 does not exist on the surface S to be inspected, there is no high brightness region to be connected, so that it is uniform. An image (g) converted into low luminance is obtained. Finally, the presence / absence of the concave / convex defect 31 can be determined by determining the presence / absence of a region of interest, for example, a high luminance region, for the processed images (f), (g), and the like.

  Further, as described above, since the direction ψ and the magnitude | i | of the movement vector i change depending on the slope angle α and the rotation direction angle β of the concavo-convex defect 31, the degree of distortion in (a) also changes depending on the degree of concavo-convex. Will be. That is, in the larger and more severe (small and fine) concavo-convex defect having the slope angle α and the rotation direction angle β, the distortion of the stripe becomes larger (smaller), and the number of pixels converted to high luminance by the calculation 1 is larger ( A reduced image (c) is obtained. Then, as a matter of course, by performing the calculation 2 on (c), when the degree of the concavo-convex defect 31 is more severe, when the concatenated and transformed high-luminance attention area image (f) is fine The image (f) in which the connected high-luminance attention area is small or cannot be connected and converted into a plurality of small areas is obtained, and the unevenness defect is obtained by analyzing the feature quantity such as area and number. The degree of 31 can be relatively evaluated.

  Note that the signal processing method described above is the most basic part for easily understanding the principle. In practice, in addition to this method, the signal processing accuracy is generally improved. It is also preferable to perform a noise removal (for example, averaging processing, median processing, smoothing processing, etc.) method or binarization method used at an appropriate stage.

  The above description has been made on the assumption that a device having a two-dimensional field of view represented by a CCD area sensor camera is used as the photoelectric conversion means 21 of the imaging means 20. However, according to another preferred embodiment of the inspection method and inspection apparatus of the present invention, a device having a one-dimensional field of view represented by a CCD line sensor camera can be used as the photoelectric conversion means 21. In general, a CCD line sensor camera has a larger number of pixels in the horizontal direction of the image than a CCD area camera, and when inspecting an object to be inspected wide in the width direction, the number of imaging means 20 can be reduced, and / or Alternatively, there is an advantage that the inspection resolution can be improved by reducing the resolution of imaging.

  When the photoelectric conversion means 21 in FIG. 3A is changed from a device having a two-dimensional field of view to a device having a one-dimensional field of view, FIGS. 2C and 2D are used for an image obtained by the imaging means 20. I will explain. Here, FIG. 2C shows an image obtained by the photoelectric conversion means 21 having a one-dimensional field of view when the unevenness defect 31 does not exist on the surface S to be inspected, and FIG. 2D shows the case where the unevenness defect 31 exists. It is an obtained image. First, when the uneven defect 31 does not exist on the surface S to be inspected, a two-dimensional image (c) obtained by arranging the one-dimensional visual field 33 of FIG. ) Will be obtained. In the image (c), bright portions and dark portions extending in the vertical direction of the image are periodically arranged at a predetermined pitch in the horizontal direction. On the other hand, when the uneven defect 31 exists on the surface S to be inspected in the same manner, a two-dimensional image obtained by arranging the one-dimensional field of view 34 of FIG. d) will be obtained. In the image (d), a bright part and a dark part extending in the vertical direction of the image are picked up, but the horizontal pitch disturbs a predetermined period due to an axial change (axial misalignment) of the incident light Li caused by the uneven defect 31. doing. Here, the reason why a change appears in the image obtained by the imaging means 20 depending on the presence or absence of the concave / convex defect 31 is the same as in the case where the device having the two-dimensional field of view already described is used. However, since there is a difference in the principle of constructing a two-dimensional image between a device having a two-dimensional field of view and a device having a one-dimensional field of view, the obtained images include (a), (c), (b), and (d). The difference appears.

  Next, a first signal processing method for determining the presence / absence and degree of the concavo-convex defect 31 from the images (c) and (d) obtained by the imaging means 20 having a one-dimensional field of view will be described with reference to FIGS. This will be described using c ′), (d), and (d ′). Here, FIG. 2 (c ′) shows luminance waveforms 43 and 44 obtained by integrating the luminance information of the image (c) and (d ′) in the vertical direction.

  First, when the uneven defect 31 does not exist on the surface S to be inspected, the luminance waveform 43 obtained from the image (c) has the bright area Pitch Pb and the dark area Pd Pd, and both are in the horizontal direction of the image. , They are regularly and periodically arranged without changing the pitch. On the other hand, when the irregularity defect 31 is present on the surface S to be inspected, the luminance waveform 44 obtained from the image (d) is not affected by the irregularity defect 31, and the irregularity defect 31 is not present. Similarly, a periodic state of Pb and Pd can be obtained. In the region where the axial deviation of the incident light Li occurs due to the influence of the concavo-convex defect 31, for example, as shown in FIG. Like the width Pm1 to Pm7 of the dark area, the Pb and Pd are arranged in the horizontal direction with different widths. That is, if the horizontal arrangement period of the bright part width Pb and the dark part width Pd is analyzed, in other words, if the presence / absence of the bright part / dark part width Pmx different from Pb and Pd is detected, the presence / absence of the uneven defect 31 on the surface S to be inspected Can be determined.

  In addition, as described above, the direction ψ and the magnitude | i | of the movement vector i change depending on the slope angle α and the rotation direction angle β of the concavo-convex defect 31, so the degree of periodic fluctuation of the luminance waveform 44 depends on the degree of concavo-convexity. Will change. That is, in the larger and more severe (small and fine) irregularity defect of the slope angle α and the rotation direction angle β, the region (distance between Pm1 to Pm7) where the periodic fluctuation occurs in the luminance waveform 44 becomes large (small), and The number of periods Pmx different from Pb and Pd will increase (decrease). By analyzing these information, the degree of the concavo-convex defect 31 can be relatively evaluated.

  In the above description, the luminance information of the images (c) and (d) is integrated in the vertical direction to obtain luminance waveforms 43 and 44. The luminance information is averaged in the vertical direction, or (c) and (d). Even if the luminance waveforms 43 and 44 are obtained from the luminance information of one line in the horizontal direction at, the same processing can be executed and the same effect can be obtained. As a method for obtaining these luminance waveforms, a technique that can reduce noise most and combine it with an actual non-inspection object to obtain high-precision signal processing may be selected.

  Further, a second signal processing method for determining the presence / absence and degree of the concavo-convex defect 31 from the image obtained by the imaging means 20 having a one-dimensional field of view will be described with reference to FIGS. . Here, FIG. 5A shows an image (master image) in which bright portions and dark portions are regularly arranged with a periodic width in the horizontal direction, and FIG. 5B shows an uneven defect 31 on the surface S to be inspected. FIG. 5C shows an image (inspection image) obtained by the photoelectric conversion means 21 having a one-dimensional field of view, and FIG. 5C is an image obtained by performing the operation 3 on FIGS. 5A and 5B. It is.

  The calculation 3 is a technique generally called difference processing, in which the luminance difference between the master image and the inspection image is calculated between pixels having the same coordinates, for example, the absolute value of the luminance difference is converted into new luminance, and the processed image , And a process of emphasizing portions with different luminance in the master image and the inspection image. When the operation 3 is performed on the master image shown in FIG. 5A and the inspection image shown in FIG. 5B, the master is obtained in the region 51 that is not affected by the axial deviation of the incident light Li due to the uneven defect 31 in the inspection image. Since there is no luminance difference from the image, it is converted as a uniform low-luminance portion as shown in region 53 of (c). On the other hand, in the region 52 affected by the axial deviation of the incident light Li due to the concave / convex defect 31, there is a portion causing a luminance difference from the master image, and a part thereof has a high luminance as shown in the region 54 of (c). Converted as part. When the uneven defect 31 does not exist on the surface S to be inspected, an image like (a) similar to the master image should be obtained as the inspection image. In this case, if the operation 3 is performed on the master image and the inspection image, an image (not shown) in which the entire image is uniformly low in brightness should be obtained. That is, if the presence or absence of a bright portion in the image (c) obtained by the calculation 3 is detected, it is possible to determine the presence or absence of the uneven defect 31 on the surface S to be inspected.

  Further, as described above, the direction ψ and the magnitude | i | of the movement vector i change depending on the slope angle α and the rotation direction angle β of the concavo-convex defect 31, so the bright / dark part of the bright / dark part appearing in the inspection image depends on the degree of the concavo-convex. The degree of periodic disturbance will also change. That is, in the larger and more severe (small and fine) uneven defect having the slope angle α and the rotation direction angle β, the region 52 in which the periodic disturbance occurs in the inspection image (b) becomes large (small) and The total number of bright / dark parts will increase (decrease). By analyzing these information, the degree of the concavo-convex defect 31 can be relatively evaluated.

  Here, a master image may be created in advance, or a second inspection image acquired at a time different from the time for acquiring the target inspection image may be used. However, when the second inspection image is used as the master image, the defect cannot be detected if the inspection image by the same unevenness defect 31 and the second inspection image as the master image are set as the processing targets of the calculation 3. For example, if five inspection images I1 to I5 are acquired at the imaging times T1 to T5, and if the concavo-convex defect 31 appears at the times T2 to T4, the calculation 3 between I2, I3, and I4 is performed. It is clear from the fact that the luminance change due to the uneven defect 31 is not detected. For this reason, it is necessary to employ | adopt the 2nd test | inspection image in the condition where the uneven | corrugated defect 31 does not exist as a master image. As the simplest method for this purpose, for example, when acquiring test images several times continuously, an inspection image acquired before a predetermined time from acquisition of an inspection image to be inspected is adopted as a master image, The master image may be updated as needed. In the above example, if the master image for the inspection image Ix is set as the inspection image Ix-1, the operation 3 is executed for the master image I1 and the inspection image I2 and for the master image I4 and the inspection image I5. Therefore, it is possible to detect the presence of the concavo-convex defect 31, and it is difficult for the defect to be overlooked.

  Note that the signal processing method described above is the most basic part for easily understanding the principle. In practice, in addition to this method, the signal processing accuracy is generally improved. It is also preferable to perform a noise removal (for example, averaging processing, median processing, smoothing processing, etc.) method or binarization method used at an appropriate stage.

  Further, the second signal processing method for determining the presence / absence and degree of the concavo-convex defect 31 from the image obtained by the imaging means 20 having a one-dimensional field of view is obtained by the imaging means 20 having a two-dimensional field of view. It can also be used as a method for determining the presence or absence and degree of the irregularity defect 31 from the obtained image.

  Moreover, in all the signal processing methods for determining the presence / absence and the degree of the concavo-convex defect 31 from the image obtained by the imaging unit 20 described above, it is assumed that the bright and dark portions can be clearly distinguished in the image before processing. It has become. Therefore, the apparent bright part width 35 and dark part width 36 shown in FIG. 2A and the apparent bright part width 37 and dark part width 38 shown in FIG. The widths of the light transmitting part 12a and the light shielding part 12b of the light / dark pattern light creating means 12 are preferably set. However, the fact that too many pixels are included in the apparent bright part widths 35 and 37 and the dark part widths 36 and 38 is due to the movement of the imaging point caused by the axial misalignment of the incident light Li due to the uneven defect 31 (generation of the movement vector i). Therefore, the pitches pa and pb of the light transmission part 12a and the light-shielding part 12b of the light / dark pattern light creating means 12 are both the above formulas 31 and 32. It is preferable to set so as to satisfy. It is also possible to increase the pixels included in the apparent bright part widths 35 and 37 and the dark part widths 36 and 38 by moving the camera position closer to the reflection point Pr and narrowing the image pickup area. Narrowing also causes disadvantages such as an increase in the number of cameras, so the pitch of the light transmitting part 12a and the light shielding part 12b of the light / dark pattern light creating means 12 is set to the apparent light part width 35, 37 and the dark part width 36, 38. It is also preferable that the included pixels are set to be within 30 pixels. Furthermore, if the above condition is satisfied, it is not necessary that the bright portion width pa and the dark portion width pb are equally spaced.

  In the above, the light / dark pattern light irradiating means 10 includes the light / dark pattern light creating means 12 in which the light transmitting part 12a and the light shielding part 12b are regularly arranged in the light emitting part of the illumination means 11 having a two-dimensional spread. Although described as being realized, when the specification of the concave / convex defect 31 to be detected is changed, a new bright / dark pattern light creating means 12 ′ in which the configurations of the light transmitting portion 12a and the light shielding portion 12b are newly changed. It is necessary to carry out a construction to replace the light / dark pattern light creation means 12 installed in the current illumination means 11 with The light / dark pattern light creating means 12 is typically constructed by applying a light-shielding paint to a transparent material such as acrylic or film. In this case, the illumination means 11 and the light / dark pattern light creating means are used. It is also conceivable that heat is easily accumulated during 12 and the luminous efficiency of the illuminating means 11 is lowered or the life is shortened. Further, if the light shielding material is disposed only in the light shielding portion 12b and no material is disposed in the light transmitting portion 12a, the intensity of the light / dark pattern light creating means 12 itself may be insufficient.

  Here, essentially, the light / dark pattern light irradiating means 10 is only required to irradiate the surface S to be inspected with the light / dark stripe pattern having a predetermined angle θ and pitch p. It is also preferable that the light / dark pattern light irradiation means 10 is settable. For example, it is conceivable to produce light and dark stripe pattern light by electrically controlling the light emission state of a display having a large and flat light emission surface represented by a plasma display panel. According to this embodiment, when the specification of the uneven defect 31 to be detected is changed, the light emission pattern of the reference surface Sl may be changed from the electrical control system, and the work for replacing the light / dark pattern light creating means 12 is performed. The burden of carrying out is eliminated. In addition, it is not necessary to consider the problem of reduction in luminous efficiency and life shortening as the illumination means 11, and the accessory part of the light / dark pattern light creating means 12 is not necessary, so that it is not necessary to consider its intensity.

  In the above description, the case where a crease generated in a transparent plastic moving in one direction in a plane as shown in FIGS. 1 and 3 has been described. However, in the process of manufacturing a sheet-like product such as a transparent plastic film, in many cases, there is a step of winding the product into a roll. Need to be detected.

  This case will be described with reference to FIG. FIG. 6 is a schematic diagram showing a second basic configuration of the present invention. When considering the detection of the concavo-convex defect 61 generated on the inspection target surface Sr on the roll inspection target 62, basically, the flat inspection target surface S moving in one plane in the above-described manner is basically described above. The principle of detecting the generated irregularity defect 31 can be applied. Although the shape of the reflecting surface has changed from a flat surface to a curved surface, there is no difference in the basic optical principle regarding the reflection point Pr, incident light Li, and light emission point Pi for one reflected light Lr, and the same optical equipment arrangement as above The concave / convex defect 61 that is a detection target can be detected by the configuration concept and signal processing.

  However, in this case, when detecting the concave / convex defect 61 generated on the inspection target surface Sr on the roll-shaped inspection target object 62 being rolled up while performing the rotational motion in the rotational direction r1, the plane is one direction in a plane. Compared with the case of detecting the concave / convex defect 31 occurring on the moving flat surface S to be inspected, the difference is that the bright / dark pattern light illuminating means 10, the imaging means 20 and the surface to be inspected during the manufacture of the product. This is a change in the optical positional relationship between S and Sr. That is, in the latter, the optical positional relationship between the light / dark pattern light illuminating means 10 and the imaging means 20 and the surface S to be inspected is unchanged during product manufacture, whereas in the former, the sheet-like object 60 to be inspected is rolled. The distance between the surface Sr to be inspected and the rotation center axis c of the roll is changed (thickening) and / or the rotation center axis c of the roll is formed while the rolled object to be inspected 62 is formed However, the distance of the winding core 63 itself relative to the position at the start of winding changes by the roll-shaped inspection object 62 being pushed and moved by the contact roll 65 for forming the roll whose position is fixed ( When the arm 64 moves), a distance change (hereinafter referred to as a roll distance change) DR is generated, and it always changes over time. In other words, with the occurrence of the distance change DR, the distance DL (hereinafter referred to as the illumination distance) DL from the light / dark pattern light illumination means 10 to the inspection target surface Sr immediately after the start of winding, and the imaging means 20 to the inspection target surface Sr. Distance (hereinafter referred to as imaging device distance) DC constantly changes during product manufacturing.

  The following two points are related to the inspection accuracy due to the change in the illumination distance DL and the imaging device distance DC described above. That is, for example, the first point is that the focus of the imaging by the light converging means 22 is disturbed by the change of the optical path length (DL + DC), and the obtained elephant is blurred. Furthermore, the direction ψ and the magnitude | i | of the movement vector i change as the optical positional relationship between the light / dark pattern light illuminating means 10, the imaging means 20 and the surface S to be inspected changes, A second point is that the amount of distortion of the stripe appearing in the stripe image obtained by the imaging means 20 changes.

  First, since the focus is disturbed with the change of the optical path length (DL + DC) generated with the roll distance change DR, it is preferable to correct this in real time.

  As a specific method, for example, as a first correction method, a condensing means 22 having a function of automatically correcting focus, for example, an autofocus function may be used. As a second correction method, the imaging means 20 may be moved in the optical axis direction m2 using the moving means 23 in response to the roll distance change DR that occurs during product manufacture. Here, the roll distance change DR may be measured in real time by a displacement measuring means (not shown) separately provided, or the information may be used if a change with time can be calculated in advance from manufacturing conditions. Further, as a third correction method, when the illumination distance DLmax at the start of winding, the imaging device distance DCmax, and the roll distance change DRmax at the completion of winding are defined, the design value of the focus is DLmax + DCmax−DRmax and Imaging means 22 having a depth of field of ± DRmax may be used. If the imaging means 22 having such a design is used, even if a roll distance change DR occurs during manufacturing, the captured image will not be blurred. By combining the above two methods, it is possible to change the focus or optical path length in stages.

  Next, with respect to the second point, when an uneven defect 61 on the surface Sr to be inspected is to be detected, the condition that the detection sensitivity is the lowest (the stripe distortion of the captured image due to the uneven defect 61 is small) is the roll Usually, it can be said geometrically that it is immediately before winding is completed (roll distance change≈DRmax). That is, this decrease in detection sensitivity is caused by the fact that the magnitude | i | of the movement vector i decreases as the illumination distance DL decreases. The angle φ of the imaging means 20 and the light / dark pattern light illumination means 10 with respect to the normal Z of the surface Sr to be inspected is also changed by changing the distance of the surface Sr to the imaging means 20 and the light / dark pattern light illumination means 10. The direction ψ and magnitude | i | of the movement vector i change accordingly, but from the viewpoint of change in detection sensitivity, the magnitude of the movement vector i | i | It has been obtained as a geometrical finding that it is almost dominant. In other words, the angle θ and the pitch p (that is, the bright portion width pa and the dark portion) of the light and dark stripe pattern described above according to the conditions immediately before the completion of the winding of the roll (roll distance change DR≈DRmax) with the lowest detection sensitivity of the uneven defect 61 The width pb) should be optimized. If these conditions are set, the inspection sensitivity becomes higher during the production of the product roll (roll distance change DR <DRmax) than immediately before the roll winding is completed (roll distance change DR≈DRmax). There is no oversight.

  In addition, the roll-shaped inspection object 62 rotates in the rotation direction r1, and accordingly, the inspection object surface Sr and the concavo-convex defect 61 also move. Therefore, the shorter the exposure time of the imaging means 20 is, the sharper the image becomes, and an image in which the bright / dark areas of the light / dark stripe pattern can be easily distinguished can be obtained. However, if the exposure time is too short, a sufficient exposure amount for image acquisition cannot be obtained. Therefore, the exposure time of the imaging means 20 is preferably set so that the rotational movement distance of the surface Sr to be inspected is 0.25 to 2.5 mm during exposure.

  More generally, the inspected object 30 with a flat surface and the inspected object 62 with a curved surface are in the horizontal direction in the field of view of the imaging means 20, that is, in the direction orthogonal to the moving direction of the inspected objects 30 and 62 If the resolution of the image pickup means 20 is increased in order to obtain a sufficient inspection sensitivity (so that the brightness of the stripe can be distinguished), a single image pickup means 20 is obtained. Therefore, it is difficult to acquire an image over the entire width of the inspection objects 30 and 62. Therefore, it is conceivable to arrange a plurality of imaging means 20 side by side in the width direction of the inspected objects 30 and 62. However, if the inspected objects 30 and 62 are wide, a significant increase in cost is caused. Therefore, one or a small number of imaging means 20 can be moved in the width direction of the inspected objects 30 and 62 by an imaging means moving means (not shown) so that the inspection is performed over the entire width of the inspected objects 30 and 62. It is also preferable to be configured.

  As described above, the method for inspecting surface irregularities according to the present invention detects a crease generated on a transparent plastic moving in one direction in a plane and a wound wrinkle generated on the surface of a transparent plastic film roll wound up in a roll shape. However, these methods can be actually configured using general optical equipment, image processing equipment, interfaces, information processing equipment, etc., and realized as surface roughness inspection equipment. It is what is done.

  The surface unevenness inspection method and inspection apparatus of the embodiment of the present invention as described above are transparent with a width of 1000 mm during movement at a speed m3 = 200 m / min in one direction in a plane shown in FIGS. 7 (a) and 7 (b). This was carried out in order to detect a crease 71 that is expected to occur on the surface S1 to be inspected of the plastic film 70. FIG. 7A is a schematic view of the transparent plastic film 70 being conveyed facing from the normal direction (direct direction) of the surface S1 to be inspected, and FIG. 7B is a schematic view of the cross section of the angle a1 to b1 portion. . Here, the slope angle α of the folding wrinkle 71 is 3 °, the rotational direction angle β is 8 °, and the wrinkle length L1 is 10 m. The wrinkle slope angle α assumed in the transparent plastic film 70 using the bent wrinkle 71 as a representative example is 2.2 ° to 3.8 °, and the rotation direction angle β is −10 ° to 10 °. The crease 71 may occur when a nip roll or the like bites a normal-direction flutter generated when the transparent plastic film is transported, and is a defective product.

  The equipment that constitutes the inspection apparatus for realizing the inspection method of the present invention includes a stage effect illumination having a light emission area of 1200 × 400 mm and a light / dark pattern light creating means 12 with the light / dark pattern irradiation means 10 as the illumination means 11. A general-purpose single-focus lens having a focal length of 35 mm as the imaging means 22 and a pixel as the photoelectric conversion means 21. A general-purpose CCD area sensor camera with a number of 640 x 480 pix was used. Further, from the mechanical relationship of the field equipment to which this apparatus is to be installed, the angle φ formed by the light / dark pattern light irradiation means 10 and the imaging means 20 with respect to the normal Z of the surface S1 to be inspected is 50 °, and the surface S1 to be inspected The distance from the light-dark pattern light irradiation means 10 to 1000 mm, and the distance from the surface S1 to be inspected to the imaging means 20 was 600 mm. Accordingly, the field of view of the camera by this optical system is 280 × 210 mm, the resolution is 438 μm / pix, and the width of the transparent plastic film 70 can be inspected in 10 divisions. In this case, the values of Ψmin and Ψmax are in the range of 0.1 ° to 18.2 °.

  Here, when the direction ψ and the magnitude | i | of the movement vector i are obtained from the above equations 27 and 28, ψ = 10.1 ° and | i | = 68.3 mm. Therefore, referring to the above-mentioned formulas 29, 30, 31, and 32, the inclination angle θ of the light and dark stripes is θ = 79.9 °, and the pitch p of the light and dark stripes is p = 13 mm (pa = 7 mm, pb = 6 mm). did. Further, as a signal processing method for determining surface irregularities from an image obtained by the imaging means, a differential processing method including the above-described calculation 1 and calculation 2 was adopted.

  The surface unevenness inspection apparatus configured as described above was installed at the production site of the transparent plastic film, and production was started. At the beginning of production, a transparent plastic film having a flat surface is produced smoothly. As shown in FIG. 2 (a), the stripe pattern of the image captured by the camera of the inspection apparatus is not disturbed, and the inspection apparatus is folded and wrinkled. Was not detected. However, because the tension of the stretching machine occurred in the middle, there was a flapping of conveyance, and this caused the nip roll to bite and the wrinkle was generated with the slope angle α = 4 ° and the rotation direction angle β = 10 °. As shown in (b), the stripe pattern of the image captured by the camera of the inspection apparatus is disturbed, and the wrinkle 71 can be detected by performing the signal processing consisting of the above-described operations 1 and 2 on the image. did it. In addition, there was a problem with the wrinkle removal device that caused even greater transport flutter than the above, and the nip roll bite into it, causing creases in the slant angle α = 3.6 ° and rotation direction angle β = 10 °. As a result, the stripe pattern of the image captured by the camera of the inspection apparatus is further disturbed, and the wrinkle 71 'can be detected by performing the signal processing including the above-described operations 1 and 2 on the image. . Also, by analyzing the processed image, specifically by analyzing the number of pixels of the area converted to high brightness, the number of areas themselves, the density, etc., the folding wrinkle 71 'is more than the previously detected folding wrinkle 71 It was relatively confirmed that it was a severe one.

  Another transparent plastic film was inspected. That is, the minimum limit of the folding wrinkle to be detected is changed to the folding wrinkle 71 ″ in which the slope angle α = 2 ° and the rotation direction angle β = 0 °. The wrinkle slope angle α assumed in the transparent plastic film 70 typified by the bent wrinkle 71 ″ is 1.6 ° to 2.4 °, and the rotation direction angle β is −4 ° to 4 °. In this case, the values of Ψmin and Ψmax are in the range of 0.1 ° to 8.1 °.

  In response to this, the direction ψ and the magnitude | i | of the movement vector i are obtained from the above equations 27 and 28, and ψ = 1.5 ° and | i | = 45 mm. Therefore, referring to the above-mentioned formulas 29, 30, 31, and 32, the inclination angle θ of the light and dark stripes is θ = 83.5 °, and the pitch p of the light and dark stripes is p = 10 mm (pa = 5 mm, pb = 5 mm). A striped pattern mask was prepared and attached to an inspection apparatus.

  Similar to the above, a bending wrinkle 71 '' with a slope angle α = 2 ° and a rotation direction angle β = 0 ° occurred due to abnormal tension in the drawing machine and problems with the wrinkle removal device during product manufacture. It was possible to detect without missing.

  Further, in order to realize an inspection region that is as wide as possible in the width direction by one set of imaging means 20 and / or to improve the inspection accuracy by reducing the resolution of imaging, the imaging means in the inspection apparatus configuration of the first embodiment is used. 20 is a configuration using a general-purpose single-focus lens with a focal length of 50 mm as the imaging means 22 and a general-purpose CCD line sensor camera with a pixel number of 2000 pix as the photoelectric conversion means 21. The distance from S1 to the imaging means 20 was 1200 mm. Therefore, the visual field width of the camera using this optical system is 623 mm, the resolution is 312 μm / pix, and the width of the transparent plastic film 70 can be inspected in three divisions. The surface unevenness to be detected was the above-described crease 71 ″ (slope angle α = 2 °, rotation direction angle β = 0 °). The wrinkle slope angle α assumed in the transparent plastic film 70 typified by the bent wrinkle 71 ″ is 1.6 ° to 2.4 °, and the rotation direction angle β is −4 ° to 4 °. In this case, the values of Ψmin and Ψmax are in the range of 0.1 ° to 8.1 °.

  As a signal processing method for determining surface irregularities from the image obtained by the imaging means, a method of analyzing the arrangement period of the light and dark pitches repeated in the horizontal direction of the image was adopted.

  The inspection apparatus for surface irregularities having the configuration changed as described above was installed at the production site of the transparent plastic film, and the production was started. As in Example 1, a transparent plastic film having a flat surface is produced smoothly at the beginning of production, and the horizontal period pitch cycle of the image captured by the camera of the inspection apparatus as shown in FIG. (... Pb, Pd, Pb ...) was not disturbed, i.e., the rectangular period of the luminance waveform shown in FIG. 2 (c ') was not disturbed, so the inspection apparatus did not detect any wrinkles. However, due to the occurrence of bending wrinkles with a slope angle α = 2 ° and a rotation direction angle β = 0 ° on the way, the repetition period of the horizontal brightness pitch of the image captured by the camera of the inspection apparatus as shown in FIG. (... Pb, Pd, Pb ...) appears in the region where the disturbance Pmx occurs, and accordingly, the rectangular period of the luminance waveform shown in FIG. '' Was detected.

  In the same configuration, the difference processing method including the above-described calculation 2 is adopted as a signal processing method for determining surface irregularities from the image obtained by the imaging means, and the master image is acquired once before the inspection image is acquired. The acquired inspection image was used. When the production of the transparent plastic film was started in the same manner as described above, a folding wrinkle 71 '' was generated between T2 and T4 during a certain period of time T1 to T5, but the inspection apparatus acquired the master image I1 ( By performing operation 3 on the inspection image I2 (with disturbance) acquired at time T2 and the master image I4 (with disturbance) acquired at time T4 and the inspection image I5 (without disturbance) acquired at time T5. In addition, it was possible to detect the brightness difference area between the master image and the inspection image and to detect the crease 71 ''. A pattern that has been created in advance and stored in the memory of the arithmetic unit may be used as the master image. In this case, the inspection images I1 to I5 and the master image I0 acquired at times T1 to T5 are used. Since the calculation 3 is performed, the crease 71 '' can be detected in the inspection images I2, I3, and I4.

  Furthermore, the light / dark pattern light irradiating means 10 is changed to a 50-inch plasma display panel in the configuration of the first embodiment, and the light emission state of the display (specifically, the angle θ and the pitch p of the light / dark stripe pattern) is changed by a dedicated arithmetic unit. ) Can be electrically controlled. With this change, when the target specification of the surface irregularities to be detected is changed, it is only necessary to change the light emission pattern of the reference surface Sl from the electrical control system, and the replacement work of the light / dark pattern light creating means 12 can be performed. The troublesome to carry out did not occur, and the production of the next transparent plastic film could be started quickly.

  Still further, the surface unevenness inspection method and inspection apparatus of the present invention has a width of 1000 mm during winding at a speed of m4 = 150 m / min in the rotation direction with the central axis CS shown in FIGS. This was carried out in order to detect winding wrinkles 81 that are expected to occur on the surface S2 to be inspected of the transparent plastic film roll 80. The transparent plastic film roll 80 operates with the arm supported by the fixed contact roll 65, winds up the transparent plastic film 70 to a thickness of 80mm, and the roll distance change DR becomes 150mm. 8A is a schematic view of the transparent plastic film roll 80, which is winding the transparent plastic film 70, viewed from the horizontal direction with respect to the floor surface, and FIG. 8B is a schematic view of the cross section of the angle a2 to b2 portion. It is. Here, the slope angle α of the wound wrinkle 81 is 2 °, the rotation direction angle β is 5 °, and the wrinkle length L2 is 100 mm. The wrinkle slope angle α assumed in the transparent plastic film roll 80 using the bent wrinkle 81 as a representative example is 1.6 ° to 2.4 °, and the rotation direction angle β is −8 ° to 8 °. The winding wrinkles 81 may occur due to improper adjustment of the contact roll 65, distortion of the winding core 63, uneven thickness of the transparent plastic film 70, and the like, which is a defective product.

  As for the equipment constituting the inspection apparatus for realizing the inspection method of the present invention, the stage effect illumination having a light emission area of 1200 × 400 mm with the light / dark pattern irradiation means 10 as the illumination means 11 and the light / dark pattern light creating means 12 This is composed of a transparent acrylic plate using a stripe pattern mask in which the light shielding part 12b is formed by printing. The imaging means 20 is used as the imaging means 22, and a lens with an autofocus function is built in, and the number of pixels is 640 x 480 pix. A CCD area sensor camera was used. Also, from the mechanical arrangement of the field equipment to which this device is to be installed, the angle φ formed by the light / dark pattern light irradiation means 10 and the imaging means 20 with respect to the normal Z of the surface S1 to be inspected is 10 °, immediately after the start of winding. The distance from the surface S1 to be inspected to the light / dark pattern light irradiation means 10 at the time was 2000 mm, and the distance from the surface S1 to be inspected to the imaging means 20 was 1200 mm. The field of view of the camera by this optical system is 280 × 210 mm, the resolution is 438 μm / pix, and the width of the transparent plastic film roll 80 can be inspected in 10 divisions. In this case, as described above, at DL = 1850 mm where the inspection sensitivity is lowest, the values of Ψmin and Ψmax are in the range of 0.3 ° to 8.5 °.

  Here, from the equations 27 and 28 described above, the direction ψ and the magnitude | i | of the movement vector i immediately before the completion of winding are determined in consideration of the roll movement distance DR = 150 mm. Ψ = 4.7 °, | i | = 127.4mm. Therefore, with reference to the above-described equations 29, 30, 31, and 32, the inclination angle θ of the light and dark stripes is θ = 85.3 °, and the pitch p of the light and dark stripes is p = 30 mm (pa = 15 mm, pb = 15 mm). did. Further, as a signal processing method for determining surface irregularities from an image obtained by the imaging means, a differential processing method including the above-described calculation 1 and calculation 2 was adopted.

  Furthermore, in order to obtain a sharp image in which the edge of the light / dark stripe pattern is sharp and the bright / dark regions are easily distinguished, the rotational movement distance of the surface Sr to be inspected becomes 1.0 mm during the exposure time of the camera. It was set to 0.4 msec.

  In addition, a function for moving the imaging means 20 with respect to the central axis CS direction (horizontal direction) of the roll by a moving means (not shown) is added to each area divided into ten equal parts in the width direction of the roll. The operation to perform the sequential inspection can be performed. Specifically, the following operations are repeated from the origin to the non-origin side, which will be described later:… → Camera movement → Camera stop → Image acquisition → Camera movement →…, and when the last position is imaged, it returns to the origin position. Thereafter, the same operation was repeated again. Signal processing and inspection result output are performed while the camera is moving.

  In addition, it is also preferable that a plurality of images are acquired and inspected if the movement is within the start time of movement to the next position at one position. In this case, in general, the time required for signal processing is sufficiently faster than the time required for movement of the moving means, and it is virtually impossible that an increase in the amount of processing becomes a burden for inspection.

  The surface unevenness inspection apparatus configured as described above was installed at the production site of the transparent plastic film, and production was started. Immediately after the start of production, a transparent plastic film having a flat surface is produced smoothly. As shown in FIG. 2 (a), the stripe pattern of the image captured by the camera of the inspection apparatus is not disturbed, and the inspection apparatus is folded and wrinkled 81. Was not detected. However, when the laminated thickness of the transparent plastic film roll 80 reaches 5 mm from the start of winding, the slope angle α = 2 °, the rotation direction angle β = 5 °, and L2 = 100 mm due to the strain that the winding core 63 originally had The wrinkle 81 was generated on the transparent plastic film roll 80 having a total width of 1000 mm at a portion 750 mm from the right edge (this position is defined as the origin) shown in FIG. At the time of occurrence of the wrinkles, the camera was present at the position closest to the origin among the positions obtained by dividing the entire width of the transparent plastic film roll 80 into 10 units of 100 mm by the moving means. However, the inspection was continued while repeating the predetermined cycle immediately after the occurrence of wrinkles, and the wound wrinkles 81 could be detected by the same image pickup and signal processing method as in Example 1.

  When this example was tried on another occasion, when the laminated thickness of the transparent plastic film roll 80 became 75 mm from the start of winding, the slant angle α = 2.2 ° and the rotation direction due to poor rotation of the contact roll 65. Wrinkles 81 ′ having an angle β = 2 ° and L2 = 100 mm occurred on a transparent plastic film roll 80 having a total width of 1000 mm at a portion 340 mm from the origin. At the time of occurrence of this wrinkle, the camera was present at the sixth position from the origin to the non-origin side among the positions obtained by dividing the entire width of the transparent plastic film roll 80 by 10 by 10 mm by moving means. . However, the inspection was continued while repeating the above-described predetermined cycle immediately after the occurrence of wrinkles, and the wound wrinkle 81 ′ could be detected by the same image pickup and signal processing method as in Example 1.

  In the two cases in this embodiment, the image can be clearly picked up without blurring by the autofocus function built in the camera, and the inspection optical path length is increased by increasing the laminated film thickness of the transparent plastic film roll 80. It was confirmed that the test can be performed normally even if there is any change. For image blur correction, it is also preferable to move the camera in the direction of m2 using moving means 23 such as a general-purpose precision stage in accordance with the movement of the surface Sr to be inspected. It is also preferable to use a single focus lens whose depth of field is designed to be sufficiently longer than the change in the inspection optical path length.

  The present invention is not limited to a method and an apparatus for inspecting creases on a transparent plastic film having a flat surface, and a method and apparatus for inspecting a wrinkle on a transparent plastic film roll having a curved surface. Although it can be applied to a method and apparatus for inspecting a wrinkle on a winding roll of a film-like substance whose surface is coated with a high light reflectance material, the application range is not limited thereto.

It is the schematic which shows the 1st basic composition of this invention. It is the schematic which shows a stripe pattern image and a periodic signal. It is the schematic which shows a stripe pattern image and a periodic signal. It is the schematic which shows a stripe pattern image and a periodic signal. It is the schematic which shows a stripe pattern image and a periodic signal. It is the schematic which shows the basic principle of this invention. It is the schematic which shows the basic principle of this invention. It is the schematic which shows the signal processing method of the acquired image by a two-dimensional imaging means. It is the schematic which shows the signal processing method of the acquired image by a one-dimensional imaging means. It is the schematic which shows the 2nd basic composition of this invention. It is the schematic which shows a transparent plastic film and folding wrinkles. It is the schematic which shows a transparent plastic film and folding wrinkles. It is the schematic which shows a plastic film roll and winding wrinkles. It is the schematic which shows a plastic film roll and winding wrinkles. It is the schematic which shows the direction of the movement vector by the slope angle of a wrinkle, and a rotation direction angle. It is the schematic which shows the magnitude | size of the movement vector by the inclination angle of a wrinkle, and a rotation direction angle. FIG. 2 is a schematic diagram illustrating the geometric elements of the present invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Bright / dark pattern light illumination means 11 Illumination means 12 Bright / dark pattern light creation means 12a Light transmission part 12b Light-shielding part 20 Imaging means 21 Photoelectric conversion means 22 Imaging means 23 Moving means 30 Inspection object 31 whose surface is flat 31 Surface is flat Irregular defect on inspection object 31a Slope of uneven defect 32 Distorted portion 33 One-dimensional field of view in non-distorted stripe image 34 One-dimensional field of view in distorted stripe image 35 Bright portion width in stripe image by two-dimensional imaging means 36 2 Dark part width in a stripe image by a three-dimensional imaging unit 37 Bright part width in a stripe image by a one-dimensional imaging unit 38 Dark part width in a stripe image by a one-dimensional imaging unit 43 Luminance waveform obtained from a stripe image when there is no uneven defect 44 If there is a stripe image Obtained luminance waveform 51 Area not affected by uneven defect in inspection image 52 Area affected by uneven defect in inspection image 53 Area where brightness difference between master image and inspection image occurs 60 Sheet-like object to be inspected Object 61 Uneven defect on the surface to be inspected having a curved surface 62 Object to be inspected in a roll shape (surface is curved) 63 Core 64 Arm 65 Contact roll 70 Transparent plastic film 71 Folding wrinkle 80 Transparent plastic film roll 81 Wrinkle a1 Angle facing the crease a2 Angle facing the crease b1 Cross section line of the crease b2 Cross section line of the crease CS Central axis CL1 Intersection line between the plane including the sight line direction and normal and the surface to be inspected CL2 Line of intersection between the plane including the reference surface and the reference surface CL3 Infinite plane including the reference surface Intersection line between the surface of the object to be inspected DC imaging device distance DL illumination distance DR roll distance change Hv stripe pattern brightness change direction Hh stripe pattern brightness boundary direction
i Movement vector I Line-of-sight direction of image pickup means L1 Total length of folding wrinkle L2 Total length of wound wrinkle Li incident light Li1 Incident light incident on reflection point on surface of inspection object Li2 Incident light incident on reflection point on slope of uneven defect Lr Reflected light m1 Moving direction of moving means m2 Moving direction of arm m3 Direction of transport of transparent plastic film m4 Direction of rotation of transparent plastic film roll p Pitch of light and dark stripes (Pa + Pb)
Pa Bright part width (= bright pitch)
Pb Dark part width (= dark pitch)
Pb The pitch of the bright area in the region where the light-dark arrangement period is not disturbed on the image Pd The pitch of the dark area in the area where the light-dark arrangement period is not disturbed on the image Pi1 Light emission which emits incident light incident on the reflection point on the surface to be inspected Point Pi2 Light emitting point that emits incident light incident on the reflection point on the slope of the irregular defect Pmx Bright / dark pitch in the region where the light / dark arrangement period is disturbed on the image Pr Reflection point Pr1 Reflection point on the surface to be inspected Pr2 Reflection point on slope of uneven defect q Reference vector r1 Roll rotation (winding) direction r2 Contact roll rotation direction S Surface to be inspected of surface to be inspected S0 Visual direction of image pickup means and inspection Plane surface including normal of target surface S1 Surface to be inspected of transparent plastic film S2 Surface to be inspected of transparent plastic film roll S1 Reference surface Sr Surface to be inspected of object to be inspected with a curved surface Z Normal to surface of object to be inspected α Slope angle of irregularity defect β Rotational angle of irregularity defect φ Bright / dark pattern light irradiation means and imaging means are methods of surface to be inspected Angle formed with respect to line ψ Direction of movement vector (angle formed by reference vector and movement vector)
θ Light / dark stripe tilt angle

Claims (22)

A regular reflection image of the inspection surface S of the object to be inspected of the reference surface Sl having the brightness of a predetermined periodic bright and dark pattern is imaged by the imaging means, and the pattern is distorted from the pattern shape distortion of the obtained regular reflection image. A surface irregularity detection method for detecting streaky irregularities on the inspection surface S, when the angle formed by the normal Z of the surface S to be inspected and the line-of-sight direction I of the imaging means is φ,
The direction of the reference vector q perpendicular to the intersecting line CL2 of the plane S0 and the reference surface Sl including the normal Z and the line-of-sight direction I is the movement vector i in the same direction as the direction Hv of the periodic change of the periodic pattern. The angle ψ formed with respect to the above satisfies the following formula 1, and the cross line between the ratio inspection surface S and the direction Hh perpendicular to the direction Hv of the periodic change of the periodic pattern and the infinite plane including the reference surface Sl A method for inspecting surface irregularities, characterized in that the reference surface and the imaging means are installed so that an angle θ formed by the following equation 2 is satisfied.

Here, Ψmin can be a value between 0.5 and 4.0 ° for the slope parameter α, and can be expressed by the following Equation 3 when the slope parameter β can take a value between -25 and 25 °. The defined minimum value of Ψ0, Ψmax is the maximum value of Ψ0.

The movement vector i having a size | i | satisfying the following expression 4 and having a bright part with a width pa and a dark part with a width pb as the periodic light / dark pattern, and the representative width pc of the pattern | The method for inspecting surface irregularities according to claim 1, wherein the reference surface and the imaging unit are installed so as to satisfy 5.

Here, | i | min is assumed that the slope parameter α can take a value between 0.5 and 4.0 °, and the slope parameter β can take a value between -25 and 25 °. The minimum value of | i | 0 defined by Equation 6 can be taken, and | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

A regular reflection image of the inspection surface S of the object to be inspected of the reference surface Sl having the brightness of a predetermined periodic bright and dark pattern is imaged by the imaging means, and the pattern is distorted from the pattern shape distortion of the obtained regular reflection image. A method of detecting surface irregularities for detecting streaky irregularities on the inspection surface S, wherein the angle formed by the normal Z of the surface S to be inspected and the line-of-sight direction I of the imaging means is φ, The inclination with respect to the surface to be inspected α, the inclination of the longitudinal direction of the assumed streaky irregularities with respect to the intersection line CL1 of the plane S0 and the surface to be inspected S including the normal Z and the line-of-sight direction I, β, And when
The angle ψ formed by the direction of the reference vector q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl with respect to the movement vector i in the same direction as the periodic change direction Hv of the periodic pattern is expressed by the following equation (7). And the angle θ formed by the intersection of the direction Hh perpendicular to the periodic change direction Hv of the periodic pattern and the infinite plane including the reference surface Sl and the specific inspection surface S is expressed by the following equation (8): A method for inspecting surface irregularities, wherein the reference surface and the imaging means are installed so as to satisfy

Here, Ψmin can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and takes a value between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value Ψmax that can be taken by the value of Ψ0 defined by the following equation 9 when it can be obtained is the maximum value that Ψ0 can take.

The movement vector i having a size | i | satisfying the following expression 10 and having the bright part with the width pa and the dark part with the width pb as the periodic bright / dark pattern, and the representative width pc is the following expression 4. The method for inspecting surface irregularities according to claim 3, wherein the reference surface and the imaging unit are installed so as to satisfy 11.

Here, | i | min can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value of | i | 0 defined by the following equation 12 when the value can be taken, | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

A stripe-type pattern is used as the periodic light / dark pattern, and an image pickup means that picks up a two-dimensional image based on the reflected light from the surface S to be inspected, and an image pickup resolution of the obtained image is obtained. The inspection according to any one of claims 1 to 4, wherein each of the apparent bright part width and the dark part width of the stripe pattern on a three-dimensional image is set to include five or more pixels. Method. 6. The inspection method according to claim 1, wherein a stripe pattern is used as the periodic light / dark pattern, and the obtained two-dimensional image is subjected to a differentiation process in the longitudinal direction of the stripe. A striped pattern is used as the periodic light / dark pattern, and a one-dimensional image is picked up based on the reflected light from the surface S to be inspected as the imaging means, and the one-dimensional image obtained at predetermined time intervals is arranged. The imaging resolution of the obtained two-dimensional image is set to include five or more pixels for each of the apparent bright part width and dark part width of the stripe pattern on the obtained two-dimensional image. The inspection method according to claim 1, wherein the inspection method is performed. The inspection method according to claim 7, wherein the uneven portion signal is detected from a light / dark period disturbance of a stripe pattern of the obtained image. 8. The inspection method according to claim 5, wherein the uneven portion signal is detected by performing a comparison operation between the stripe pattern of the obtained image and a predetermined master image. An inspection method for detecting irregularities on a surface to be inspected from a pattern shape distortion of an image obtained by imaging a light and dark pattern projected on a surface to be inspected S, wherein the light and dark pattern is used to detect the light and darkness of each part in a light emitting surface. Is projected onto the surface S to be inspected by light / dark pattern light irradiation means that can be arbitrarily set. An inspection method for detecting irregularities on the surface to be inspected from the pattern shape distortion of the image obtained by imaging the light and dark pattern projected on the surface to be inspected S, the object to be inspected performing a rotational movement in a roll shape, The distance between the surface S to be inspected and the rotation center axis of the roll changes with time, and / or the rotation center axis of the roll moves, and the distance between the surface S to be inspected and the rotation center axis of the roll A method for inspecting surface irregularities, wherein blurring of a captured image associated with a change and / or movement of a rotation center axis of a roll is corrected. 12. The inspection method according to claim 11, wherein the exposure time of the imaging means is set so that the rotational movement of the surface S to be inspected during exposure is 0.25 to 2.5 mm. Pattern shape distortion of the obtained regular reflection image, comprising a reference surface Sl having a predetermined periodic brightness / darkness pattern brightness and an imaging means for capturing a regular reflection image of the inspection object on the inspection surface S A surface irregularity detecting device for detecting streaky irregularities on the surface S to be inspected, and when the angle formed by the normal Z of the surface S to be inspected and the line-of-sight direction I of the imaging means is φ,
The direction of the reference vector q perpendicular to the intersecting line CL2 of the plane S0 and the reference surface Sl including the normal Z and the line-of-sight direction I is the movement vector i in the same direction as the direction Hv of the periodic change of the periodic pattern. The angle ψ formed with respect to this satisfies the following equation (13), and the intersection line between the ratio inspection surface S and the infinite plane including the direction Hh perpendicular to the periodic change direction Hv of the periodic pattern and the reference surface Sl An inspection apparatus for surface irregularities, wherein the reference surface and the imaging means are installed so that an angle θ formed by

Here, Ψmin can be a value between 0.5 and 4.0 ° for the slope parameter α, and the following equation 15 when the slope parameter β can take a value between -25 and 25 °. The defined minimum value of Ψ0, Ψmax is the maximum value of Ψ0.

The magnitude | i | of the movement vector i satisfies the following expression 16 and the periodic bright / dark pattern has a bright part with a width pa and a dark part with a width pb. The surface irregularity inspection apparatus according to claim 13, wherein the reference surface and the imaging unit are installed so as to satisfy 17.

Here, | i | min is assumed to be a value when the slope parameter α can take a value between 0.5 and 4.0 °, and when the slope parameter β can take a value between -25 and 25 °. The minimum value of | i | 0 defined by Equation 18 can be taken, and | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

Pattern shape distortion of the obtained regular reflection image, comprising a reference surface Sl having a predetermined periodic brightness / darkness pattern brightness and an imaging means for capturing a regular reflection image of the inspection object on the inspection surface S A surface irregularity detecting device for detecting streaky irregularities on the surface S to be inspected, wherein the angle formed by the normal Z of the surface S to be inspected and the line-of-sight direction I of the imaging means is φ, and the assumed streaks The inclination of the unevenness with respect to the surface S to be inspected is α, the inclination of the assumed streaky unevenness in the longitudinal direction with respect to the intersection line CL1 between the surface S0 to be inspected and the plane S0 including the normal Z and the line-of-sight direction I. Is β,
The angle ψ formed by the direction of the reference vector q perpendicular to the intersection line CL2 of the plane S0 and the reference surface Sl with respect to the movement vector i in the same direction as the direction Hv of the periodic change of the periodic pattern is expressed by the following equation (19): And an angle θ formed by an intersecting line between an infinite plane including the reference surface Sl and the ratio inspection surface S and a direction Hh perpendicular to the direction Hv of the periodic change of the periodic pattern An inspection apparatus for surface irregularities, wherein the reference surface and the imaging means are installed so as to satisfy

Here, ψmin can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and takes a value between the assumed minimum value αmin and the maximum value αmax of the slope β. The minimum value Ψmax that can be taken by the value of Ψ0 defined by the following equation 21 when it is assumed to be Ψ0 is the maximum value that Ψ0 can take.

The magnitude | i | of the movement vector i satisfies the following expression 22 and the periodic bright / dark pattern has a bright part having a width pa and a dark part having a width pb, and the representative width pc is expressed by the following expression: The surface unevenness inspection apparatus according to claim 15, wherein the reference surface and the imaging unit are installed so as to satisfy 23.

Here, | i | min can take a value between the assumed minimum value αmin and the maximum value αmax of the slope α, and between the assumed minimum value βmin and the maximum value βmax of the slope β. The minimum value of | i | 0 defined by the following equation 24 when the value can be taken, | i | max is the maximum value of | i | 0. The representative width pc is the larger of the widths pa and pb.

The periodic light / dark pattern is a stripe pattern, and the imaging means is an area sensor camera in which photoelectric conversion elements are two-dimensionally arranged, and the imaging resolution is the appearance of the stripe pattern on the obtained two-dimensional image. The inspection apparatus according to claim 13, wherein the inspection apparatus is set so as to include five or more pixels for each of the bright part width and the dark part width. The periodic bright / dark pattern is a stripe pattern, and the imaging means is a line sensor camera in which photoelectric conversion elements are arranged one-dimensionally, and the imaging resolution is an apparent appearance on the two-dimensional image of the obtained stripe pattern. The inspection apparatus according to claim 13, wherein the inspection apparatus is set to include five or more pixels for each of the bright part width and the dark part width. An apparatus for inspecting surface irregularities, wherein the illumination means and the light / dark pattern light creating means are composed of light / dark pattern light irradiation means capable of arbitrarily setting the light and darkness of each part in the light emitting surface. An inspection apparatus provided with illumination means, light / dark pattern light creation means, imaging means, and image processing means, in which the object to be inspected is in a roll shape and performs a rotational motion with the roll center axis as the rotation center, and time As the distance elapses, the distance between the surface S to be inspected and the rotation center axis of the roll changes, and / or the rotation center axis of the roll moves from the position before the start of rotation. An apparatus for inspecting surface irregularities, comprising a mechanism for compensating for blurring of a captured image that accompanies a change in distance of a rotation center axis and / or movement of a rotation center axis of a roll. 21. The inspection apparatus according to claim 20, wherein the exposure time of the imaging means is set to 1/60 to 1/3000 sec. It includes a step of inspecting a transparent plastic film or a plastic film on which a metal thin film is deposited using the inspection method according to any one of claims 1 to 12 or the inspection device according to any one of claims 13 to 21. Manufacturing method of plastic film.

Images