JP2020106295A - Sheet defect inspection device - Google Patents

Abstract

To detect an anisotropic defect (for example, a film flaw and a flux flaw of a nonwoven fabric) which can be imaged only in specific optical conditions depending on a shape and an occurrence position of the defect.SOLUTION: A sheet defect inspection device comprises: illumination means for irradiating a sheet with illumination light; imaging means for imaging light generated when the illumination light irradiated from the illumination means is reflected or transmitted on the sheet; movable means connected to one or more of an imaging device provided in the imaging means and an irradiation device provided in the illumination means, and for switching a relative positional relation between the irradiation device and the imaging device with respect to the sheet, with the imaging region subjected to imaging by the imaging means maintained to be the same; image processing means for processing the image data imaged by the imaging means on the basis of each of a plurality of positional relations switched by the movable means and obtaining luminance feature amount data indicating feature amounts on luminance corresponding to a plurality of respective image data; and defect determination means for determining whether the imaging region includes a defect from the luminance feature amounts.SELECTED DRAWING: Figure 1

Description

The present invention relates to a method for automatically inspecting an appearance defect of a sheet-like material such as a non-woven fabric made of a resin film or a collection of short fibers by an optical method.

For sheet-shaped products such as films used for displays and magnetic recording materials, and non-woven fabrics used for substrates for filters and electrodes, one of the quality requirements to guarantee the performance and characteristics of the final product is the sheet surface. The management of defects and foreign substances existing in the. These are required to be managed with higher precision as the final products to which they are applied have higher performance and precision, and at present, it is necessary to detect minute defects and foreign substances of several to several tens of microns.

In order to maintain and improve the quality and control the production process as described above, sheet-like product manufacturing lines are inspected for defects and foreign substances on the sheet surface, and defects and foreign substances are generated in each process. It is important to check if there is any and maintain and improve product quality.
Under these circumstances, the following inspection technology for fine defects in the sheet-like product manufacturing process is known.

For example, Patent Document 1 discloses a technique for automatically inspecting appearance defects of a continuously running sheet-like object by an optical method and quantitatively and accurately detecting a plurality of types of defects having different characteristics. ..

For defect inspection of a normal sheet-shaped product, as in Patent Document 1, illumination for irradiating an object to be inspected including a defect with light and light received by the object to be inspected for receiving or irradiating light are received. An optical method for determining the presence or absence of a defect by arranging the image pickup device to be fixed to the inspection object and detecting a change in the characteristics of light reflection or transmission in the defect portion is used.

However, among the defects of sheet-shaped products such as films and nonwoven fabrics, there are defects that can be visualized only under specific optical conditions with respect to the shape of the defects. For example, a flaw having linear irregularities in a resin film or a defect in which agglomerated short fibers in a non-woven fabric form a bundle by aligning the fiber orientation in one direction and are exposed on the sheet surface (hereinafter, for convenience, For example, because the directivity of the light reflected or transmitted from the defect greatly differs depending on the irradiation direction of the light (there is anisotropy), the defect shape is irradiated with light from a specific direction. Then, an image capable of defect detection (a contrast can be obtained between a normal portion and a defective portion) can be obtained only under an optical condition in which reflected or transmitted light is imaged from a specific direction.

Several techniques have been disclosed as inspection techniques for such defects (hereinafter referred to as anisotropic defects for convenience) in which the directivity of such reflected/transmitted light differs depending on the incident direction of illumination light with respect to the defect. ..

For example, in Patent Document 2, an illumination optical system that supplies light to a sample, a first detection optical system that is arranged at a first elevation angle, and a second elevation angle that is higher than the first elevation angle are arranged. A plurality of second detection optical systems are provided to efficiently capture the light scattered by the directional defect.

In Patent Document 3, the surface of the sample to be inspected is irradiated with a focused laser beam, scattered light generated on the surface of the sample to be inspected is condensed in multiple directions, and the condensed scattered light is photoelectrically converted. In the device for inspecting defects existing on the surface of the sample to be inspected, the detection signals detected in multiple directions are added to detect minute defects, and the detection signals detected in multiple directions are individually processed to obtain anisotropy. Techniques for detecting defects are disclosed.

In addition, in Patent Document 4, anisotropic illumination and isotropic illumination are performed on an object to capture a plurality of images, and a first image obtained by anisotropic illumination and an isotropic image are obtained. There is disclosed an inspection apparatus capable of inspecting various defects by generating an inspection image from a second image obtained by sexual illumination and performing processing for inspecting an object based on the inspection image.

Further, in Patent Document 5, with respect to the inspection area of the inspection target surface, the direction of the inspection target surface is fixed, and the reflected light of the light emitted from different directions is received to capture the surface image, and the illumination direction is determined. There is disclosed a technique of detecting a defect from pixel information according to the above.

Further, Patent Document 6 discloses an automatic non-destructive inspection of a mechanical part using an apparatus including an image sensor, an illumination light source, and a means for moving the image sensor, the illumination light source, and the mechanical part with respect to each other. Method is disclosed.

Japanese Patent No. 6183875 JP, 2017-133830, A Japanese Patent No. 4394707 JP, 2017-67633, A JP, 2013-7589, A Special table 2012-514193 gazette

In order to visualize anisotropic defects, one inspection area is irradiated with light from all directions, and imaging is performed under many optical conditions such as imaging from all directions. It is necessary to determine the presence/absence of a defect from the acquired image, and in the inspection technique in which the irradiation device and the image pickup device are fixedly arranged as described in Patent Document 1, it is possible to miss the anisotropic defect as described above. is there.

In the techniques described in Patent Documents 2 and 3, the detection optical system is fixedly arranged under a plurality of conditions, detection signals of scattered light detected in multiple directions are added to detect a micro defect, and each detection signal is individually detected. It is possible to avoid overlooking the anisotropic defect by performing the processing. However, the irradiation direction of the illumination light is one direction, and the detection optical system installed under a plurality of angle conditions is also a fixed arrangement, so a certain limitation occurs in the optical conditions, and the anisotropic defect generation direction Depending on the positional relationship between the illumination/detection optical system, it may not be possible to obtain an image in which a contrast is obtained between the normal portion and the defective portion.

According to the techniques described in Patent Documents 4 and 5, a plurality of captured images obtained by controlling the light emitting point of one lighting device and switching the irradiation direction of light with respect to one inspection region are determined according to the irradiation direction. It is possible to detect a defect from image information. However, the imaging direction is fixed with respect to the object to be inspected, and similar to the above-described known technique, a certain limit is imposed on the optical condition. Therefore, the anisotropic defect generation direction and the position of the illumination/imaging device Depending on the relationship, it may not be possible to acquire an image in which a contrast is obtained between the normal portion and the defective portion.

According to the technique described in Patent Document 6, the image sensor, the illumination light source, and the object to be inspected are moved relative to each other based on the shape of the object to be inspected obtained by the three-dimensional measuring means, and the surface of the object to be inspected is moved. Although it is possible to determine the operation path for acquiring an image by following the change in shape, it is not possible to obtain the optimum image for defect detection by changing the optical conditions for detecting anisotropic defects.

Therefore, in the present invention, in the defect inspection of a sheet-shaped product, in order to detect an anisotropic defect that was difficult to detect by the conventional technique, the irradiation device and the image pickup device are relative to the inspection area of the inspection object. Provide an inspection apparatus that can automatically inspect whether or not a defect is included in an inspection area by performing image pickup under a large number of optical conditions while switching between different positional relationships and using image processing from a plurality of captured images with different optical conditions. ..

In order to achieve the above object, the sheet defect inspection apparatus of the present invention,
Illumination means for irradiating the sheet with illumination light, and imaging means for imaging the light reflected or transmitted by the illumination light emitted from the illumination means on the sheet,
The irradiation of the sheet is performed by being connected to at least one of the sheet, the image pickup device provided in the image pickup means, and the irradiation device provided in the illumination means, and the area taken by the image pickup means remains the same. Movable means for switching the relative positional relationship between the device and the imaging device;
The image data captured by the image capturing unit is processed in each of the plurality of positional relationships switched by the movable unit, and the brightness feature amount data that is the feature amount regarding the brightness corresponding to each of the plurality of image data is obtained. Image processing means,
Defect determining means for determining whether or not a defect is included in the imaging region based on the brightness feature amount data.

Note that “brightness” described here is a pixel gradation value on an image, and is described as “brightness” in the same meaning in the specification and claims unless otherwise specified.

The sheet defect inspection apparatus of the present invention preferably has any one of the following configurations (1) to (8).
(1) The imaging area is determined with reference to coordinate data of defects existing on the sheet obtained by another sheet defect inspection apparatus.
(2) The movable means continues to operate without stopping at each of the positional relationships.
(3) The movable means rotates the imaging means and the illuminating means in the same rotation direction by the same rotation angle about a rotation axis extending from the imaging area in a direction perpendicular to the surface of the sheet.
(4) In addition to (3) above, an incident angle that is an angle (acute angle) formed by the irradiation direction of the illumination means and the rotation axis, and an angle formed by the imaging direction of the image pickup means and the rotation axis (acute angle). ) Is equal to the imaging angle.
(5) The movable means is a vertical articulated robot having four or more degrees of freedom.
(6) The defect determining means determines the brightness feature amount Lk satisfying the following condition (a) and condition (b) in the plurality of brightness feature amount data L1... Ln (n is an integer of 3 or more). If it exists, it is determined that the image data Ik corresponding to the luminance feature amount Lk includes a defect.
Condition (a): In the luminance feature amount data L1... Ln, there is another luminance feature amount data whose difference (absolute value) with the luminance feature amount Lk is larger than a predetermined threshold value.
Condition (b): The number of other brightness feature amount data satisfying the condition (b) is larger than the predetermined number.
(7) When the movable means includes a defect in any of the plurality of image data, a plurality of new positions are set with reference to the positional relationship when the image data including the defect is captured. Work in a relationship.
(8) When none of the plurality of image data includes a defect, the movable means operates in a plurality of new positional relationships different from the plurality of positional relationships.

In the sheet defect inspection apparatus of the present invention, when the sheet to be inspected is a transparent or semi-transparent material and inspects whether or not there is a defect in which linear concaves or convexes are exposed on the surface of the sheet with a constant length. Suitable for

Further, the sheet defect inspection apparatus of the present invention, the sheet to be inspected is a non-woven fabric composed of short fibers, or a non-woven fabric composed of short fibers and resin, the aggregated short fibers unidirectionally its fiber orientation. It is also suitable for inspecting whether or not there are defects that are aligned in a bundle and exposed on the sheet surface.

According to the sheet defect inspection apparatus of the present invention, the movable portion connected to at least one of the imaging device and the irradiation device switches the relative positional relationship between the irradiation device and the imaging device with respect to the sheet to be inspected. A plurality of image data obtained by illuminating the sheet with illumination light and capturing the light reflected or transmitted by the sheet is processed, and defects are detected in the imaging area based on the feature amount data related to the brightness corresponding to each image data. By determining whether or not is included, it is possible to accurately inspect defects exposed on the surface of the sheet-shaped product including anisotropic defects. As a result, it is possible to achieve the effect of realizing control of the quality of the sheet-shaped product and the production process and preventing missing of defective products.

FIG. 1 is a schematic view showing an example of the sheet defect inspection apparatus of the present invention. FIG. 2 is a schematic view showing an example of the operation path of the movable part of the sheet defect inspection apparatus of the present invention. FIG. 3 is a schematic diagram showing an example of operation path parameters of the movable part of the sheet defect inspection apparatus of the present invention. FIG. 4 is a diagram showing an example of an image processing flow of the sheet defect inspection apparatus of the present invention.

A preferred embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic diagram schematically showing an example of a sheet defect inspection apparatus according to an embodiment of the present invention. A sheet defect inspection apparatus 1 according to the present invention includes an illumination device 3 that irradiates a sheet 2 that is an inspection object with illumination light 3a, and an imaging device 4 that images reflected light, scattered light, and transmitted light 2a from the sheet 2. An image processing device 5 that determines whether or not a defect (not shown) is present in the sheet 2 from the brightness feature amount data of the image obtained by performing image processing on the captured image captured by the imaging device 4, the illumination device 3, and the imaging It comprises a movable part 6 (6′, 6″) connected to each of the devices 4 and a movable control part 6a for controlling the operation of the movable part 6.

In general, when a defect or a bundle defect of foreign matter/fiber (not shown) which is treated as a defect on a product exists on the surface of the sheet 2, when the defect light is irradiated with the illumination light 3a, the defect is reflected. The light/scattered light/transmitted light 2a behaves differently from the reflected light/scattered light/transmitted light 2a when there is no defect on the surface of the sheet 2. Specifically, the intensity of the reflected light/scattered light/transmitted light 2a, the directivity (degree of diffusion), the reflection direction, and the like are different between the area where the defect exists and the area where the defect does not exist. It is possible to inspect whether or not there is a defect in the sheet 2 by imaging the change in the transmitted light 2a with the imaging device 4 and imaging it. Here, the illumination device 3 and the imaging device 4 do not necessarily have to be arranged on only one side of the sheet 2. For example, by disposing the illuminating device 3 and the imaging device 4 so as to face each other with the sheet 2 interposed therebetween, it is possible to image the light that has passed through the sheet 2, and it is possible to detect pinhole defects and the like existing in the sheet 2. Become. Further, the lighting device 3 and the imaging device 4 used are not limited to one each. For example, by installing the illuminating device 3 on the front and back of the sheet 2, it becomes possible to inspect for scratches and bundle defects and pinhole defects existing on the surface at the same time.

The sheet 2 to be inspected is, for example, continuously or intermittently conveyed by a conveying device 7 having a winding-out portion and a winding portion of the sheet 2, and a movable portion 6 installed in any of the conveying devices 7. The illumination light 3a is emitted from the illumination device 3 connected to. The reflected light/scattered light/transmitted light 2a from the sheet 2 is imaged by the imaging device 4 such as a line sensor or an area sensor connected to the movable portion 6 installed in any of the transport devices 6. At this time, the relative positional relationship P between the illumination device 3 and the imaging device 4 with respect to the sheet 2 in a certain imaging region 8 is determined by the movable unit 6 controlled by the movable control unit 6a. This positional relationship P can be arbitrarily set by the movable control unit 6a, and the movable unit 6 is controlled to switch the respective positions of the illumination device 3 and the imaging device 4, so that one imaging region 8 Thus, a plurality of positional relationships P1... Pn can be constructed. Further, by acquiring the image data I1... In of the imaging region 8 with the constructed positional relationships P1... Pn, respectively, it is possible to acquire an image for one imaging region 8 under a plurality of optical conditions. The image processing device 5 processes the plurality of image data I1... In obtained by the image pickup device 4 to obtain luminance feature amount data, and from the obtained result, a defect exists in the image pickup region 8 of the sheet 2. It is determined whether or not.

The sheet 2 does not necessarily have to be in a conveyed state, and a sheet cut out to a certain size may be placed in a stationary state. Further, by connecting the seat 2 to the movable portion 6 and switching the position and angle of the seat 2 and the distance between the seat 3 and the illumination device 3 and the image pickup device 4, the illumination device 3 and the image pickup device 4 in the image pickup area 8 can be moved relative to the seat 2. The physical positional relationship P may be constructed. For connecting the sheet 2 and the movable portion, for example, a method of joining a mold matching the size of the sheet 2 to the movable portion 6 in addition to a clipper, a suction hand, or the like is used. On the other hand, when the sheet 2 is continuously conveyed by the conveying device 7, in order to acquire the image data I1...In with a plurality of positional relationships P1... In order to prevent the imaging area 8 in 2 from changing, it is necessary to control the movable portion 6 by following the conveyance speed of the sheet 2. In this case, the movable control unit 6a must add the amount of movement associated with the conveyance of the sheet 2 itself to the positional relationship P of the movable unit 6, which complicates the construction of the control program. Further, when the conveying speed of the sheet 2 is too fast with respect to the moving speed of the movable portion, the sheet 2 passes through the movable range of the movable portion before the positional relationship P1... Pn is established by the movable portion 6. There are cases. Therefore, the intermittent operation of repeatedly carrying and stopping the sheet 2 is applied to inspect the inspection area 9 of the sheet 2 that can be imaged in the movable range of the movable portion 6 while the transportation is stopped, and then the sheet 2 is moved to the next inspection area. It is preferable to carry a fixed amount to 9a and stop.

It is not necessary to connect all of the illumination device 3 and the imaging device 4 used to the movable part 6. For example, one in which only the imaging device 4 is connected to the movable portion 6 and a plurality of lighting devices 3 are arranged so that the illumination light 3a can be emitted from all directions, or the light emitting point of the lighting device 3 can be arbitrarily controlled. If, for example, is used, an image can be acquired for one imaging region 8 under a plurality of optical conditions simply by switching the position to the imaging device 4 connected to the movable portion 6.

A preferred example of the sheet 2 of the inspection object and the defect of the detection object in the present invention is (a) a transparent or semi-transparent material such as a resin film, and the linear concaves or convexes have a constant length and have a sheet surface. Defects exposed to the surface, (b) non-woven fabric composed of short fibers or non-woven fabric composed of short fibers and resin, and aggregated short fibers are aligned in one direction to form a bundle and exposed on the sheet surface Defect (bundle defect) and the like. In the defects as shown in (a) and (b), the directivity of the light reflected or transmitted from the defect portion greatly differs depending on the irradiation direction of the light to the defect (there is anisotropy).

As an example, a scratch defect (linear protrusion) existing in a transparent resin film will be described. Since the surface of the film is smooth, the irradiated light is specularly reflected on the surface of the film in the case of a normal portion where no defect exists. When the film having a flaw defect is imaged in a state where the image pickup device is arranged at a position where the specular reflection light can be picked up, if the line direction of the flaw defect is orthogonal to the direction of irradiating the light, Since the light is scattered at the convex portion, the reflected light from the flaw defect portion that can be received by the imaging device becomes darker than the reflected light from the normal portion, and can be imaged as a dark defect. However, if the line direction of the flaw defect is parallel to the direction of irradiating the light, the light is hardly scattered in the flaw defect portion and almost specularly reflected, so that the difference in the image between the normal portion and the defect portion cannot be seen.

Such a defect (hereinafter referred to as an anisotropic defect for convenience) in which the directivity of reflected/transmitted light differs depending on the incident direction of the illumination light with respect to the defect causes light to be emitted from a specific direction with respect to the shape of the defect. An image capable of defect detection can be acquired only under the optical condition of irradiating and imaging reflected or transmitted light from a specific direction. In order to visualize this anisotropic defect, one inspection region is irradiated with light from all directions, and imaging is performed under a large number of optical conditions such that imaging is performed from all directions. It is necessary to determine the presence or absence of a defect from a plurality of acquired images. On the other hand, as in the present invention, the irradiation device 3 and the image pickup device 4 are connected to the movable part, and the relative positional relationship P between the illumination device 3 and the image pickup device 4 with respect to the sheet 2 is switched to one image pickup area 8. On the other hand, a method of realizing image acquisition under a plurality of optical conditions by acquiring image data I1... In with a plurality of positional relationships P1... Pn is preferable.

The irradiation device 3 irradiates the sheet 2 with the illumination light 3a so that the defective portion of the sheet 2 can be visualized on the image. The type and form of the irradiation device 3 is not limited as long as it can irradiate the surface of the sheet 2, but it is preferable that the irradiation device 3 can uniformly irradiate the illumination light 3a in a wider range. The wavelength of the illumination light 3a to be applied is not particularly limited, but the wavelength of the illumination light 3a includes the visible light region because the wavelength of the light received by the sensor of the general image pickup device 4 is often visible light. Is preferred. Further, in order to obtain reproducibility for each image pickup, it is preferable that the illuminance of the illumination light 3a has a constant value. The imaging device 4 includes a light receiving element group, receives the reflected light/scattered light/transmitted light 2a from the sheet 2, and images the sheet 2. The light receiving element group included in the imaging device 4 may be, for example, a light receiving element having a single light receiving sensitivity such as a monochrome camera, or a plurality of light receiving elements having different light receiving sensitivities such as a color camera. You may use what is.

The imaging area 8 is an area to be imaged by the imaging device 4 included in the inspection area 9. While the imaging area 8 is gradually shifted so as to image the entire inspection area 9, a plurality of positional relationships P1 are set in each imaging area 8. Although the image data may be acquired by Pn, if a sheet inspection apparatus B different from the sheet defect inspection apparatus according to the present invention is installed, the position of the defect candidate detected by this sheet inspection apparatus B It is also possible to inspect with the sheet inspection apparatus of the present invention as the image pickup area 8. For example, when a sheet inspection apparatus B installed upstream detects a defect area to be detected as well as a normal area having no defect as a defect (so-called over-detection), a person normally inspects the sheet inspection apparatus. The area on the sheet determined to have a defect in B is visually checked again and it is determined again whether there is a defect to be detected. With the sheet defect inspection apparatus of the present invention, the sheet inspection apparatus B is used. By making the position of the detected defect candidate as a clearer image as the imaging region 8 and inspecting it in detail, visual re-determination by a person can be eliminated.

The movable unit 6 determines an image pickup area 8 and a relative positional relationship P between the illumination device 3 and the image pickup apparatus 4 in the image pickup area 8 with respect to the sheet 2, that is, an optical condition for picking up an image of the image pickup area 8. The positional relationship P can be arbitrarily determined by the movable control unit 6a connected to the movable unit 6. The movable portion 6 does not necessarily have to be stopped in each of the positional relationships P1... PN. For example, while continuing to image the inspection area 8, the positional relationship P1, P2,... Pn between the illumination device 3 and the imaging device 4 with respect to the seat 2 is stopped along the operation path 10 shown by the broken line in FIG. You may switch continuously without it. As described above, the movable portion sequentially changes the positional relationship P while capturing an image in a moving image, so that an image of the imaging region P under a larger number of optical conditions can be acquired.

The operation path 9 of the movable part 6 for constructing the positional relationship P1, P2,... Pn between the illumination device 3 and the imaging device 4 with respect to the imaging region 8 is set in advance by the movable control part 6a. (Teaching), or as parameters for constructing the positional relationships P1, P2,... Pn, for example, the following (A), (B), (C), etc. are used as the inspection target sheet 2 The operation path 9 may be determined by setting the type and type of the product, or for each inspection (see FIG. 3).
(A) Distances L1 and L2 between the imaging area 8 and the lighting device 3 or the imaging device 4
(B) How many times the illumination device 3 and the imaging device 4 are rotated in the rotation direction 12 about the rotation axis 11 extending in the direction perpendicular to the surface of the sheet 2.
(C) An incident angle θ1 which is an angle (acute angle) formed by the irradiation direction of the illumination device 3 and the rotation axis 11 and an imaging angle θ2 which is an angle (acute angle) formed by the imaging direction of the image pickup apparatus 4 and the rotation axis 11 are defined. How many times will it change?

Particularly, as a preferable example of determining the positional relationship P1... Pn for detecting anisotropic defects, the angle (B) is rotated by the same rotation angle in the same rotation direction 12 in the illumination device 3 and the imaging device 4. However, the image of the imaging region 8 is acquired, and more preferably, the angle of (B) is changed in the same rotation direction 12 in the illumination device 3 and the imaging device 4 by the same rotation angle by 10 to 15°. While acquiring the image of the imaging region 8. Further, as a preferable example of determining the positional relationship P1... Pn for the detection of the surface flaw defect of the film, which is a typical product of the inspection target sheet 2, and the bundle defect of the non-woven sheet, the anisotropic defect described above In addition to the operation (B) in the preferred example of determining the positional relationship P1... Pn for the detection of (1), it is preferable to make θ1 and θ2 of (C) equal to each other.

A robot, for example, can be used as the movable portion 6, and orthogonal coordinate type, polar coordinate type, cylindrical coordinate type vertical multi-joint type, etc. can be applied, but it has a high degree of freedom of movement and can secure a wide movable range. It is preferable to use a vertical articulated robot. Further, one movable part 6 may be configured by combining a plurality of robots and actuators.

The movable unit 6 moves the illumination device 3 and the image pickup device 4 so as to have a predetermined positional relationship P1... Pn, but it is not always necessary to perform imaging only in the operation path of the positional relationship P1... Pn. Absent. For example, depending on the determination result (described later) of the image data group I1... In of the imaging region 8 captured by each of the positional relationships P1... It is also possible to set an operation path for constructing a simple positional relationship P+. At this time, when image data Iz including a defect is present in the image data group I1... In of the imaging region 8 imaged at each of the positional relationships P1... Pn, the defect is more clearly imaged and the size of the defect is increased. As an example for enabling more accurate measurement of characteristics such as shape, density, and the like, it is preferable to use the positional relationship Pz obtained by capturing the image data Iz as the reference position when constructing a new positional relationship P+. .. Further, even if the image data group I1... In of the imaging area 8 imaged with each of the positional relationships P1... Pn does not include a defect, it is rechecked whether the imaging area 8 is really defective. For the purpose of, for example, a new positional relationship P++ different from the positional relationship P1... Pn may be constructed to acquire an image of the same imaging region 8 under different optical conditions.

The image processing device 5 processes the image data group I1...In acquired for each of the plurality of positional relationships P1...Pn constructed by the movable unit 6 with respect to the imaging region 8, and acquires the brightness feature amount data. Then, based on the obtained result, it is determined whether or not there is a defect in the imaging region 8 of the sheet 2. The image processing device 5 may use a PC equipped with dedicated image processing software, or hardware for image processing, or the like, instead of which the imaging device 4 has a CPU and can perform imaging and image processing integrally. May be used. FIG. 4 shows an example of an image processing flow of defect detection and determination performed by the image processing device 5. First, in the brightness feature amount calculation process of U1, for the image data group I1...In acquired by imaging the imaging region 8 at each of a plurality of positional relationships P1...Pn, the brightness data L1...Ln of the image The process of calculating is executed. For the brightness data L calculated here, for example, the average brightness of the image, the maximum/minimum brightness, the standard deviation of the brightness, and the like are used, and the brightness data L is obtained by combining filter processing, expansion/contraction processing, subtraction processing, and the like on the image data. The residual area and the number of white pixels when the preprocessed image is binarized can also be used. In particular, the latter method is suitable when the brightness of the entire image varies among the images in the image data group I1... In. Further, the brightness data L does not necessarily have to be one, and a plurality of data obtained by the above-described image processing may be used.

When a defect is included in the imaging region 8 and a defect appears in any one or more of the acquired image data groups I1...In, the luminance change is large in the image data group I1...In. On the other hand, when the imaging area 8 does not include any defect and no defect is shown in any of the acquired image data groups I1...In, the luminance change is small in the image data group I1...In.

As an example for determining these, in the luminance feature amount difference processing of U2, the image data Ik of interest among the image data group I1... In obtained by imaging the imaging region 8 with the positional relationship P1... Pn The difference (absolute value) between the luminance data Lk and the other luminance data L1...Lk (excluding Ik) of the other image data I1...Ik (excluding Ik) excluding Lk is calculated to obtain Lk. The difference data S1... In the luminance feature amount difference data determination process of U3, the difference data S1... Sn calculated in U2 is compared with a predetermined first threshold th1, and the difference data S larger than th1 is counted as defect candidate data. In U4, the number G determined to be defect candidate data among S1... Sn is compared with a predetermined second threshold value th2, and when the number G of defect candidate data is larger than th2, a defect is found in the image data Ik. Is included, and it is determined that there is a defect in the imaging area 8 of the image data Ik on the sheet 2.

For the imaged defect, the feature amount related to the shape, area, density, etc. is calculated by image processing, etc., and the standard or threshold value for the calculated feature amount is set to determine whether or not the defect affects the product quality of the sheet 2. You can also do it. In addition, for example, a defect that affects the product quality of the sheet 2 is determined by using machine learning or the like in which an image in which a defective part of an image including a defect that affects product quality is labeled or a normal image in which a defective part does not exist is used as a teacher. It is also possible to determine whether or not.

INDUSTRIAL APPLICABILITY The present invention can be suitably used for appearance inspection of a sheet-shaped material.

DESCRIPTION OF SYMBOLS 1 sheet defect inspection apparatus 2 sheet 3 irradiation apparatus 4 imaging apparatus 5 image processing apparatus 6 (6', 6'') movable part 7 transport apparatus 8 imaging area 9 inspection area 10 operation path 11 rotating shaft 12 rotating direction

Claims (11)

Illumination means for illuminating the sheet with illumination light,
Imaging means for imaging the light reflected or transmitted by the sheet, the illumination light emitted from the illumination means,
Any one or more of the sheet, the image pickup device provided in the image pickup unit, and the irradiation device provided in the illumination unit are connected, and the image pickup region picked up by the image pickup unit remains the same. Movable means for switching the relative positional relationship between the irradiation device and the imaging device;
The image data captured by the image capturing unit is processed in each of the plurality of positional relationships switched by the movable unit, and the brightness feature amount data that is the feature amount regarding the brightness corresponding to each of the plurality of image data is obtained. Image processing means,
A sheet defect inspection device, comprising: a defect determination unit that determines whether or not a defect is included in the imaging region based on the brightness feature amount data. The sheet defect inspection apparatus according to claim 1, wherein the image pickup area is determined with reference to coordinate data of a defect existing on the sheet obtained by another sheet defect inspection apparatus. 3. The sheet defect inspection apparatus according to claim 1, wherein the movable means continues to operate without stopping at each of the positional relationships. 4. The movable means rotates the imaging means and the illuminating means in the same rotation direction by the same rotation angle about a rotation axis extending in a direction perpendicular to the surface of the sheet from the imaging area. Sheet defect inspection device of any of. The incident angle that is an angle (acute angle) formed by the irradiation direction of the illumination unit and the rotation axis is equal to the imaging angle that is an angle (acute angle) formed by the imaging direction of the imaging unit and the rotation axis. 4 sheet defect inspection device. The sheet defect inspection apparatus according to claim 1, wherein the movable means is a vertical articulated robot having four or more degrees of freedom. In the case where the defect determining means has a brightness feature amount Lk satisfying the following condition (a) and condition (b) in the plurality of brightness feature amount data L1... Ln (n is an integer of 3 or more): 7. The sheet defect inspection apparatus according to claim 1, wherein the image data Ik corresponding to the luminance feature amount Lk is determined to include a defect.
Condition (a): In the brightness feature amount data L1... Ln, there is another brightness feature amount data whose difference (absolute value) in size from the brightness feature amount Lk is larger than a predetermined threshold value.
Condition (b): The number of other brightness feature amount data satisfying the condition (b) is larger than the predetermined number. When any of the plurality of image data includes a defect, the movable unit operates in a plurality of new positional relationships based on the positional relationship when the image data including the defect is captured. The sheet defect inspection apparatus according to claim 1, wherein 9. The movable unit operates according to a plurality of new positional relationships different from the plurality of positional relationships when a defect is not included in any of the plurality of image data. Sheet defect inspection device. The sheet to be inspected is a transparent or translucent material, and it is inspected whether or not there is a defect in which linear concaves or convexes are exposed on the surface of the sheet with a constant length, and the inspection is performed. Sheet defect inspection system. The sheet to be inspected is a non-woven fabric composed of short fibers or a non-woven fabric composed of short fibers and a resin, and the aggregated short fibers are aligned in one direction to form a bundle and exposed on the sheet surface. 10. The sheet defect inspection device according to claim 1, which inspects whether or not there is a defect.

Images