JP5842479B2 - Ground inspection device and ground inspection method - Google Patents

Description

  The present invention relates to a formation inspection apparatus and a formation inspection method for inspecting formation of a thin sheet-like inspection object such as paper or nonwoven fabric.

  Thin sheet-like products such as paper and non-woven fabrics are generally manufactured while being inspected for moisture, thickness, texture (fiber distribution state), etc., in order to increase production efficiency and maintain a certain quality. It is. Conventionally, the inspection of the formation, which is one of the inspections performed on the products, has been performed by visual inspection by a skilled inspector. In recent years, the inspection of the formation that enables automatic inspection of the formation is possible. Equipment has been developed.

  A conventional formation inspection device performs image processing on image data obtained by photographing a product as an inspection object, and evaluates formation index (formation from the standard deviation of brightness, variance, average value, etc. The index used for the inspection is generally determined. For example, the image data is subjected to image processing to obtain the standard deviation and average value of luminance (product transmittance and absorbance), and the uniformity is obtained by dividing the standard deviation of luminance by the average value. The product is inspected using the formation index. For details of such a conventional ground inspection device, see, for example, the following Patent Documents 1 to 6.

  However, the above-mentioned formation index does not reflect the particle size (information indicating how light areas (dark areas) gather due to fiber unevenness). For this reason, a divergence may occur between the inspection result obtained by the ground inspection apparatus disclosed in Patent Documents 1 to 6 and the inspection result by a skilled inspector. Therefore, a ground inspection device that performs ground inspection by reflecting the granularity in the ground index has also been proposed.

  Specifically, in Patent Document 7 below, image data is processed so that a region to be inspected is divided into a plurality of regions having the same size as the granularity to be considered, and the uniformity of each of the divided regions is determined. By obtaining the formation index, the granularity is reflected in the formation index. Further, in Patent Document 8 below, the granularity is reflected in the formation index by using the existence ratio of frequency components that are considered to be sensitive to humans when determining formation.

Japanese Patent No. 2692299 Japanese Patent No. 2738066 Japanese Patent No. 2751410 Japanese Patent No. 2797474 Japanese Patent No. 3227852 Japanese Patent Laid-Open No. 06-050906 Japanese Patent Laid-Open No. 11-281589 JP 2000-009660 A

  By the way, human vision recognizes that the uniformity and the average luminance are the same even if the granularity is different, or that the average luminance is different even if the uniformity and the granularity are the same, the formation is different. be able to. FIG. 28 is a diagram illustrating images having the same uniformity and average luminance and different particle sizes, and FIG. 29 is a diagram illustrating images having the same uniformity and granularity and different average luminances. 28 shows an image in which the granularity increases in the order of (a), (b), (c), and FIG. 29 shows the average luminance in the order of (a), (b), (c). The image which becomes high is illustrated.

  Referring to the example shown in FIG. 28, it can be recognized that the image shown in (c) is the best, the image shown in (a) is good, and the image shown in (b) is the worst. . In addition, referring to the example shown in FIG. 29, it is recognized that the image shown in (a) is the best, the image shown in (b) is good, and the image shown in (c) is the worst. Can do. Thus, human vision can recognize the difference in formation according to the granularity and brightness (average luminance).

  However, the formation inspection devices disclosed in Patent Documents 1 to 6 described above use only the uniformity as the formation index, and do not use the formation index that reflects the granularity. For this reason, there is a problem that it is impossible to detect a difference between images having the same uniformity and average luminance as shown in FIG. In addition, the formation inspection devices disclosed in Patent Documents 7 and 8 described above use a formation index that reflects the granularity in addition to the uniformity, but use a formation index that takes lightness into consideration. Not a translation. For this reason, there is a problem that it is impossible to detect a difference between images having the same uniformity and granularity and different average luminances as shown in FIG.

  In addition, the formation index used in the conventional formation inspection apparatus is not unified, and different ones are used for each inspection apparatus. For example, a certain geological inspection device uses a geological index that decreases in value with a better geological formation, while other geological inspection devices use a geological index that increases in value with a better geological formation. It is a condition such as using. For this reason, if the type of the geological inspection device is different, even if the same product or inspection object is inspected, a different geological index is obtained. There is a problem that you can not.

  Furthermore, even in the same type of ground inspection device, instrumental differences between devices and equipment used for inspection (for example, differences in lighting devices, image input devices, optical components, etc.) and differences in inspection conditions (for example, If there is a difference in lighting conditions and shooting conditions), a different formation index will be obtained. For this reason, even if the same kind of ground inspection device is used, if there is no influence of instrumental differences and the inspection conditions are not exactly the same, the ground is objectively compared and evaluated only by the ground index. There is a problem that you can not.

  The present invention has been made in view of the above circumstances, and objectively compare the difference in texture according to the granularity and brightness similar to human vision using the texture index regardless of the difference in inspection conditions, etc. An object is to provide a ground inspection device and a ground inspection method that can be evaluated.

In order to solve the above problems, the ground inspection apparatus according to the present invention is a ground obtained by performing image processing on image data obtained by imaging at least one of one side and the other side of a sheet-like inspection object (P). In the ground inspection devices (1 to 4) for inspecting the formation of the inspection object using a composite index (DX), first data indicating a contrast for each spatial frequency by performing an FFT operation on the image data The first calculation means (55) for obtaining the first data obtained by the first calculation means, the observation viewing distance parameter for defining the distance for observing the inspection object, and the luminance for observing the inspection object. and a second data indicating the contrast sensitivity for each spatial frequency in the case of observing the test object by the observation condition defined by the observed brightness parameters defined for the by adding by multiplying each spatial frequency It is characterized in that it comprises a second calculating means for calculating a slip index (57).
According to the present invention, the FFT operation is performed on the image data indicating the image of the sheet-like inspection object to obtain the first data indicating the contrast for each spatial frequency, and the first data and the sheet-like inspection object are obtained. The formation index is calculated by multiplying and adding the second data indicating the contrast sensitivity for each spatial frequency in consideration of the brightness of the image of the object and the observation visual field for the image for each spatial frequency.
In addition, in the ground inspection apparatus according to the present invention, the second calculation means may calculate the value obtained by multiplying the first data obtained by the first calculation means and the second data for each spatial frequency. Thus, a predetermined weight is added and added.
In addition, the ground inspection apparatus of the present invention is characterized by comprising conversion means (51) for converting the image data into image data indicating the physical brightness of the image.
The ground inspection apparatus according to the present invention further includes an adjusting unit (52) that adjusts the image data converted by the converting unit so that the physical luminance of the image becomes a preset reference luminance. It is characterized by.
In addition, the ground inspection apparatus according to the present invention includes a setting unit (54) for setting a plurality of local regions for the image data converted by the conversion unit, and for each local region set by the setting unit, And a statistical calculation means (59) for calculating the formation index by performing a predetermined statistical calculation on the local formation index obtained by the calculation by each of the first and second calculation means. .
In addition, the ground inspection apparatus according to the present invention is configured to multiply the first data and the second data for each spatial frequency to obtain the third data obtained by the second calculation means at a predetermined spatial frequency. A third operation means (81n) for calculating a formation index for each class by adding each class as a region is provided.
Here, the ground inspection apparatus of the present invention is characterized in that the third calculation means performs a predetermined weighting defined for each class on the third data, and adds the weights for each class. Yes.
Alternatively, the ground inspection device according to the present invention is configured such that the third data obtained by the second calculation means by multiplying the first data and the second data for each spatial frequency is for each class defined in advance. 4th calculating means (81n) which calculates the formation index | exponent for every said class by multiplying and adding the weight data prescribed | regulated to is characterized by the above-mentioned.
Here, in the ground inspection apparatus according to the present invention, the fourth calculation means performs a predetermined weight defined for each class on a value obtained by multiplying the third data and the weight data. And adding for each class.
The formation inspection method of the present invention acquires image data indicating at least one image of one side and the other side of a sheet-like inspection object (P), and performs image processing on the image data to obtain the formation. A ground inspection method for inspecting the surface of the inspection object using an index (DX), wherein first data is obtained by performing an FFT operation on the image data and indicating contrast for each spatial frequency. Space when the inspection object is observed under the observation conditions defined by the step, the observation viewing distance parameter that defines the distance for observing the inspection object, and the observation luminance parameter that defines the luminance for observing the inspection object A second step of obtaining second data indicating contrast sensitivity for each frequency, the first data obtained in the first step, and the second data obtained in the second step are represented by a spatial frequency. It is characterized by a third step of calculating the formation index by adding and multiplying the.

  According to the present invention, the first data indicating the contrast for each spatial frequency is obtained by performing the FFT operation on the image data indicating the image of the sheet-shaped inspection object, and the first data and the sheet-shaped inspection object are obtained. When the human actually inspects the inspection object by multiplying and adding the second data indicating the contrast sensitivity for each spatial frequency in consideration of the brightness of the image and the observation visual field for the image for each spatial frequency The formation index reflecting the conspicuousness of the formation is calculated. For this reason, it is possible to objectively compare and evaluate the difference in texture according to the granularity and brightness similar to human vision regardless of the inspection condition and the like using the texture index.

It is a block diagram which shows the structure of the formation inspection apparatus by 1st Embodiment of this invention. It is a figure for demonstrating the visual MTF characteristic. It is a figure which shows an example of the MTF characteristic which shows the relationship between the spatial frequency of a fringe pattern, and contrast sensitivity. It is a figure which shows an example of a two-dimensional MTF characteristic. It is a figure which shows an example of the two-dimensional MTF characteristic which considered the anisotropy. It is a block diagram which shows the internal structure of the image process part with which the formation inspection apparatus by 1st Embodiment of this invention is provided. It is a figure which shows the formation distribution image and the formation index which show the two-dimensional distribution of the local formation index calculated | required from the image shown in FIG. It is a figure which shows the formation distribution image and the formation index which show the two-dimensional distribution of the local formation index calculated | required from the image shown in FIG. It is a figure for demonstrating the difference of the effect by the case where weighting calculation is not performed, and the case where weighting is performed when calculating a local formation index. It is a figure which shows an example of the formation pattern used in order to demonstrate the difference of the effect shown in FIG. It is a block diagram which shows the structure of the formation inspection apparatus by 2nd Embodiment of this invention. It is a block diagram which shows the structure of the formation inspection apparatus by 3rd Embodiment of this invention. It is a block diagram which shows the structure of the formation inspection apparatus by 4th Embodiment of this invention. It is a block diagram which shows the internal structure of the image process part with which the formation inspection apparatus by 5th Embodiment of this invention is provided. It is a block diagram which shows the internal structure of the image process part with which the underground inspection apparatus by 6th Embodiment of this invention is provided. It is a figure which shows an example of the classification method of the class in 6th Embodiment of this invention. It is a figure which shows an example of the image observed in 6th Embodiment of this invention. It is a figure which shows the formation distribution image calculated | required from the image shown in FIG. It is a figure which shows the formation distribution image classified by class calculated | required from the image shown in FIG. It is a figure which shows the several image which observed in 6th Embodiment of this invention. It is a figure which shows the formation index according to class calculated | required from each of the image shown in FIG. It is a block diagram which shows the internal structure of the image process part with which the formation inspection apparatus by 7th Embodiment of this invention is provided. It is a block diagram which shows the internal structure of the image process part with which the formation inspection apparatus by 8th Embodiment of this invention is provided. It is a figure which shows an example of the class influence degree determination image calculated | required in 8th Embodiment of this invention. It is process drawing which shows the outline of the production process of the paper as a test subject. It is a figure which shows the example of a change of the classification | category formation index DXn accompanying the change of operation conditions. It is a figure for demonstrating the correlation with the result of visual evaluation performed after printing, and the formation index calculated | required before printing, or the formation index according to a class. It is a figure which shows the image from which a uniformity and average brightness | luminance are the same, and a particle size differs. It is a figure which shows the image with the same uniformity and granularity, and different average brightness | luminance.

  Hereinafter, a ground inspection apparatus and a ground inspection method according to an embodiment of the present invention will be described in detail with reference to the drawings. In the following, in order to facilitate understanding, the inspection object is assumed to be paper, and a ground inspection device and a ground inspection method for inspecting the paper texture will be described as an example. However, the present invention can also be applied to a ground inspection apparatus and a ground inspection method for inspecting a ground of a thin sheet-like inspection object such as a nonwoven fabric in addition to paper.

[First Embodiment]
<Configuration of ground inspection equipment>
FIG. 1 is a block diagram showing the configuration of the ground inspection apparatus according to the first embodiment of the present invention. As shown in FIG. 1, the ground inspection device 1 of the present embodiment includes a lighting device 10, an image input device 20, and a control processing device 30, and paper P (inspection object) conveyed in the conveyance direction Z <b> 1. Image processing is performed on the image data obtained by photographing the image, and the formation of the paper P is inspected based on the formation index obtained by this image processing. The ground inspection apparatus 1 according to the present embodiment is a transmitted light type ground inspection apparatus that inspects the formation of the paper P from an image (transmitted light image) of light transmitted through the paper P.

  The illumination device 10 includes a light source 11 and an optical element 12 and is disposed on the back side of the paper P and irradiates the back surface of the paper P with illumination light. The light source 11 emits illumination light (for example, white light) to be applied to the paper P. As the light source 11, any of a point light source, a line light source, and a surface light source can be used, but a line light source is preferable to a point light source, and a surface light source is preferable to a line light source. Further, when the paper P is transported at high speed in the transport direction Z1, a light source that emits strobe light is desirable. The optical element 12 is a lens or the like, for example, and projects the illumination light emitted from the light source 11 onto the back surface of the paper P. The optical element 12 can be omitted if not necessary.

  The image input device 20 includes an optical element 21 and an imaging device 22, and is an image (one-dimensional image or two-dimensional image) that is arranged on the surface side of the paper P and is obtained by taking an image of the surface of the paper P. Output image data. Specifically, the image input device 20 is arranged such that a region photographed by the imaging device 22 overlaps a region irradiated with illumination light by the illumination device 10 with the paper P interposed therebetween.

  The optical element 21 is a lens or the like, for example, and guides light (transmitted light) transmitted through the paper P to the imaging device 22. The imaging device 22 includes an imaging device such as a photodiode, a CCD (Charge Coupled Device), and a CMOS (Complementary Metal Oxide Semiconductor), and outputs image data of a photographed image. To do. The image sensor 22 can be either one that outputs image data of a one-dimensional image or one that outputs image data of a two-dimensional image. In this embodiment, the image sensor 22 outputs image data of a two-dimensional image. Will be described as an example.

  The control processing device 30 includes an illumination control unit 31, an image capturing unit 32, a storage unit 33, an image processing unit 34, an inspection control unit 35, and an inspection result output unit 36, and illumination light irradiation control on the paper P. And image processing on the image data output from the image input apparatus 20 is performed. Then, the formation of the paper P is inspected using the formation index obtained by image processing, and the inspection result is output.

  Under the control of the inspection control unit 35, the illumination control unit 31 performs light amount control of illumination light irradiated on the paper P, control of irradiation timing (for example, control of timing of strobe light emission), and the like. The image capturing unit 32 is an image board, for example, and captures image data output from the image input device 20 under the control of the inspection control unit 35. The storage unit 33 is realized by a memory such as a RAM (Random Access Memory), for example, and stores the image data captured by the image capturing unit 32.

  Under the control of the inspection control unit 35, the image processing unit 34 reads the image data stored in the storage unit 33, performs image processing, and calculates a formation index for evaluating the formation of the paper P. Here, the formation index calculated by the image processing unit 34 reflects the degree of conspicuousness of the formation when the person actually inspects the paper P, and is a human visual perception limit (JND: Just This is a quantitative expression based on Noticeable Difference. Details of the formation index used in the present embodiment and details of image processing performed by the image processing unit 34 will be described later.

  The inspection control unit 35 controls the blocks provided in the control processing device 30 to control the operation of the ground inspection device 1 in an integrated manner. Specifically, the illumination control unit 31 is controlled to control the amount of illumination light irradiated on the paper P, and the image capture unit 32 is controlled to perform image data capture control. In addition, the image processing unit 34 is controlled to perform execution control of the image processing, and an inspection result using the formation index calculated by the image processing unit 34 is output. The inspection result output unit 36 outputs the calculation result of the formation index and the inspection result of the paper P using the formation index to, for example, a display, a database server, a recorder, and the like.

<Details of formation index>
Next, the details of the formation index used in the present embodiment will be described. In order to express the formation index on the basis of the human visual perception limit (JND), in this embodiment, the visual spatial frequency sensitivity characteristic, that is, the MTF (Modulation Transfer Function) characteristic is used. FIG. 2 is a diagram for explaining visual MTF characteristics.

  As shown in FIG. 2, the visual MTF characteristic indicates the relationship between the spatial frequency of the fringe pattern and the contrast sensitivity when a human observes the fringe pattern whose brightness changes in a sine wave shape. This is equivalent to the contrast sensitivity function (CSF) of the visual system. Here, the contrast sensitivity is the reciprocal of a limit contrast (hereinafter referred to as JND contrast) that can be sensed when a human observes a fringe pattern whose brightness changes in a sine wave shape.

Now, a street, a maximum luminance L _MAX shown in FIG. 2, the minimum luminance is L _MIN, consider a case of observing a fringe pattern background luminance is L _A. When the luminance difference ΔL that can be detected by humans when such a fringe pattern is observed, the JND contrast C _JND is expressed by the following equation (1), and the contrast sensitivity M is expressed by the following equation (2). expressed. Incidentally, the background luminance _{L A} and luminance difference ΔL is, for example, the following (3), respectively represented by equation (4).

  FIG. 3 is a diagram illustrating an example of the MTF characteristic indicating the relationship between the spatial frequency of the fringe pattern and the contrast sensitivity. Referring to FIG. 3A, it can be seen that the contrast sensitivity changes according to the spatial frequency of the fringe pattern to be observed, and becomes maximum at a certain spatial frequency f0. In the images shown in FIGS. 28A to 28C (images having the same uniformity and average luminance but different particle sizes), it is observed that the formation of the image shown in FIG. 28B is conspicuous. This is due to the difference in contrast sensitivity according to the spatial frequency.

  Here, the MTF characteristic also changes depending on the background luminance (brightness) and the viewing viewing angle. FIG. 3B is a diagram showing the MTF characteristics when the background luminance is changed, and FIG. 3C is a diagram showing the MTF characteristics when the observation viewing angle is changed. In FIG. 3B, the background luminance increases in the order of graphs L14 to L11, and in FIG. 3C, the observation viewing angle increases in the order of graphs L24 to L21.

  Referring to FIG. 3B, it can be seen that the contrast sensitivity increases as the background luminance increases. This means that a small contrast difference can be sensed when the background brightness is higher than when the background brightness is low. It can also be seen that the frequency at which the contrast sensitivity is maximized tends to increase as the background luminance increases. In the images shown in FIGS. 29A to 29C (images having the same uniformity and granularity but different average luminance), it is observed that the formation of the image shown in FIG. 29C is conspicuous. This is due to the difference in contrast sensitivity according to the background luminance. In addition, referring to FIG. 3C, it can be seen that the contrast sensitivity increases as the observation field increases.

In general, the visual MTF characteristic is expressed by a Gaussian function G shown in the following equation (5).
However, the variable u in the above equation (5) is a spatial frequency, and the variable σ is a variable that determines the spread of the Gaussian function G. The unit of the spatial frequency u is cpd (cycle per degree: [number of cycles / viewing angle]) indicating the number of cycles per viewing angle of 1 °.

FIG. 3 (b), the as described with reference to FIG. 3 (c), MTF characteristic (contrast sensitivity) is not only varies in accordance with the spatial frequency u, varies depending background luminance L _A and observation viewing angle d . Such spatial frequency background brightness in addition to u L _A and observation viewing angle d one-dimensional MTF characteristic in consideration of (contrast sensitivity M _B1) is represented by, for example, by the following equation (6).

However, the function X of (6) wherein is a function representing a decrease in contrast sensitivity M _B1 at high frequencies. The function phi ₁ is a function showing a characteristic of contrast sensitivity M _B1 is proportional to the square root of the background luminance L _A, function phi ₂ is a function representing a monotonic increase in the contrast sensitivity M _B1 at low frequencies is there. In the above equation (6), k is a predetermined constant. In the above (6), the Gaussian function G is used as a function representing a decrease sensitivity of the high frequency by the ocular lens system, the variable σ is used as a variable indicating the magnitude of the pupil determined by the background luminance L _A.

The details of the above formula (6) are disclosed in, for example, the following documents.
“PHYSICAL MODEL FOR THE CONTRAST SENSITIVITY OF THE HUMAN EYE”, Barten.PGJ, SPIE Proceedings, vol.1666, 57-64, (1992)
The above equation (6) is an international standard for DICOM (Digital Imaging and Communications in Medicine) called “Digital medical imaging and communication”. Part 14 “Image display function for unifying gradation characteristics of display devices” This model is used for standard design. The contrast sensitivity M _B1 may be a function that can approximate a one-dimensional MTF characteristic, and is not limited to the above equation (6). However, the function is preferably a function that takes the background luminance into consideration, and more preferably a function that takes the observation viewing angle into consideration.

The formation inspection apparatus 1 of the present embodiment inspects formation by taking a two-dimensional image of the paper P. Since the above expression (6) represents a one-dimensional MTF characteristic, it is necessary to use an expression representing a two-dimensional MTF characteristic in the present embodiment. When representing the spatial frequency u in the horizontal direction of the spatial frequency u _h and vertical spatial frequencies u _v, the following equation (7) a one-dimensional MTF characteristic represented by the above (6) formula (contrast sensitivity M _B1) The two-dimensional MTF characteristic (contrast sensitivity M _B1 ) shown can be extended.
However, the function u (u _h , u _v ) in the above equation (7) is a function shown in the following equation (8).

  FIG. 4 is a diagram illustrating an example of a two-dimensional MTF characteristic. The two-dimensional MTF characteristic shown in FIG. 4 is a concentric circle obtained by rotating the one-dimensional MTF characteristic shown in FIG. 3A around a DC component (a component having a frequency of 0). Therefore, for example, when the MTF characteristic along the broken line from the center point O to the end point A of the two-dimensional MTF characteristic shown in FIG. 4 is obtained, the one-dimensional MTF characteristic shown in FIG. 3A can be obtained. .

  Here, it is known that the two-dimensional MTF characteristic has anisotropy depending on the arrangement direction of the stripe pattern. FIG. 5 is a diagram illustrating an example of a two-dimensional MTF characteristic in consideration of anisotropy. As shown in FIG. 5, the two-dimensional MTF characteristic considering anisotropy has a white portion (a portion with high contrast sensitivity) extending from the center in the horizontal direction and the vertical direction. This is because the contrast sensitivity is high for a stripe pattern in which light and dark appear alternately in the horizontal direction (0 degree stripe pattern) and a stripe pattern in which light and dark appear alternately in the vertical direction (90 degree stripe pattern). This means that the contrast sensitivity is low for a stripe pattern (45-degree stripe pattern) in which light and dark appear alternately along an oblique direction with respect to both the horizontal direction and the vertical direction.

  The MTF characteristic considering the anisotropy shown in FIG. 5 is closer to the MTF characteristic of human vision than the MTF characteristic not taking the anisotropy shown in FIG. For this reason, by using the MTF characteristic considering the anisotropy shown in FIG. 5, it is possible to realize the same inspection as when a person inspects the formation of the paper P.

The correction function D that expresses anisotropy according to the arrangement direction of the fringe pattern is expressed by, for example, the following expression (9).
However, δ, ε, ζ, and η in the above equation (9) are constants, and the function θ (u _h , u _v ) is a function represented by the following equation (10).

The two-dimensional MTF characteristic (contrast sensitivity M ₂ ) in consideration of anisotropy shown in FIG. 5 is obtained by multiplying the above-described equation (7) by the above-described equation (9) as shown in the following equation (11). can get.

Here, as described above, the formation index calculated by the image processing unit 34 reflects the degree of conspicuousness of the formation when the person actually inspects the paper P, and the human visual perception limit. It is expressed quantitatively with (JND) as a reference. Now, has sinusoidal fringe pattern background luminance is L _A spatial frequency a u is formed on the paper P as the inspection object, consider the case being observed this in observation viewing angle d.

When the contrast of the stripe pattern with respect to the background luminance L _A is C and the contrast sensitivity of the spatial frequency u is M _u based on the MTF characteristics at the background luminance L _A and the observation viewing angle d, the conspicuous degree J is expressed by the following equation (12): Is required.

Referring to the above equation (12), the conspicuous degree J is obtained by multiplying the contrast C and the contrast sensitivity _Mu . Here, as shown in the above-described equation (2), the contrast sensitivity is the reciprocal of the JND contrast, which is the limit contrast that can be sensed when a human observes a fringe pattern whose brightness changes in a sine wave shape. is there. For this reason, as shown in the above equation (12), the conspicuous degree J means that the contrast C is also divided by the JND contrast C _JND , and therefore means “how many times the contrast C is the JND contrast C _JND ”. To do. In the present embodiment, a conspicuous degree J in which the JND contrast C _JND is a unit (hereinafter, this unit is expressed as [JND]) is used as the formation index.

  The conspicuous degree J takes a value of 0 [JND] or more. For example, when the value of the conspicuous degree J is 0 [JND], it means that the inspection object is completely uniform, and when it is 1 [JND], it indicates that the inspection object is at the detection limit. means. Further, when the value of the conspicuous degree J is smaller than 1 [JND], it means that it cannot be detected by humans, and when it is 1 [JND] or more, it means that humans can detect it. In addition, since the formation becomes conspicuous as the value of the conspicuous degree J increases, it means that the formation is bad.

Incidentally, as shown in equation (12), to determine the formation index (noticeable degree J), it is necessary to obtain a contrast C to the background luminance L _A of the fringe pattern. Here, the image data obtained by photographing the paper P as the inspection object is in a state where a plurality of spatial frequency components are overlapped. For this reason, a Fast Fourier Transform (FFT) is performed on the image data to obtain a contrast distribution image indicating the contrast for each spatial frequency. In order to easily obtain a contrast distribution image, the image data output from the image input device 20 is converted into image data indicating the physical brightness of the image obtained by the image input device 20, and then the FFT operation is performed. Is desirable.

<Configuration of image processing unit>
Next, a specific configuration of the image processing unit 34 that calculates the above-described formation index (conspicuous degree J) will be described. FIG. 6 is a block diagram showing an internal configuration of an image processing unit provided in the ground inspection apparatus according to the first embodiment of the present invention. As shown in FIG. 6, the image processing unit 34 includes a measurement region setting unit 41, an image correction unit 42, and a formation index calculation unit 43, and the image input condition data D <b> 11 output from the inspection control unit 35 and the observation In accordance with the condition data D12, the image data D1 read from the storage unit 33 is subjected to image processing to calculate the formation index DX.

  Here, the image input condition data D11 output from the inspection control unit 35 eliminates the influence of instrumental differences between the illumination device 10 and the image input device 20 that are used to obtain an image of the paper P as the inspection object. For example, data including a luminance conversion characteristic parameter, an exposure time parameter, a noise characteristic parameter, and an actual size conversion parameter. On the other hand, the observation condition data D12 output from the inspection control unit 35 is data for setting the observation conditions of the paper P as the inspection object to the same observation conditions as when the human observes the paper P. This is data including an observation viewing distance parameter and an observation luminance parameter.

The luminance conversion characteristic parameter included in the image input condition data D11 is a luminance value (for example, the unit is [cd / m] with the image data output from the image input device 20 (image data D1 read from the storage unit 33) as a physical quantity. ² ] is a parameter used for conversion to a luminance value). The exposure time parameter is a parameter indicating the exposure time when the paper P is photographed by the imaging device 22 included in the image input device 20.

  As the above-described luminance conversion characteristic parameter, for example, the image input device 20 captures image data by obtaining the image data while changing the luminance of a light source having a stable luminance (for example, a light source including an integrating sphere) in advance. The value of the conversion table created based on the gradation characteristics of the image data with respect to the exposure time or the coefficient of the conversion formula is used.

Here, generally, the gradation V of the image data output from the image input device 20 is proportional to the exposure time T, and is proportional to the input luminance L when the image input device 20 is for industrial use. When the image input device 20 is for consumer use, it is proportional to the power of the input luminance L. For this reason, the above conversion equation is expressed, for example, by the following equation (13), assuming that the exposure time when photographing a light source with stable luminance is T ₀ .
The coefficients α, β, γ, and T ₀ in the above equation (13) are used as the above-described luminance conversion characteristic parameters. Note that the coefficient β in the equation (13) has a value of “1” when the image input device 20 is for industrial use.

  The noise characteristic parameter is a parameter for removing the influence of light shot noise and dark current of the image input device 20. Here, the light shot noise is generally proportional to the square root of the signal amount, and increases as the pixel value of the image data increases. As the noise characteristic parameter, for example, the image input device 20 captures image data by obtaining the image data while changing the luminance of a light source having a stable luminance (for example, a light source including an integrating sphere) in advance. A conversion table value created based on the characteristics of gradation variations or a coefficient of a conversion equation is used.

The above conversion equation is expressed by, for example, the following equation (14), where N is the amount of variation due to noise and V is the gradation of the image data.
The coefficients a, b, and c in the above equation (14) are used as noise characteristic parameters.

  The actual size conversion parameter is a parameter for setting an observation viewing angle that determines the MTF characteristics together with the observation viewing distance parameter included in the observation condition data D12. The actual size conversion parameter is a parameter used to convert the size of an image captured by the image input device 20 into an actual size on the paper P. For example, the actual size conversion parameter is captured by one pixel of the imaging device 22 provided in the image input device 20. A value indicating the size on the paper P that can be used is used. This actual size conversion parameter is used in order to eliminate the influence of differences in installation conditions and instrumental differences of the image input apparatus 20.

  The observation viewing distance parameter included in the observation condition data D12 is a parameter that defines the distance (viewing distance) for observing the paper P as the inspection object, and can be defined as an arbitrary value. This observation viewing distance parameter is for setting an observation condition in which a human observes the paper P as an inspection object from the viewing distance. Here, since the viewing angle changes when the viewing distance changes, the degree of conspicuousness that the human feels also changes. In order to maintain the objectivity of the calculated ground index (degree of conspicuity), the observation viewing distance parameter is fixed for each type of inspection object, for each paper production line, for each office, or for each company. It is desirable to be fixed with.

  The observation luminance parameter is a parameter that defines the luminance for observing the paper P as the inspection object, and an arbitrary value can be defined. This observation luminance parameter is used to calculate the degree of conspicuousness when observing the paper P in a state where the transmitted light amount is adjusted to a predetermined reference light amount (reference luminance) regardless of the type of the paper P or the type of apparatus. belongs to. Here, when there is a difference in transmittance depending on the type of the paper P, or when there is a difference in the amount of light depending on the apparatus, the degree of conspicuousness that the human feels also changes. For this reason, in order to maintain the objectivity of the calculated formation index (degree of conspicuity), the observation luminance parameter is fixed for each type of inspection object, for each paper production line, for each office, or for each company. It is desirable to be fixed within the industry.

  The measurement area setting unit 41 reads the image data D1 stored in the storage unit 33, and sets an area for calculating the formation index DX for the image indicated by the read image data D1. Specifically, an area where the paper P as the inspection object is photographed is extracted from the image indicated by the image data D1 read from the storage unit 33, and the formation index DX is calculated for the extracted area. An area (hereinafter referred to as an inspection area) is set.

  The image correction unit 42 removes noise superimposed on the image data in which the inspection region is set by the measurement region setting unit 41. For example, high-frequency noise superimposed on the image data is removed, illuminance unevenness caused by the optical characteristics of the illumination device 10 and the image input device 20 and low-frequency shading caused by the opening characteristics of the image input device 20 are removed. The corrected image data from which noise has been removed is output.

Specifically, the image correction unit 42 removes high-frequency noise caused by light shot noise, dark current, and the like of the image input device 20 from the noise characteristic parameters included in the image input condition data D11 output from the inspection control unit 35. Noise corrected image data I _NC is generated. For example, the average image data I _OA is obtained by performing a moving average process with a size of 3 × 3 pixels on the data in the inspection area (hereinafter referred to as original image data I _O ) in the image data D1, and the above-mentioned (14) for generating a high-frequency noise amount image I _N indicating a noise amount for each pixel of the average image data I _OA using equation. Then, a pixel-by-pixel difference between the original image data I _O and the average image data I _OA is obtained, and when the difference value is smaller than the high-frequency noise amount image I _N , the difference is regarded as noise and the average image data the value of I _OA is a true value, the noise correction image by when the larger high-frequency noise amount image I _N is the value obtained by removing the high frequency noise amount image I _N from the original image data I _O the true value Data I _NC is generated.

Then, the image correcting unit 42, with respect to the noise corrected image data I _NC above, the moving average processing or FFT operation, the shading image data I _S by an optical simulation or the like, the value V _B of the reference area of the shading image data I _S acquired to generate the shading correction data I _C by using an arithmetic expression shown below, for example of formula (15). Here, the value V _B of the reference region, for example, the image center of the value of the shading image data I _S, or using the average value or the maximum value of the area including the image center. It said shading correction data I _C is output from the image correcting unit 42 as the corrected image data.

  The formation index calculation unit 43 includes a luminance conversion unit 51 (conversion unit), an observation luminance conversion unit 52 (adjustment unit), a local size calculation unit 53, a local region setting unit 54 (setting unit), and an FFT calculation unit 55 (first unit). A calculation unit), an MTF calculation unit 56, a local formation index calculation unit 57 (second calculation unit), a formation distribution image generation unit 58, and a formation distribution statistical calculation unit 59 (statistic calculation unit). The formation index calculation unit 43 having such a configuration uses the image input condition data D11 and the observation condition data D12 output from the inspection control unit 35 to calculate the formation index DX from the corrected image data output from the image correction unit 42. calculate.

The luminance conversion unit 51 uses the luminance conversion characteristic parameter and the exposure time parameter included in the image input condition data D11, and uses the corrected image data from the image correction unit 42 as a physical value as a luminance value (for example, the unit is [cd / m ² ] is converted into luminance image data. Specifically, the luminance conversion characteristic parameters (α, β, γ, T ₀ ) and the exposure time parameter (T) are respectively substituted into the equation (13), and the corrected image data I _C output from the image correction unit 42 is used. by obtaining the input luminance L of each pixel into the variable V in the equation (13) (gray scale value of each pixel), it converts the corrected image data I _C on the luminance image data I _L.

Observation luminance conversion unit 52 uses the observation luminance parameter included in the observation condition data D12, and converts the luminance image data I _L from the luminance conversion unit 51 to the observation luminance image data I _LS. Here, the observation luminance image data I _LS is data indicating luminance when the paper P is observed in a state where the transmitted light amount is adjusted to a predetermined reference light amount. The reason for performing the conversion process is to realize an objective inspection of the formation of the paper P regardless of the type of the paper P or the type of the apparatus.

Specifically, the observation luminance conversion unit 52 calculates the average luminance L _A of the luminance image data I _L from the luminance conversion unit 51 and is included in the luminance image data I _L , the average luminance L _A , and the observation condition data D12. the observation brightness parameter L _V, obtains the observation luminance image data I _LS by substituting, for example, in the following equation (16).

The local size calculation unit 53 indicates the size of the local area to be set in the examination area from the actual size conversion parameter included in the image input condition data D11 and the observation viewing distance parameter included in the observation condition data D12. R _h and R _v and local region viewing angles d _h and d _v indicating the viewing angle of the local region are calculated. Here, when a person visually inspects the paper P, the paper P is observed with a limited visual field of, for example, a viewing angle of about 2 °, and the entire paper P is inspected while moving the gazing point. The local area is an area that is set to realize an inspection similar to an inspection performed by a human, and corresponds to an area that is observed in a limited field of view of the human.

Specifically, the local size calculation unit 53 uses the actual size conversion parameter included in the image input condition data D11 and the observation viewing distance parameter included in the observation condition data D12, for example, to calculate the number of pixels corresponding to a viewing angle of 2 ° locally. The area sizes are calculated as R _h and R _v . Here, when the local region sizes R _h and R _v are expressed by the number of pixels, the viewing angle of the local region having the local region sizes R _h and R _v is almost shifted from 2 °. For this reason, the local size calculation unit 53 uses the actual size conversion parameter included in the image input condition data D11 and the observation viewing distance parameter included in the observation condition data D12 to represent the local region size R _h , expressed by the number of pixels. The viewing angle of the local region having the magnitude of R _v is calculated as the local region viewing angles d _h and d _v .

The local region setting unit 54 sets a local region in the examination region using the local region sizes R _h and R _v calculated by the local size calculation unit 53. Specifically, for example, the size in the horizontal direction is the local region size _Rh , the size in the vertical direction is the local region size _Rv , and is arranged in a matrix so that the ends of the adjacent local regions overlap. A plurality of rectangular local areas are set in the inspection area.

The FFT calculator 55 calculates a contrast distribution image C _i (first data) indicating the contrast for each spatial frequency. Specifically, from the observation luminance image data I _LS converted by the observation luminance conversion unit 52, data in one local region set by the local region setting unit 54 (hereinafter referred to as local region observation luminance image data I _LSi ). ) _Is extracted, a two-dimensional FFT operation is performed on the local region observation luminance image data _ILSi to obtain a power spectrum image, and the normalized power is obtained by dividing by the size of the local region (R _h × R _v ). Obtain a spectral image. Then, the contrast distribution image C _i is calculated by dividing the normalized power spectrum image by the value of the center point.

Here, the value of the center point of the normalized power spectrum image represents the average luminance L _Ai of the local region from which the local region observation luminance image data I _LSi is extracted. Therefore, the contrast distribution image can be obtained by dividing the normalized power spectrum image by the value of the center point. In general, in the FFT operation, a component that does not actually exist in the power spectrum generated due to the discontinuity of the signal appears. For this reason, it is desirable to correct using, for example, a window function called an Akaike window.

The MTF calculation unit 56 uses the local region viewing angles d _h and d _v set by the local size calculation unit 53 and the average luminance L _Ai used by the FFT calculation unit 55 to generate an MTF image M _2i (second data). Generate. This MTF image M _2i shows the contrast sensitivity for each spatial frequency when the paper P is observed under the observation conditions defined by the observation viewing distance parameter and the observation luminance parameter included in the observation condition data D12.

Specifically, the MTF calculator 56 divides the spatial frequencies f _h and f _v used in the contrast distribution image calculated by the FFT calculator 55 by the local region viewing angles d _h and d _v , respectively. Is converted into spatial frequencies u _h and u _v with [cycle number / viewing angle]. Then, the MTF image M _2i is generated by substituting the spatial frequencies u _h , u _v , the average luminance L _Ai used in the FFT calculation unit 55, and the local region viewing angles d _h , d _v into, for example, the equation (11). To do.

The local formation index calculation unit 57 calculates the local formation index J _i using the contrast distribution image C _i obtained by the FFT calculation unit 55 and the MTF image M _2i obtained by the MTF calculation unit 56. Specifically, as shown in the equation (12), the local distribution is obtained by multiplying the contrast distribution image C _i and the MTF image M _2i for each spatial frequency and adding the obtained values. The index J _i is calculated.

The formation distribution image generation unit 58 generates a formation distribution image indicating a two-dimensional distribution of the local formation index J _i from the local formation index J _i calculated by the local formation index calculation unit 57. That is, a plurality of local regions are set in the examination region by the local region setting unit 54, and the local formation index for each of the local regions is calculated by the FFT operation unit 55, the MTF operation unit 56, and the local formation index operation unit 57. Since J _i is obtained, the formation distribution image generation unit 58 generates a formation distribution image indicating a two-dimensional distribution of the local formation index J _i in the examination region.

The formation distribution statistical calculation unit 59 performs statistical calculation on the formation distribution image indicating the two-dimensional distribution of the local formation index J _i generated by the formation distribution image generation unit 58 to obtain the formation index DX. Ask. Specifically, the formation distribution statistical calculation unit 59 performs an averaging process, a maximum value calculation process, a standard deviation calculation process, and the like on the formation distribution image indicating the two-dimensional distribution of the local formation index J _i , for example. The value obtained by these processes is obtained as the formation index DX. When only one local region is set in the examination region by the local region setting unit 54, the local formation index J _i calculated by the local formation index calculation unit 57 becomes the formation index DX.

<Operation of the ground inspection device>
Next, the operation of the ground inspection apparatus having the above configuration will be described. When the inspection is started, the illumination control unit 31 is controlled by the inspection control unit 35, and the illumination light is irradiated from the illumination device 10 to the back surface of the paper P as the inspection object. A part of the illumination light applied to the paper P is transmitted through the paper P, and the rest is reflected or scattered on the back surface.

  The surface of the paper P is photographed by the imaging device 22 included in the image input device 20 at a predetermined timing while the illumination light is irradiated on the paper P (while there is a part of light that passes through the paper P). Image data of the image is output from the image input device 20. The image data output from the image input device 20 is captured by the image capturing unit 32 and stored in the storage unit 33 under the control of the inspection control unit 35.

  When the image data is stored in the storage unit 33, the image input condition data D11 and the observation condition data D12 are output from the inspection control unit 35 to the image processing unit 34. Under the control of the inspection control unit 35, the storage unit The image data D1 stored in 33 is read by the image processing unit 34, and the process of calculating the formation index DX is started. Here, the image input condition data D11 and the observation condition data D12 are output from the inspection control unit 35 to the image processing unit 34 after the image data is stored in the storage unit 33. The output of the image input condition data D11 and the observation condition data D12 to the processing unit 34 may be performed in advance before starting the inspection.

When the processing of the image processing unit 34 is started, first, an area where the paper P is photographed is extracted from the image indicated by the image data D1 read from the storage unit 33, and an inspection area is extracted from the extracted area. Is set by the measurement area setting unit 41. Next, a process of removing noise superimposed on the image data in which the inspection area is set is performed by the image correction unit 42, and the shading correction data I generated using the arithmetic expression shown in the above-described expression (15). _C is output to the formation index calculation unit 43 as corrected image data.

When the corrected image data is input to the formation index calculation unit 43, first, using the luminance conversion characteristic parameter and the exposure time parameter included in the image input condition data D11, the corrected image data is converted into a luminance value (for example, unit). There process of converting the luminance image data I _L indicating a [cd / m ^2] brightness value is) is performed in the luminance conversion unit 51. Then, using the observed brightness parameter included in the observation condition data D12, the process of converting the luminance image data I _L from the luminance conversion unit 51 to the observation luminance image data I _LS is performed in the observation luminance conversion unit 52. Thereby, since the observation luminance image data _ILS indicating the luminance when the paper P is observed in a state where the transmitted light amount is adjusted to a predetermined reference light amount is obtained, regardless of the type of the paper P or the type of the apparatus. In addition, an objective inspection of the formation of the paper P is realized.

When the observation luminance image data I _LS is obtained, the size of the local region to be set in the examination region is determined from the actual size conversion parameter included in the image input condition data D11 and the observation viewing distance parameter included in the observation condition data D12. The local size calculation unit 53 performs processing for calculating the local region sizes R _h and R _v to be shown and the local region viewing angles d _h and d _v showing the viewing angle of the local region. Then, the local region setting unit 54 performs processing for setting a local region in the examination region using the local region sizes R _h and R _v calculated by the local size calculating unit 53.

When the above process is completed, the FFT calculation unit 55 performs a process of calculating a contrast distribution image C _i indicating the contrast for each spatial frequency with respect to the local area set in the examination area (first step). In addition, the MTF calculation unit 56 performs processing for generating the MTF image M _2i using the local region viewing angles d _h and d _v set by the local size calculation unit 53 and the average luminance L _Ai used by the FFT calculation unit 55. Performed (second step).

Then, the process of calculating the local formation index J _i using the contrast distribution image C _i obtained by the FFT operation unit 55 and the MTF image M _2i obtained by the MTF operation unit 56 is a local formation index operation unit. 57 (third step). Specifically, as shown in the above-described equation (12), the contrast distribution image C _i obtained by the FFT computation unit 55 and the MTF image M _2i obtained by the MTF computation unit 56 are each spatial frequency. The local formation index J _i is calculated by multiplying and adding the values obtained thereby.

The above processing of the FFT calculation unit 55, the MTF calculation unit 56, and the local formation index calculation unit 57 is performed for each of the local regions set in the examination region. Here, the processing of the FFT calculation unit 55, the MTF calculation unit 56, and the local formation index calculation unit 57 may be sequentially performed for each local region, or may be performed in parallel for a plurality of local regions. Also good. By performing processing on a plurality of local regions in parallel, the time required to calculate the local formation index J _i can be shortened.

When the local formation index J _i is calculated for all the local areas set in the examination area, the formation distribution image generation unit 58 generates a formation distribution image indicating the two-dimensional distribution of the local formation index J _i. Generated. Then, the formation distribution statistical calculation unit 59 performs an averaging process, a maximum value calculation process, a standard deviation calculation process, and the like on the formation distribution image indicating the two-dimensional distribution of the local formation index J _i . The value obtained by the processing is obtained as the formation index DX.

  The formation index DX calculated by the above processing is output from the image processing unit 34 to the inspection control unit 35, and the quality of the formation is inspected based on the formation index DX. For example, when the value of the formation index DX is 1 [JND] or more and 50 [JND] or less, the inspection result is “good”, and the formation index DX is larger than 50 [JND]. In some cases, the inspection result is “bad”. The formation index DX thus obtained and the inspection result of the inspection control unit 35 are output to the inspection result output unit 36.

  FIG. 7 is a diagram showing a formation distribution image and a formation index showing a two-dimensional distribution of local formation indices obtained from the image shown in FIG. 28 (images having the same uniformity and average luminance and different granularities). is there. FIG. 8 shows a formation distribution image and a formation index showing a two-dimensional distribution of local formation indexes obtained from the image shown in FIG. 29 (images having the same uniformity and granularity and different average luminance). FIG. Note that the formation distribution images and formation indexes shown in FIGS. 7 and 8 are obtained with the distance of 10 times the image height as the observation viewing distance with respect to the images shown in FIGS. In addition, the formation distribution images shown in FIGS. 7 and 8 are illustrated so as to become brighter as the value of the formation index increases.

  28, the image shown in (c) is the best, the image shown in (a) is good, and the image shown in (b) can be recognized visually by humans. It was a thing. Referring to FIG. 7, the bright portion has the smallest formation distribution image shown in (c), then the smallest formation distribution image shown in (a), and the largest formation formation image shown in (b). In addition, the formation index shown in (c) is the smallest, the one shown in (a) is the smallest, and the one shown in (b) is the largest.

  29, the image shown in (a) is the best, the image shown in (b) is good, and the image shown in (c) is the worst. It was something that could be done. Referring to FIG. 8, the bright part has the smallest formation distribution image shown in (a), then the smallest formation distribution image shown in (b), and the largest formation distribution image shown in (c). In addition, the formation index shown in (a) is the smallest, the one shown in (b) is small, and the one shown in (c) is the largest.

  As described above, the brightness and darkness of the formation distribution image for both images having the same uniformity and average luminance and different granularity and images having the same uniformity and granularity and different average luminance. The magnitude of the index value is consistent with the way people feel visually. This means that the inspection result when the paper is inspected by the ground inspection apparatus 1 of the present embodiment matches the inspection result when the human (skilled inspector) visually inspects.

Here, in the above-described embodiment, the local formation index calculation unit 57 provided in the image processing unit 34 performs the contrast distribution image C _i obtained by the FFT calculation unit 55 and the MTF image obtained by the MTF calculation unit 56. An example has been described in which the local formation index J _i is calculated by multiplying M _2i for each spatial frequency and adding the values obtained thereby. However, as shown in the following equation (17), the local formation index calculation unit 57 may perform a weighting calculation to calculate the local formation index J _i .

Specifically, the value obtained by multiplying the contrast distribution image C _i and the MTF image M _2i for each spatial frequency is calculated and added to the power (w is an arbitrary real number). The 1 / w power value of the value may be obtained and used as the local formation index J _i . By performing such weighting calculation, the accuracy of the formation index can be further improved.
In the above equation (17), if the value of w is set to “1”, the same processing as the processing described in the above-described embodiment (processing for calculating the local formation index J _i without performing weighting calculation). Is done.

  FIG. 9 is a diagram for explaining a difference in effect between the case where the weighting calculation is not performed and the case where the weighting is performed when the local formation index is calculated. FIG. 9A illustrates the case where the weighting calculation is not performed. (B) is a figure which shows the effect at the time of performing a weighting calculation. Moreover, FIG. 10 is a figure which shows an example of the formation pattern used in order to demonstrate the difference of the effect shown in FIG. The formation pattern shown in FIG. 10A is a fringe pattern (hereinafter referred to as a sine wave pattern) whose brightness changes in a sine wave shape in the horizontal direction, and the formation pattern shown in FIG. 10B has a circular shape. Is a block pattern (hereinafter referred to as block pattern).

  9A and 9B, the horizontal axis represents the human visual inspection result (experimental value), and the vertical axis represents the inspection result (predicted value) of the ground inspection apparatus 1. However, since it is difficult to directly obtain the formation index from human visual inspection, the sine wave pattern shown in FIG. 10A is visually observed with different frequencies (number of horizontal light and dark) and background luminance. Or an experiment for obtaining the JND contrast by visually observing the block pattern shown in FIG. 10B with different sizes and background luminances, and obtaining the JND contrast [%] obtained by this experiment. Is the “experimental value” on the horizontal axis.

  In addition, when the same experiment as the above experiment is performed with the ground inspection apparatus 1, the calculated ground index value is almost “1” and the difference is small. It is difficult to evaluate. For this reason, in FIGS. 9A and 9B, the contrast C (contrast of the sinusoidal pattern shown in FIG. 10A or the block pattern shown in FIG. 10B) is obtained from the above-described equation (12). The JND contrast [%] obtained by dividing by the formation index J (the formation index calculated without performing the weighting operation or the formation index calculated by performing the weighting operation) is expressed as “predicted value”. "

Referring to FIG. 9 (a), when observing the sinusoidal pattern shown in FIG. 10 (a), the predicted and experimental values correlation coefficient of the approximate straight line Q1 R ² becomes "0.94" It can be seen that there is a high correlation between the two. However, in addition to the case of observing the sinusoidal pattern shown in FIG. 10 (a), when including the case of observing the bulk pattern shown in FIG. 10 (b), the correlation coefficient R ² of the approximate straight line Q2 " 0.40 "indicating that there is no high correlation between the experimental value and the predicted value.

On the other hand, referring to FIG. 9B, in addition to the case where the sinusoidal pattern shown in FIG. 10A is observed, the case where the block pattern shown in FIG. the correlation coefficient R ² is "0.94" and the straight line Q3, irrespective of the difference in shape of the observation pattern, there can be seen a high correlation between predicted and experimental values. The result shown in FIG. 9B is obtained when the value of w in the above-described equation (17) is set to “1.3”. Thus, the accuracy of the formation index can be further improved by performing the weighting calculation to obtain the formation index.

  As described above, in the present embodiment, the contrast distribution image calculated by the FFT calculation unit 55 and the MTF image obtained by the MTF calculation unit 56 are multiplied for each spatial frequency, and the values obtained thereby are added. Thus, the formation index DX reflecting the degree of conspicuousness of the formation when the human actually inspects the paper P is calculated. For this reason, the difference in formation according to the granularity and lightness similar to human vision can be inspected. As a result, the degree of formation can be communicated to the other party only by notifying the formation index by, for example, a document or a telephone.

  In addition, when calculating the formation index DX, the transmitted light amount is adjusted to a predetermined reference light amount (reference luminance) regardless of the type of paper P and the type of apparatus, so that the difference in inspection conditions, etc. Regardless, it becomes possible to inspect objectively using the formation index. This makes it possible to calculate a stable formation index regardless of the inspection conditions and to compare the formations of the papers P manufactured on different production lines.

[Second Embodiment]
FIG. 11 is a block diagram showing the configuration of the ground inspection apparatus according to the second embodiment of the present invention. As shown in FIG. 11, the ground inspection device 2 of the present embodiment includes the illumination device 10, the image input device 20, and the control processing device 30, similarly to the ground inspection device 1 shown in FIG. 1, and is transported. Image processing is performed on image data obtained by photographing the paper P (inspection object) conveyed in the direction Z1, and the formation of the paper P is inspected based on the formation index obtained by this image processing. .

  However, in the ground inspection apparatus 1 shown in FIG. 1, the image input device 20 is arranged on the opposite side of the lighting device 10 with the paper P interposed therebetween, and an image (transmitted light image) by light transmitted through the paper P is used. It was a transmitted light type ground inspection device for inspecting the paper P. On the other hand, in the ground inspection device 2 of the present embodiment, the image input device 20 is arranged on the same side as the lighting device 10 with respect to the paper P, and an image by light reflected or scattered on the surface of the paper P This is a reflected light type ground inspection device for inspecting the ground of paper P from (reflected light image).

  The ground inspection device 2 according to the present embodiment is different from the ground inspection device 1 shown in FIG. 1 only in that the formation of the paper P is inspected from the reflected light image. The operation is the same as that of the ground inspection apparatus 1 shown in FIG. For this reason, the detailed description of a structure and operation | movement of the ground inspection apparatus 2 is abbreviate | omitted. The ground inspection device 2 of the present embodiment also calculates the ground index DX that reflects the degree of conspicuousness of the ground when the human actually inspects the paper P. The effect of being able to inspect the difference in formation according to the lightness is obtained.

[Third Embodiment]
FIG. 12 is a block diagram showing the configuration of the ground inspection apparatus according to the third embodiment of the present invention. As shown in FIG. 12, the ground inspection device 3 of the present embodiment includes an illumination device 10 a and an image input device 20 a arranged on the front surface side of the paper P, an illumination device 10 b and an image input device arranged on the back surface side of the paper P. The apparatus 20b and the control processing device 30 are provided, and image processing is performed on each of the image data obtained by photographing both sides of the paper P (inspection object) conveyed in the conveyance direction Z1, and obtained by this image processing. The formation of the paper P can be inspected simultaneously on both sides based on the formed formation index.

  The illuminating devices 10a and 10b respectively include a light source 11a and an optical element 12a, a light source 11b, and an optical element 12b similar to the light source 11 and the optical element 12 included in the illuminating device 10 illustrated in FIG. The image input devices 20a and 20b respectively include an optical element 21a and an imaging device 22a, an optical element 21b, and an imaging device 22b that are the same as the optical element 21 and the imaging device 22 included in the image input device 20 illustrated in FIG.

  The control processing device 30 includes image capturing units 32a and 32b, storage units 33a and 33b, image processing units 34a and 34b, an inspection result output unit 36, an illumination control unit 61, and an inspection control unit 62. In addition to performing illumination light irradiation control on the image data, image processing is performed on image data output from the image input device 20. Then, the formation of the paper P is inspected using the formation index obtained by image processing, and the inspection result is output.

  The image capturing unit 32a, the storage unit 33a, and the image processing unit 34a are the same as the image capturing unit 32, the storage unit 33, and the image processing unit 34 shown in FIG. Below, image data output from the image input device 20a is captured, stored, and processed. The image capturing unit 32b, the storage unit 33b, and the image processing unit 34b are the same as the image capturing unit 32, the storage unit 33, and the image processing unit 34 shown in FIG. Below, image data output from the image input device 20b is captured, stored, and processed.

  Under the control of the inspection control unit 62, the illumination control unit 61 performs control of the amount of illumination light applied to each of the front and back surfaces of the paper P, control of irradiation timing (for example, timing control of strobe light emission), and the like. . The inspection control unit 62 controls the illumination control unit 61 to control the amount of illumination light applied to each of the front and back surfaces of the paper P, and controls the image capture units 32a and 32b to control image data capture. The image processing units 34a and 34b are controlled to perform image processing execution control. When the formation indexes are calculated by the image processing units 34a and 34b, the inspection control unit 62 outputs the formation indexes and the inspection results to the inspection result output unit 36.

  The ground inspection device 3 according to the present embodiment inspects the texture of the surface of the paper P from the reflected light image on the surface of the paper P, and simultaneously inspects the texture of the back surface of the paper P from the reflected light image of the back surface of the paper P. (Simultaneous inspection on both sides) is possible, but the processing performed by the control processing device 30 on the image data indicating each reflected light image is the same as that of the ground inspection device 1 shown in FIG. . For this reason, the detailed description of a structure and operation | movement of the ground inspection apparatus 3 is abbreviate | omitted.

  The ground inspection device 3 of the present embodiment also calculates the ground index DX reflecting the degree of conspicuousness of the ground when the human actually inspects the paper P. The effect of being able to inspect the difference in formation according to the lightness is obtained. In addition, since the formation inspection device 3 according to the present embodiment can separate the formation on the front surface and the formation on the back surface of the paper P, for example, the distribution of the formation on the front surface and the back surface and the trend in the transport direction Z1 are analyzed. By doing so, management of the front side and the back side of the paper P can be performed individually.

[Fourth Embodiment]
FIG. 13 is a block diagram showing the configuration of the ground inspection device according to the fourth embodiment of the present invention. As shown in FIG. 13, the ground inspection device 4 according to the present embodiment includes illumination devices 10 a and 10 b arranged so as to sandwich the paper P, an image input device 20 arranged on the front side of the paper P, and a control processing device. 30. The image input device 20 may be disposed on the front surface side of the paper P or may be disposed on the back surface side of the paper.

  That is, the ground inspection device 4 according to the present embodiment has a combination of the ground inspection device 1 according to the first embodiment and the ground inspection device 2 according to the second embodiment, and is transported in the transport direction Z1. Inspection of the transmission ground and reflection ground of the paper P (inspection object) to be performed is possible. However, although the illumination control part 31 provided in the ground inspection apparatuses 1 and 2 by 1st, 2nd embodiment controls one lighting apparatus 10, the ground inspection apparatus 4 of this embodiment is used. The illumination control unit 31 provided in is different from the illumination control unit 61 provided in the ground inspection device 3 of the third embodiment in that the two illumination devices 10a and 10b are controlled.

  Since the ground inspection apparatus 4 of this embodiment arrange | positions the illuminating devices 10a and 10b so that the paper P may be pinched | interposed, by changing the light emission timing of the illuminating devices 10a and 10b, for example, from the image input device 20 The image data of the transmitted light image and the image data of the reflected light image are output separately. For this reason, if the same image processing as in the first and second embodiments is performed on the image data of the transmitted light image and the image data of the reflected light image, the transmission condition and the reflection condition of the paper P are inspected. can do.

  The ground inspection device 4 of the present embodiment is different from the ground inspection devices 1 and 2 of the first and second embodiments in that both the transmission ground and the reflection ground can be inspected. Specific image processing performed by the control processing device 30 is the same as that of the ground inspection devices 1 and 2. For this reason, detailed description of the image processing performed in the ground inspection apparatus 4 is omitted. The ground inspection device 4 of the present embodiment also calculates the ground index DX that reflects the degree of conspicuousness of the ground when the human actually inspects the paper P. The effect of being able to inspect the difference in formation according to the lightness is obtained.

[Fifth Embodiment]
FIG. 14 is a block diagram showing an internal configuration of an image processing unit provided in the ground inspection apparatus according to the fifth embodiment of the present invention. As shown in FIG. 14, the image processing unit 34 includes a formation degradation region extraction unit 71 and a defect determination unit 72 in addition to the measurement region setting unit 41, the image correction unit 42, and the formation index calculation unit 43. In addition to the calculation of the formation index DX, it is possible to detect defects in the paper P. The image processing unit 34 shown in FIG. 14 includes a transmitted light type ground inspection device 1 shown in FIG. 1, a reflected light type ground inspection device 2 shown in FIG. 11, and a double-sided simultaneous inspection shown in FIG. It can be applied to both the joint inspection apparatus 3 and the ground inspection apparatus 4 capable of performing the transmission method and the reflection method inspection shown in FIG.

  The formation degradation region extraction unit 71 sets a threshold value for the formation distribution image generated by the formation distribution image generation unit 58, the region where the local formation index is larger than the threshold value, and the local formation index An area that is equal to or less than the threshold is extracted by binarization processing or the like. The formation deterioration area extraction unit 71 performs labeling on the extracted area by performing a labeling process or the like.

  The defect determination unit 72 cuts out a region corresponding to each region extracted by the formation degradation region extraction unit 71 from the observation luminance conversion image generated by the observation luminance conversion unit 52, and uses it as a determination usage image. The characteristics are evaluated to determine the defect area and defect type. As an image used for the determination, a corrected image generated by the image correcting unit 42 or a luminance image generated by the luminance converting unit 51 may be used instead of the observation luminance converted image.

  Here, the defect area can be regarded as an area having a large deviation from the average value of the determination use images. For this reason, for example, the defect determination unit 72 calculates an average value of the determination use images, and determines whether the defect is a defect region by extracting a region higher or lower than the average value. The defect determination unit 72 classifies whether the extracted area is a bright defect or a dark defect depending on whether the extracted area is a high area or a low area with respect to the average value. Furthermore, the defect type is determined by determining whether the extracted region is dot-like or streak-like by filtering processing or learning.

  In the above configuration, when the image data D1 stored in the storage unit 33 is read and input to the image processing unit 34, the same processing as in the first embodiment is performed to calculate the formation index DX. Specifically, for example, the FFT calculation unit 55 performs processing for calculating a contrast distribution image for each local region, and the MTF calculation unit 56 performs processing for generating an MTF image for each local region. Then, a process of calculating a local formation index for each local region from the contrast distribution image and the MTF image is performed by the formation index calculation unit 57, and a formation distribution image showing a two-dimensional distribution of the local formation index is generated. Is performed by the formation distribution image generation unit 58.

  When the formation distribution image is generated by the formation distribution image generation unit 58, the formation distribution image is input to the formation deterioration region extraction unit 71, and binarization processing using a predetermined threshold is performed, so that the local formation index is obtained. A region larger than the threshold value and a region where the local formation index is equal to or smaller than the threshold value are extracted. Next, from the observation luminance conversion image generated by the observation luminance conversion unit 52, regions corresponding to the respective regions extracted by the formation degradation region extraction unit 71 are cut out as determination use images, and the defect determination unit 72 performs defect regions, A defect type is determined.

  As described above, in the present embodiment, a region where the local formation index is larger than the threshold using the formation distribution image showing the two-dimensional distribution of the local formation index reflecting the degree of conspicuousness, and the local formation index are An area that is equal to or less than the threshold is extracted by binarization processing or the like, and an area corresponding to the extracted area is cut out from the observation luminance conversion image as a determination use image to determine a defect area and a defect type. For this reason, an area conspicuous when viewed by a human can be detected as a defect regardless of the shape of the defect, so that an unexpected defect shape can be detected.

[Sixth Embodiment]
FIG. 15 is a block diagram showing an internal configuration of an image processing unit provided in the ground inspection apparatus according to the sixth embodiment of the present invention. As shown in FIG. 15, the image processing unit 34 includes a measurement region setting unit 41, an image correction unit 42, and a formation index calculation unit 80, and image input condition data D11 output from the inspection control unit 35 and the like. In addition to the formation index DX, the formation index DXn for each class is calculated using the observation condition data D12 and the class-specific analysis condition data D21. When the total number of classes is K, the variable n is an integer that satisfies 1 ≦ n ≦ K.

  The image processing unit 34 shown in FIG. 15 includes a transmitted light type ground inspection device 1 shown in FIG. 1, a reflected light type ground inspection device 2 shown in FIG. 11, and a double-sided simultaneous inspection shown in FIG. It can be applied to both the joint inspection apparatus 3 and the ground inspection apparatus 4 capable of performing the transmission method and the reflection method inspection shown in FIG.

  Here, in the first to fifth embodiments described above, since the formation index DX that reflects the degree of conspicuousness of the formation when the person actually inspects the paper P is calculated, It is possible to inspect the difference in formation according to the same granularity and brightness. However, the formation index DX alone cannot identify the type of formation conspicuous and the cause of its occurrence, and may be insufficient as information for improving the formation. For example, when there is an area on the paper P that is different from the surrounding area, the area can be detected by using the formation index DX, but the specific cause of the difference in formation ( This is a case where it is not possible to specify that a large number of small black spots are gathered in the region, or that a large unevenness occurs on the paper P).

Therefore, in the present embodiment, the multiplication result obtained by the local formation index computation unit 57 (contrast distribution image C _i obtained by the FFT computation unit 55 and the MTF image M _2i obtained by the MTF computation unit 56 for each spatial frequency. Multiplication result: third data) is classified into a plurality of classes, and the formation index DXn for each class is calculated. By using the formation index DX and the formation index DXn for each class, the type of how the formation is conspicuous And identification of the cause of the occurrence. The classification of the above classes, the (frequency space shown in FIG. 4 or the like which is represented by the spatial frequency in the horizontal direction u _h and vertical spatial frequencies u _v) for each frequency or each direction two-dimensional frequency space Examples include a classification method for classifying into a plurality of regions and a classification method according to weighting.

  FIG. 16 is a diagram illustrating an example of a class classification method according to the sixth embodiment of the present invention. In the example shown in FIG. 16A, the two-dimensional frequency space is classified into eight concentric ring-shaped regions (classes C1 to C8). In the two-dimensional frequency space, a direct current component is arranged at the center portion, and a component having a higher frequency is arranged as the distance from the center portion increases. For this reason, by performing the classification shown in FIG. 16A, the frequency components of the two-dimensional frequency space can be classified for each frequency.

The class-specific analysis condition data D21 is data that defines the classification of the class, and is output from the inspection control unit 35 and the like together with the image input condition data D11 and the observation condition data D12. Specifically, the class-specific analysis condition data D _< b _> 21 is data including frequency weight data W _n that defines weights for multiplication results for each spatial frequency obtained by the local formation index calculation unit 57. Specifically, when the classification shown in FIG. 16A is performed, for each class, binary frequency weight data W having a value “1” inside and a value “0” outside. _n is defined. For example, focusing on class C2, as shown in FIG. 16 (b), the inside (the part not hatched) has a value of “1” and the outside (the part shaded) has a value. Frequency weight data W _n that is “0” is defined. The frequency weight data W _n can define an arbitrary value, but is preferably set so that the sum of the values for all classes is “1”.

As a class classification method, in addition to the method of classifying into a ring-shaped region shown in FIG. 16, a method of classifying a two-dimensional frequency space into a radial region according to the direction can be considered. When such classification is performed, the angular resolution is worse in the vicinity of the central portion (the portion where the DC component is disposed) in the two-dimensional frequency space than in the peripheral portion (the portion where the high component is disposed). In such a case, it is desirable to use the frequency weight data W _n that has a small value for the central part with poor angular resolution and gradually increases as it goes to the peripheral part.

The formation index calculation unit 80 of the image processing unit 34 includes a class-specific local formation index calculation unit 81n (third calculation means) in addition to the luminance conversion unit 51 to the formation distribution statistical calculation unit 59 shown in FIGS. , Fourth calculation means), a class-specific formation distribution image generation unit 82n, and a class-specific formation distribution statistical calculation unit 83n. In FIG. 15, in order to facilitate understanding, the local formation index calculation unit 57 includes a multiplication unit 57a and an addition unit 57b. The multiplier 57a is configured to multiply the contrast distribution image C _i obtained by the FFT calculator 55 and the MTF image M _2i obtained by the MTF calculator 56 for each spatial frequency, and the adder 57b is a multiplier It shows a configuration for determining the local formation index J _i by adding the value obtained by 57a.

By class local formation index calculation unit 81n on the basis of the class-specific analysis condition data D21, and calculates the local formation index at a class-specific local formation index J _ni for each class. Specifically, the multiplication result (C _i · M _2i : third data) for each spatial frequency obtained by the multiplication unit 57a of the local formation index calculation unit 57 is defined in the analysis condition data D21 for each class. The class-specific local formation index J _ni is calculated by multiplying the frequency weight data W _n and adding the values obtained thereby.

Incidentally, by class local formation index calculation unit 81n, similarly to the local formation index calculation unit 57 may be one performing weighting operation to calculate a class-by-class local formation index J _ni. Specifically, as shown in the following equation (18), the multiplication result for each spatial frequency of the contrast distribution image C _i and the MTF image M _2i is multiplied by the frequency weight data W _n to the power of w. (Where w is an arbitrary real number) may be calculated and added, and a 1 / wth power value of the obtained value may be obtained and used as the class-specific local formation index J _ni .
By performing the above weighting calculation, the accuracy of the class-specific formation index can be further improved.

By class formation distribution image generating unit 82n on the basis of the class-specific analysis condition data D21, from the class-specific local formation CLASS locally calculated by the exponential calculation unit 81n formation index J _ni, by class topical formation index A class-specific formation distribution image indicating a two-dimensional distribution of J _ni is generated. In other words, the local formation index J _i of the plurality of local regions (local regions set in the inspection region by the local region setting unit 54) by the FFT calculation unit 55, the MTF calculation unit 56, and the local formation index calculation unit 57 is obtained. As determined for each, the class-specific formation distribution image generation unit 82n generates a formation distribution image indicating a two-dimensional distribution of the class-specific local formation index J _ni in the examination region.

Based on the class-specific analysis condition data D21, the class-specific formation distribution statistical calculation unit 83n displays the two-dimensional distribution of the class-specific local distribution index J _ni generated by the class-specific formation distribution image generation unit 82n. A statistical calculation is performed on the distribution image to obtain a class-specific formation index DXn. Specifically, by class formation distribution statistic calculation unit 83n, for example an averaging process with respect to each class local formation index J _ni of formation distribution image showing a two-dimensional distribution, the maximum value calculation process, the standard deviation calculation process Etc., and the values obtained by these processes are obtained as class-specific formation indexes DXn. When only one local region is set in the examination region by the local region setting unit 54, the class-specific local formation index J _ni calculated by the class-specific local formation index calculation unit 81n is used as the class-specific region. The combined index DXn is obtained.

In the ground inspection apparatus having the above configuration, the luminance conversion unit 51 to the ground distribution statistical calculation unit 59 provided in the measurement region setting unit 41, the image correction unit 42, and the ground index calculation unit 80 of the image processing unit 34 are used. The same process as that of the first embodiment is performed to calculate the formation index DX. In parallel with this processing, the multiplication result for each spatial frequency obtained by the multiplication unit 57a of the local formation index calculation unit 57 is output to the class-specific local formation index calculation unit 81n, and the class-specific local formation index J _ni. Is calculated.

Specifically, the multiplication result for each spatial frequency obtained by the multiplication unit 57a of the local formation index calculation unit 57 is multiplied for each class by the frequency weight data W _n defined in the class-specific analysis condition data D21. Then, the value obtained thereby is added for each class, whereby the class-specific local formation index J _ni is calculated. Note that the processing by the class-specific local formation index calculation unit 81n is performed on each of the local regions in the examination region set by the local region setting unit 54 in the same manner as the processing by the local formation index calculation unit 57. .

When the class-specific local formation index J _ni is calculated for all the local areas set in the examination area, the class-specific local distribution index generation unit 82n generates a two-dimensional distribution of the class-specific local formation index J _ni. The formation distribution images shown are respectively generated. Then, the class-specific texture distribution statistic calculation unit 83n, an averaging process with respect to formation distribution image showing a two-dimensional distribution of each class local formation index J _ni, the maximum value calculation process, the standard deviation calculation processing and the like, respectively The value obtained by these processes is obtained as the class-specific formation index DXn.

  FIG. 17 is a diagram showing an example of an image observed in the sixth embodiment of the present invention. 18 is a diagram showing the formation distribution image obtained from the image shown in FIG. 17, and FIG. 19 is a diagram showing the class-specific formation distribution image obtained from the image shown in FIG. In addition, Fig.19 (a)-(h) has each shown the classification | category formation distribution image about class C1-C8.

  The image shown in FIG. 17 is an image in which a large unevenness exists in the upper left portion E1, and a small black spot exists in the lower right portion E2. When the image shown in FIG. 17 is input to the formation inspection apparatus of this embodiment, the formation distribution image generation unit 58 obtains the formation distribution image shown in FIG. Referring to FIG. 18, the upper left portion E11 corresponding to the upper left portion E1 where large unevenness exists in FIG. 17 and the lower right portion E12 corresponding to the lower right portion E2 where small black dots exist in FIG. White circles indicating that both formation indices are high appear, and it is difficult to distinguish them.

  On the other hand, referring to FIG. 19, in the class-specific formation distribution images of classes C1 and C2 shown in FIGS. 19A and 19B, a clear white circle appears in the lower right portion E12. It is difficult to say that a clear white circle appears in the upper left portion E11. On the contrary, in the class-specific formation distribution images of classes C6 to C8 shown in FIGS. 19 (f) to (h), a clear white circle appears in the upper left portion E11, but in the upper right portion E12. It is hard to say that a clear white circle appears. In addition, although illustration is abbreviate | omitted in FIG. 19, the classification | category formation index DX1-DX8 from which a value mutually differs is calculated | required from these classification | category formation distribution images. In this way, even if it is difficult to identify the type of formation conspicuousness and the cause of its occurrence with the formation index DX alone, the type of conspicuousness is calculated by calculating and evaluating the formation index DXn for each class. And the cause of the occurrence can be easily identified.

  FIG. 20 is a diagram showing a plurality of images observed in the sixth embodiment of the present invention. FIG. 21 is a diagram showing the class-specific formation index obtained from each of the images shown in FIG. 20, wherein (a) shows the class-specific formation index DXn itself, and (b) shows the class. It is a figure which shows the value which divided another formation index DXn by the formation index DX. In FIG. 21, since the classification similar to the classification of the class shown in FIG. 16 is performed, the frequency component becomes higher toward the class C1, and the frequency component becomes lower toward the class C9. In addition, legends “a” to “g” in FIGS. 21A and 21B indicate class-specific formation indexes obtained from the images shown in FIGS. 20A to 20G, respectively. Yes.

  Referring to FIG. 20, FIGS. 20 (a) and 20 (b) are images obtained when a granular formation (fine floc) is observed, and FIGS. 20 (c), (e) and (f). Is an image obtained when a cloud-like formation (large flock) is observed. Referring to FIG. 21 (a), when the image of FIG. 20 (f) having the largest cloud-like formation among the images shown in FIGS. 20 (a) to (g) is observed, FIG. It can be seen that a class-by-class formation index DXn is obtained that is far greater than the case of observing other images other than the image of f).

  In addition, referring to FIG. 21B, when a granular formation is observed (for example, when the images shown in FIGS. 20A and 20B are observed), a high frequency class (for example, a class) is used. The value of C1 and class C2) tends to increase. On the other hand, when a cloud-like formation is observed (for example, when images shown in FIGS. 20C, 20E, and 20F are observed), a low-frequency class (for example, class C5). And the value of class C6) tends to increase. For this reason, the formation shape (granularity) of the inspection object can be intuitively determined by referring to the formation index DXn by class.

[Seventh Embodiment]
FIG. 22 is a block diagram showing an internal configuration of an image processing unit provided in the ground inspection apparatus according to the seventh embodiment of the present invention. The image processing unit 34 shown in FIG. 22 is similar to the image processing unit 34 shown in FIG. 15 in that the transmitted light type ground inspection device 1 shown in FIG. 1 and the reflected light type ground inspection device 2 shown in FIG. 12 can be applied to both the ground inspection apparatus 3 capable of performing the double-sided simultaneous inspection illustrated in FIG. 12 and the ground inspection apparatus 4 capable of performing the transmission method and the reflection method illustrated in FIG.

  As illustrated in FIG. 22, the image processing unit 34 includes a measurement region setting unit 41, an image correction unit 42, and a formation index calculation unit 80, similar to the image processing unit 34 illustrated in FIG. 15. However, in the image processing unit 34 shown in FIG. 1, the multiplication result of the multiplication unit 57a is output to the addition unit 57b and the class-specific local formation index calculation unit 81n, and the calculation result of the class-specific local formation index calculation unit 81n In contrast to being output only to the class-specific formation distribution image generation unit 82n, in the image processing unit 34 shown in FIG. 22, the multiplication result of the multiplication unit 57a is output only to the class-specific local formation index calculation unit 81n. The difference is that the calculation result of the class-specific local formation index calculation unit 81n is output to the class-specific formation distribution image generation unit 82n and the addition unit 57b.

The difference in the above configuration is to shorten the time required for calculating the formation index DX and the class-specific formation index DXn as compared with the image processing unit 34 shown in FIG. That is, in the image processing unit 34 shown in FIG. 15, the multiplication result of the multiplication unit 57a is added by the addition unit 57b to obtain the local formation index J _i and the multiplication result of the multiplication unit 57a is classified for each class. The class-specific local formation index J _ni was obtained by adding each time. Here, as described with reference to FIG. 16, the frequency weight data W _{n that} is classified into the classes of the regions where the two-dimensional frequency spaces do not overlap and is defined for each class is a binary value (for example, the value “ In the case of taking “0” and the value “1”), it is considered that the addition processing performed in the addition unit 57b and the addition processing performed in the class-specific local formation index calculation unit 81n overlap. In the present embodiment, the time required for calculating the formation index DX and the class-specific formation index DXn is reduced by omitting such overlapping addition processing.

In the above configuration, the multiplication result for each spatial frequency obtained by the multiplication unit 57a of the local formation index calculation unit 57 is output to the class-specific local formation index calculation unit 81n and is defined in the class-specific analysis condition data D21. frequency weight data W _n is multiplied for each class, values obtained by this class-specific local formation index J _ni is calculated by being added for each class. When this class-specific local formation index J _ni is output to the class-specific formation distribution image generation unit 82n, the class-specific formation distribution image generation unit 82n and the class-specific formation distribution image calculation unit 83n are the same in the sixth embodiment. The same processing as described above is performed to obtain the class-specific formation index DXn.

Further, the class-specific local formation index J _ni calculated by the class-specific local formation index calculation unit 81 n is output to the addition unit 57 b of the local formation index calculation unit 57. Then, the adding unit 57b performs a process of adding the class-specific local formation indexes J _ni of all classes, thereby calculating the formation indexes J _i . As described above, in this embodiment, the class-specific local index J _ni calculated by the class-specific local pattern index calculation unit 81n is added to obtain the group index J _i , and duplicate addition processing (multiplication) it is possible to omit the process) for obtaining the formation index J _i a multiplication result for each spatial frequency determined by the parts 57a all adding to, to calculate the formation index DX and by class formation index DXn The time required can be shortened.

[Eighth Embodiment]
FIG. 23 is a block diagram showing an internal configuration of an image processing unit provided in the ground inspection apparatus according to the eighth embodiment of the present invention. As shown in FIG. 23, in addition to the measurement region setting unit 41, the image correction unit 42, and the formation index calculation unit 80, the image processing unit 34 includes a class-specific binary image data generation unit 91n and a class influence calculation unit 92. In addition to calculating the formation index DX and the class-specific formation index DXn, it is possible to calculate the influence of each class on how the formation is conspicuous. Note that, by using the influence level of each class, for example, it is possible to detect a defect of the paper P.

  Similarly to the image processing unit 34 shown in FIGS. 15 and 22, the image processing unit 34 shown in FIG. 23 also has a transmitted light type ground inspection apparatus 1 shown in FIG. 1 and a reflected light type ground formation shown in FIG. 11. The present invention can be applied to any of the inspection apparatus 2, the ground inspection apparatus 3 capable of performing the double-sided simultaneous inspection illustrated in FIG. 12, and the ground inspection apparatus 4 capable of performing the transmission method and the reflection system inspection illustrated in FIG.

  The class-specific binary image data generation unit 91n sets a threshold for each class-specific formation distribution image generated by the class-specific formation distribution image generation unit 82n. For example, the class-specific local formation index is lower than the threshold. A region to be enlarged and a region in which the class-specific local formation index is equal to or less than the threshold are extracted by binarization processing or the like. Then, class-specific binary image data that is data indicating each area extracted for each class is generated.

The class influence degree calculation unit 92 includes, from the class-specific binary image data generated by the class-specific binary image data generation unit 91n and the class-specific formation distribution image generated by the class-specific formation distribution image generation unit 82n, Calculate the degree of influence of each class on how conspicuous. Specifically, the class-specific binary image data generated by the class-specific binary image data generation unit 91n is _Bn , and the class-specific formation distribution image generated by the class-specific formation distribution image generation unit 82n is D. _{Assuming n} , a class influence degree determination image H indicating the degree of influence of each class on the conspicuousness of the formation is calculated from the following equation (19).

  In the above configuration, when the image data D1 stored in the storage unit 33 is read out and input to the image processing unit 34, the same processing as that in the sixth embodiment is performed, and the formation index DX and the classification-by-class formation are performed. An index DXn is calculated. Here, before the class-specific formation index DXn is calculated, the class-specific formation distribution image generation unit 82n generates a class-specific formation distribution image indicating a two-dimensional distribution of the class-specific local formation index. Done.

  The class-specific formation distribution images generated by the class-specific formation distribution image generation unit 82n are respectively input to the class-specific binary image data generation unit 91n, and binarization processing using a predetermined threshold is performed. Each class of binary image data is generated. Next, in the class influence calculation unit 92, the class-specific ground distribution image generated by the class-specific ground distribution image generation unit 82n and the class-specific binary image data generated by the class-specific binary image data generation unit 91n. Is used to calculate the class influence determination image H from the above-described equation (19).

  FIG. 24 is a diagram showing an example of a class influence degree determination image obtained in the eighth embodiment of the present invention. Note that the class influence degree determination image shown in FIG. 24 is obtained by observing the image shown in FIG. Further, in FIG. 24, a legend indicating the shading in each class is attached to the right side of the image. Referring to FIG. 24, it can be seen that, for example, the white circle appearing in the upper left portion E11 corresponding to the upper left portion E1 in which large unevenness exists in FIG.

  As described above, in the present embodiment, the class influence degree determination image H indicating the degree of influence of each class on the conspicuousness of the formation is calculated, so which class is affected by the conspicuous area. Can be easily determined. Note that the same processing as the processing for determining the defect area and defect type performed in the fifth embodiment is performed on the class influence determination image shown in FIG. Defect detection becomes possible.

[Paper manufacturing process]
FIG. 25 is a process diagram showing an outline of a production process of paper as an inspection object. As shown in FIG. 25, the paper production process includes a blending process S1, a forming process S2, a pressing process S3, a drying process S4, a calendar process S5, a reel process S6, and a cutting process S7. The blending step S1 is a step of blending a paper raw material liquid, and the forming step S2 is a wet paper that is dehydrated by a dehydrating device by ejecting the blended paper raw material liquid on a wire or between a pair of wires from a head box. It is a process.

  The pressing step S3 is a step of squeezing the wet paper from the forming step S2 with felt, and the drying step S4 is a step of drying the compressed wet paper. The calendar process S5 is a process for adjusting the paper thickness, paper smoothness and the like of the dried paper, and the reel process S6 is a process for winding the paper whose paper thickness and paper smoothness are adjusted. The cutting step S7 is a step of cutting the paper wound up in the reel step S6. When the paper wound in the reel process S6 is shipped as a product, the cutting process S7 is omitted.

  The ground inspection apparatus according to the first to eighth embodiments described above can be used as an off-line apparatus in which a manufactured product is an inspection object. As shown in FIG. 25, an intermediate product in the production process is an inspection object. It can also be used as an online device. In particular, when used as an online device, it can be installed in each process from the forming process S2 to the reel process S6 in FIG.

  Each of the ground inspection devices according to the first to eighth embodiments sets a local region in the inspection region, and generates a ground distribution image indicating a two-dimensional distribution of the local ground index obtained in each of the local regions. ing. For this reason, for example, from the viewpoint of quality control, it is possible to obtain the support effect in the cutting process S7 such as determining the area to be cut according to the level of formation of the paper P, and to recycle the bad area to the blending process. However, by preventing the product from being put on the market, the effect of improving the raw material cost and stabilizing the quality of the shipped product can be obtained. In addition, any of the ground inspection devices according to the first to eighth embodiments can manage the quality of the cut product for each ground level. For this reason, in the cutting process S6, it becomes possible to classify the use of the product for each level of formation.

  From the viewpoint of control management, for example, by analyzing the trend of change in formation at each point in the width direction of the inspection object, the change in formation index distribution over time is obtained, and the width direction of the inspection object in the formation area The effect of improving the quality can be obtained by obtaining the saliency with respect to and the periodicity with respect to the flow direction of the inspection object. For example, the periodicity with respect to the flow direction of the inspection object is obtained, the adjustment target of the previous process is extracted, and the adjustment target is adjusted (for example, speed adjustment and distribution adjustment of the head box, roller position adjustment and roller pressure of each process) Etc.), an effect of improving quality can be obtained, and the formation index can be used as a control factor.

  In addition, the ground inspection devices according to the sixth to eighth embodiments calculate the ground-based formation index DXn in addition to the formation index DX. For this reason, if the inspection object to be conveyed is continuously measured, the temporal change of the class-specific formation index DXn can be obtained, and the relationship between the temporal change of the class-specific formation index DXn and the operating conditions of the papermaking process. Sex can be easily grasped. This enables feedback to papermaking process control and development support for feedback methods.

  FIG. 26 is a diagram illustrating a change example of the class-specific formation index DXn according to the change of the operation condition. In the example shown in FIG. 26, the class-specific formation index DXn generally decreases due to the change of the operation condition that occurred at the time t1, and the class-specific formation index DXn becomes the time before the time t1 due to the change of the operation condition that occurred at the time t2. It has risen to the same level as the value of. Here, from experience, for example, in the blending step S1, the cloud-like formation (large flock) increases due to the increase in the density of the head box, but the granular formation (fine flock) tends to decrease, and the forming is performed. In step S2, it is known that the intensity of the intermediate frequency tends to change greatly due to an increase in the ejection speed from the head box. The formation of the paper can be arbitrarily controlled by referring to the temporal change in the formation-specific formation index DXn shown in FIG. 26 and feeding back to the control of the blending step S1 and the forming step S2.

  Further, the formation index DX and the class-specific formation index DXn calculated by the formation inspection apparatuses according to the sixth to eighth embodiments are the degree of conspicuousness (that is, the appearance quality) of the inspection object. For this reason, by measuring and comparing and evaluating the same position of the inspection object before and after each process in the papermaking process (for example, before and after the dry process S4 or before and after the calendar process S5), Relevance to appearance quality can be ascertained for each class. This enables feedback to papermaking process control and development support for feedback methods.

  Further, in the printing process, the same position of the inspection object is measured and compared and evaluated, so that the influence of the printing conditions on the appearance quality can be grasped for each class. This makes it possible to analyze the compatibility between the characteristics of the object to be inspected (for example, paper) that is to be printed and the side to be added by printing (for example, ink), and extract factors for improving printability Can help.

  FIG. 27 is a diagram for explaining the correlation between the formation index or class-specific formation index obtained before printing and the result of visual evaluation performed after printing, and (a) shows the formation index DX. It is a figure which shows the correlation with the result of visual evaluation, and (b) is a figure which shows the correlation with the formation index DXn classified by class and the result of visual evaluation. In FIG. 27B, only the correlation between the class-specific formation indices DX3, DX4, DX9 and the visual evaluation results for the classes C3, C4, C9 is shown for the sake of simplicity. .

  The above-mentioned formation index or the formation-by-class formation index is obtained by inspecting, for example, plain paper that has not been printed with the formation inspection apparatus according to the sixth to eighth embodiments of the present invention. is there. In addition, the result of the visual evaluation after the printing is obtained by the inspector visually evaluating the quality of the printed gray paper on the plain paper inspected by the ground inspection apparatus. It is what This visual evaluation is an eight-step evaluation from “1” to “8”, and the evaluation is better as the value is smaller, and is worse as the value is larger.

First, referring to FIG. 27 (a), the correlation coefficient R ² of the approximate curve Q11 showing a correlation between results of formation index DX and visual inspection obtained by the formation testing device is an "0.79" It can be seen that there is a high correlation. Referring now to FIG. 27 (b), the correlation coefficient R ² of the approximate curve Q21 showing a correlation between the results of each class formation index DX3 and visual evaluation is "0.79", by class formation the correlation coefficient R ² of the approximate curve Q22 showing a correlation between the results of the exponent DX4 and visual evaluation is "0.81", it can be seen that there is a respective high correlation.

In contrast, the correlation coefficient R ² of the approximate curve Q23 showing a correlation between the results of each class formation index DX9 and visual evaluation, it is apparent there is no correlation little since a value close to "0". As described above, a class having a high correlation with a class correlation index DXn obtained before printing and a result of visual evaluation performed after printing can be extracted, and a class having a low correlation can be extracted, which contributes to improving printability. It is possible.

  As mentioned above, although the ground inspection apparatus and the ground inspection method by embodiment of this invention were demonstrated, this invention is not necessarily limited to the said embodiment, In the range of this invention, it can change freely. For example, the image processing unit 34 included in the above-described formation inspection apparatus can be realized by hardware, but can also be realized by software. That is, a program recorded on a recording medium (a program that realizes the function of the image processing unit 34) is read and installed in a computer, or a program similar to the program recorded on the recording medium via a network such as the Internet. By downloading and installing in the computer, the image processing unit 34 can be realized by software.

  Moreover, in the said embodiment, the ground inspection apparatus which processes the image data output from the image input device 20, and calculates | requires the ground index DX and the class-specific ground index DXn was demonstrated. However, the ground inspection device is not necessarily provided with the image input device 20. For example, an image to be inspected is captured by an image input device separate from the ground inspection device, and the image data of the image captured by the image input device is processed by the ground inspection device, and the ground index DX Alternatively, the class-specific formation index DXn may be obtained.

In the above-described embodiment, the MTF calculation unit 56 uses the local area viewing angles d _h and d _v set by the local size calculation unit 53 and the average luminance L _Ai used by the FFT calculation unit 55, and uses the MTF image M. An example of generating _2i has been described. However, the MTF calculation unit 56 may be omitted, and the MTF image may be output from the inspection control unit 35 to the local formation index calculation unit 57 of the image processing unit 34.

1-4 Ground check device 20 Image input device 51 Luminance conversion unit 52 Observation luminance conversion unit 54 Local region setting unit 55 FFT calculation unit 56 MTF calculation unit 57 Local formation index calculation unit 59 Ground distribution statistical calculation unit 81n By class Local formation index calculation part DX formation index P paper

Claims (10)

In the ground inspection apparatus for inspecting the formation of the inspection object using the formation index obtained by performing image processing on image data obtained by imaging at least one of the one surface and the other surface of the sheet-like inspection object ,
First computing means for performing FFT computation on the image data to obtain first data indicating contrast for each spatial frequency;
Observation conditions defined by the first data obtained by the first computing means, an observation viewing distance parameter that defines a distance for observing the inspection object, and an observation luminance parameter that defines the luminance for observing the inspection object And second calculation means for calculating the formation index by multiplying and adding the second data indicating the contrast sensitivity for each spatial frequency when the inspection object is observed at each spatial frequency. A ground inspection device.   The second computing means performs a predetermined weighting on the value obtained by multiplying the first data obtained by the first computing means and the second data for each spatial frequency, and adds the values. The ground inspection apparatus according to claim 1, wherein   The ground inspection apparatus according to claim 1, further comprising conversion means for converting the image data into image data indicating physical luminance of the image.   The ground inspection apparatus according to claim 3, further comprising an adjustment unit that adjusts the image data converted by the conversion unit so that the physical luminance of the image becomes a preset reference luminance. Setting means for setting a plurality of local regions for the image data converted by the conversion means;
Statistics for calculating the formation index by performing a predetermined statistical calculation on the local formation index obtained by the calculation by each of the first and second calculation means for each local region set by the setting means The ground inspection device according to claim 4, further comprising: an arithmetic unit.   By adding the third data obtained by the second computing means by multiplying the first data and the second data for each spatial frequency for each class that is a region of a predefined spatial frequency, The ground inspection device according to any one of claims 1 to 5, further comprising third calculation means for calculating a ground index for each class.   The ground inspection apparatus according to claim 6, wherein the third calculation means performs a predetermined weighting defined for each of the classes on the third data and adds the weights for each of the classes.   Multiplying the third data obtained by the second computing means by multiplying the first data and the second data for each spatial frequency by multiplying by weight data defined for each predefined class The ground inspection apparatus according to any one of claims 1 to 5, further comprising a fourth calculation unit that calculates a ground index for each class.   The fourth computing means performs a predetermined weighting defined for each class on a value obtained by multiplying the third data and the weight data, and adds the weight for each class. The ground inspection apparatus according to claim 8. Image data indicating at least one image of one side and the other side of the sheet-like inspection object is acquired, and the formation of the inspection object is determined using a formation index obtained by performing image processing on the image data. A ground inspection method for inspecting,
A first step of performing an FFT operation on the image data to obtain first data indicating contrast for each spatial frequency;
For each spatial frequency when the inspection object is observed under an observation condition defined by an observation viewing distance parameter that defines a distance for observing the inspection object and an observation luminance parameter that defines a luminance for observing the inspection object. A second step of obtaining second data indicating contrast sensitivity;
A third step of calculating the formation index by multiplying and adding the first data obtained in the first step and the second data obtained in the second step for each spatial frequency. A ground inspection method characterized by that.