JP6958136B2 - Image inspection device and image inspection method - Google Patents

Description

The present invention relates to an image inspection device and an image inspection method.

Conventionally, there is known a method of performing segmentation of a color image that classifies the color of each pixel of a color image into a basic color in a color space (see, for example, Patent Document 1).
In the method described in Patent Document 1, an input image is converted into a color system such as CIE-LAB or CIE-LUV that is close to human perception, and then a clustering process is performed. In such a segmentation process, for example, in an inspection on a production line or the like, when the imaging process and the segmentation process are repeatedly performed, a parameter indicating a range of which color is classified into which basic color is set in advance. By setting it, stable segmentation processing can be performed.

Here, when performing the color image segmentation processing, the color image acquisition environment may be different from each other. In this case, it is necessary to appropriately adjust the parameter indicating the range of which color is classified into which basic color (that is, the range of colors indicated by the basic color). Here, a method using the Mahalanobis distance as the segmentation of a color image is known (see, for example, Patent Document 2). In this method using the Mahalanobis distance, it is possible to perform segmentation of a color image by adjusting parameters in consideration of the spread of distribution for each classification item (basic color).

Japanese Unexamined Patent Publication No. 2003-162718 Japanese Unexamined Patent Publication No. 2013-107269

By the way, there are various variation factors (differences in color appearance depending on the environment) to be considered depending on the detection content, and it is difficult to adjust parameters corresponding to the variation factors by the method of Patent Document 2. For example, if there is an adjustable first parameter (eg R range for the base color), a second parameter (eg G range for the base color), and a third parameter (eg B range for the base color). In addition, it becomes necessary to adjust each of the three parameters in order to change the allowable range of brightness.

An object of the present invention is to provide an image inspection apparatus and an image inspection method capable of easily performing parameter adjustment for performing image segmentation.

The image inspection apparatus according to an application example of the present invention has an image information acquisition unit having a plurality of pixels and acquiring image information in which signal values are recorded for each of the pixels, and the signal value with respect to a reference color. The reference signal acquisition unit that acquires the reference signal value, and the signal value of the pixel of the image information are converted into an uvw coordinate system using a rotation matrix, and further, the converted coordinate value of the uvw coordinate system is converted. A signal conversion unit that converts sphere coordinates to acquire sphere coordinate values, a parameter acquisition unit that is set for each of a plurality of classification categories that divide colors and acquires parameters indicating the distribution of colors belonging to the classification categories, and sphere coordinates. The rotation is provided with an index calculation unit that calculates an index value for classifying the pixels based on the color classification reference value for each classification category set in the space, the parameter, and the sphere coordinate value. The matrix is characterized in that the coordinate values when the reference signal value is converted into the uvw coordinate system are the coordinate values on the w axis.

In this application example, the image information acquisition unit acquires the image information in which the signal values (for example, R, G, B gradation values) for each pixel are recorded, and the signal conversion unit rotates the signal value of each pixel. It is converted into an uvw coordinate system using a matrix, and the coordinate values of the converted uvw coordinate system are further converted into spherical coordinate values. At this time, the signal conversion unit uses a rotation matrix in which u = v = 0 when the reference signal value acquired by the reference signal acquisition unit is converted into the uvw coordinate system. Then, the index calculation unit calculates an index value for classifying colors based on the spherical coordinate values, the color classification reference value set for each classification classification, and the parameters that can be adjusted by the parameter acquisition unit. ..
In such an image inspection device, φ in the spherical coordinate values (θ, φ, r) is the angle formed by the direction from the origin to the converted coordinates and the w axis when the signal value is converted into the uvw coordinate system. .. Further, θ is an angle formed by the v-axis when the transformed coordinates are projected onto the uv plane in the uvw coordinate system. These θ and φ indicate the difference in chromaticity from the reference color. Further, r in the spherical coordinate values (θ, φ, r) appears as a difference in brightness from the reference color. Therefore, when the parameter acquisition unit sets the parameters for the spherical coordinate values, the parameters for these θ, φ, and r may be set. For example, when adjusting the permissible range of the chromaticity to be classified into each classification, it is sufficient to adjust the parameters related to the angular coordinates θ or φ, and it is not necessary to adjust the parameters related to the radial coordinates r. Further, when adjusting the permissible range of brightness classified into each classification category, it is sufficient to adjust the parameters related to the radial coordinates r, and it is not necessary to adjust the parameters related to the angular coordinates θ and φ. As described above, in this application example, parameter adjustment for performing image segmentation can be easily performed.

In the image inspection device of this application example, the index calculation unit, the spherical coordinates (θ, φ, r), the color classification reference value _{(θ m, φ m, r} m), said parameter (sigma It is preferable to calculate the index value D _m by the following equation (1) _{as θm} , σ _φm , σ _rm).

In this application example, an index value based on the difference between the color classification reference value and the spherical coordinate value is calculated as in the equation (1). In this case, the _{smaller the index value D m} , the closer the pixel color is to the color of the classification category, and it becomes possible to easily perform image segmentation. _{Furthermore, by setting parameters (σ θm} , σ _φm , σ _rm ) corresponding to each of the spherical coordinate values θ, φ, and r, it is possible to calculate an index value in consideration of the spread of the distribution for each classification category.

In the image inspection apparatus of this application example, it is preferable that the reference signal acquisition unit acquires the signal value for the predetermined pixel in the image information acquired by the image information acquisition unit as the reference signal value.
In this application example, the reference signal acquisition unit acquires the signal value for a part of the pixels in the image information as the reference signal value. As a part of the image information, for example, it may be a pixel selected by the user, a pixel arranged on the outer circumference of the image information, or the like. Further, a part may be one pixel or a plurality of pixels. In the case of a plurality of pixels, the average value of the signal values of each pixel may be used. The image information may be image information including the object to be inspected, or may be image information not including the object.
As the reference color selection position, for example, the pixels of the background portion in the image information including the inspection target, the tray on which the inspection target is placed, the uncolored portion of the inspection target, and the like can be exemplified.
In this application example, as described above, by selecting the pixel of the portion to be used as the reference from the image information, it is possible to save the trouble of the user manually inputting the reference color at the time of segmentation.

The image inspection apparatus in this application example includes a classification unit that classifies the color of the pixel based on the index value, and the classification unit may classify the pixel into the classification category that minimizes the index value. preferable.
In this application example, the image inspection apparatus includes a classification unit, and this classification unit classifies each pixel based on an index value. As a result, one of the classification categories is assigned to each pixel of the image information, and image segmentation can be performed.

The image inspection device according to an application example of the present invention is an image inspection method for inspecting an image by a computer, in which the computer has a plurality of coordinates and a signal value is recorded for each of the coordinates. The image information acquisition step for acquiring image information, the reference signal acquisition step for acquiring the reference signal value which is the signal value for the reference color, and the signal value of the pixel of the image information are uvw coordinates using a rotation matrix. A signal conversion step of converting to a system and further converting the coordinate values of the converted uvw coordinate system to obtain the sphere coordinate values and a plurality of classification categories for classifying colors are set in the above classification categories. Based on the parameter acquisition step for acquiring the parameter indicating the distribution of the color to which the color belongs, the color classification reference value for each classification category set in the sphere coordinate space, the parameter, and the sphere coordinate value, the pixel is An index calculation step for calculating an index value to be classified is performed, and the rotation matrix used in the signal conversion step has a coordinate value on the w-axis when the reference signal value is converted into the uvw coordinate system. It is characterized in that it is a matrix that becomes.
In this application example, as in the application example described above, for example, when adjusting the permissible range of colors to be classified into each classification category, the parameters related to the values of the angular coordinates θ or φ in the spherical coordinate values (θ, φ, r) are set. It may be adjusted, and it is not necessary to adjust the parameters related to the radial coordinates r. Further, when adjusting the permissible range of brightness classified into each classification category, it is sufficient to adjust the parameters related to the radial coordinates r, and it is not necessary to adjust the parameters related to the angular coordinates θ and φ. As described above, in this application example, parameter adjustment for performing image segmentation can be easily performed.

The block diagram which shows the schematic structure of the image inspection apparatus of one Embodiment which concerns on this invention. The block diagram which shows the functional structure of the arithmetic circuit part of this embodiment. The flowchart which shows the image inspection method of this embodiment. The flowchart which shows the parameter setting process in the image inspection method of this embodiment.

Hereinafter, an embodiment according to the present invention will be described.
[Configuration of image inspection equipment]
FIG. 1 is a block diagram showing a configuration of an image inspection device 1 according to the present embodiment.
This image inspection device 1 is provided on a production line such as a factory, captures a work W placed on a tray 21 of a shooting table 2, and classifies the color of each pixel of the captured image information (color image). Then, it is a device for inspecting the type and number of work W.
As shown in FIG. 1, the image inspection device 1 includes a light source unit 11, an image pickup device 12, a monitor 13, an input operation unit 14, and a control device 15.

The light source unit 11 is composed of a light source 111 and a light source driving unit 112.
The light source 111 is a light emitting element such as an LED, and irradiates the work W with illumination light.
The light source driving unit 112 is a driving circuit that drives the light source 111, and applies a driving voltage to the light source 111 to irradiate the illumination light from the light source 111. The light source driving unit 112 can appropriately set the voltage to be output to the light source 111, thereby irradiating the work W with illumination light of a desired amount of light.

The image pickup apparatus 12 outputs the image information obtained by imaging the photographing table 2 to the control device 15. As the image pickup apparatus 12, a general image sensor such as a CCD can be used, and for example, it is configured to include RGB filters arranged in a Bayer array and a plurality of light receiving elements that receive light transmitted through each filter. ing. That is, the image pickup device 12 generates image information in which the signal value of each color (RGB) is recorded for each pixel based on the light receiving signal output from each light receiving element, and transmits the image information to the control device 15.

The monitor 13 is connected to the control device 15 and displays a predetermined image under the control of the control device 15. As the image to be displayed, for example, the segmentation result of the captured image can be exemplified.
The input operation unit 14 is composed of an input device such as a mouse or a keyboard, and is connected to the control device 15. When the input operation unit 14 is operated by the user, the input operation unit 14 outputs an operation signal according to the operation content to the control device 15. In the present embodiment, the input of the operation signal by the input operation unit 14 is illustrated, but the present invention is not limited to this, and for example, a predetermined operation signal is input to the control device 15 via a network such as the Internet. May be.

The control device 15 can be configured by a computer such as a personal computer. As shown in FIG. 1, the control device 15 includes a light source control unit 151, an image pickup control unit 152, a monitor control unit 153, a storage unit 154, an arithmetic circuit unit 155, and the like.

The light source control unit 151 is composed of a driver circuit or the like for controlling the light source unit 11, such as turning on or off the light source 111, controlling the amount of illumination light output from the light source 111 (voltage command applied to the light source 111), and the like. To carry out.
The image pickup control unit 152 is composed of a driver circuit or the like for controlling the image pickup device 12, and gives an image pickup command to the image pickup table by the image pickup device 12. Further, the captured image (image information) input from the imaging device 12 is acquired and stored in the storage unit 154.
The monitor control unit 153 is composed of a driver circuit or the like for controlling the monitor 13, and causes a predetermined image to be displayed on the monitor 13 based on a command from the arithmetic circuit unit 155.
The storage unit 154 includes, for example, an auxiliary storage device composed of a hard disk or the like, or a main storage device composed of a memory or the like. Various data and various programs are stored in the storage unit 154.

FIG. 2 is a functional block diagram showing a functional configuration of the arithmetic circuit unit 155.
As shown in FIG. 2, the arithmetic circuit unit 155 reads and executes various profiles stored in the storage unit 154, and as shown in FIG. 2, the image information acquisition unit 155A, the image processing unit 155B, the signal acquisition unit 155C, the signal conversion unit 155D, It functions as a parameter acquisition unit 155E, an index calculation unit 155F, a classification unit 155G, an inspection unit 155H, a display control unit 155I, and the like.

The image information acquisition unit 155A acquires the image information captured by the image pickup device 12. Specifically, the image information imaged by the image pickup apparatus 12 and stored in the storage unit 154 is read out.
The image processing unit 155B performs various processes on the acquired image information. For example, shading correction of the acquired image information is performed.
The signal acquisition unit 155C also functions as a reference signal acquisition unit of the present invention, and acquires the signal value of a specific pixel of each pixel of the image information.
The signal conversion unit 155D converts the signal value of the image information based on the RGB space (photographed signal space) into the spherical coordinate value (θ, φ, r) based on the spherical coordinate space.
The parameter acquisition unit 155E acquires parameters for each classification category for classifying colors based on the operation signal input from the input operation unit 14, and stores the parameters in the storage unit 154.
The index calculation unit 155F calculates an index value for classifying colors into a plurality of classification categories.
The classification unit 155G classifies each pixel into one of a plurality of classification categories based on the index value.
The inspection unit 155H detects the inspection target included in the image information based on the classification result and executes the inspection process.
The display control unit 155I causes the monitor 13 to display the captured image information, the result of the segmentation process, and the like.
A detailed description of each function will be described later.

[Image inspection method]
Next, the image inspection method in the image inspection apparatus 1 as described above will be described.

FIG. 3 is a flowchart showing the image inspection method of the present embodiment.
In the image inspection process of the present embodiment, the color of each pixel in the captured image (image information) obtained by capturing the work W is classified into one of a plurality of classification categories to perform color discrimination.
Specifically, first, the image information acquisition unit 155A of the image inspection device 1 controls the light source unit 11 and the image pickup device 12 to perform imaging of the work W, and acquires the captured image information (step S1). ..
As for the acquisition of the image information, the image pickup apparatus 12 may perform one image pickup process to acquire the imaged image information, or may perform the image pickup process a plurality of times. When performing the imaging process a plurality of times, for example, a representative value of each signal value obtained by the plurality of imagings is used as the signal value of each pixel. The representative value may be, for example, an average value, a median value, a mode value, or a root mean square.

Further, the image processing unit 155B performs shading correction of the acquired image information (step S2).
In the shading correction, the brightness distribution in the image information is corrected based on the white reference image and the black reference image (dark image). The white reference image and the dark image may be measured in advance prior to the main measurement. In the imaging of the white reference image, for example, a standard plate (for example, an 18% standard reflector, a white reference plate, etc.) whose reflectance for each wavelength is known is placed on the photographing table 2 and imaged by the imaging device 12. You can get it. For the dark image, the light source 111 is turned off to make the shooting space a dark space, and then the image pickup device 12 captures the captured image. Here, it is assumed that the gradation value of the captured image information is substantially linear with respect to the light receiving intensity of the image pickup device 12.
In the shading correction, the signal value (RGB value) of each pixel before the shading correction is (R ₀ , G ₀ , B ₀ ), and the signal value (RGB value) of each pixel in the white reference image is (R _0W , G _0W , B _0W ), the signal value (RGB value) of each pixel of the dark image is (R _0D , G _0D , B _0D ), and the image information acquisition unit 155A sets the signal value (R) of each pixel of the image information after shading correction. , G, B) is calculated by the following formula (2).

In the equation _{(2), (R W,} G W, B W) is a signal value of each pixel in the white reference image after the shading _{correction, (R D, G D,} B D) , the shading correction It is a signal value of each pixel in a later dark image, and these are values that are preset and stored in the storage unit 154.
By performing the shading correction as described above, the change in brightness in the captured image information is mainly caused by the shape of the work W and the like, and the accuracy of the segmentation process can be improved.

Next, the classification unit 155G extracts pixels to be excluded from the segmentation process in the acquired image information (step S3). Specifically, the signal values (R, G, B) for a pixel either is equal to or greater than a predetermined upper limit threshold value T _MAX of a saturated signal exceeding the received upper-limit value in the imaging device, the pixel Exclude from the target of segmentation processing. In this case, the classification unit 155G determines that the pixel is an overflow pixel, and sets, for example, the classification ID to "01".
Further, it is determined that the pixels in which all the signal values (R, G, B) are _{equal to or less than the predetermined lower limit threshold value TMIN} are underflow pixels (pixels in the vicinity of dark), and the classification ID is set to "02", for example. Set.

After that, the signal conversion unit 155D converts the signal values (R, G, B) of each pixel into spherical coordinate values (θ, φ, r) for the pixels not excluded in step S3 (step S4). ).
Specifically, the signal conversion unit 155D, first, using a rotation matrix _{M rot,} as shown in the following formula (3), the signal values (R, G, B) and, uvw coordinate system signal value of (u , V, w).

Here, the rotation matrix _{M rot,} the reference signal value corresponding to a reference color _{(R A, G A, B} A) to when converted by the rotation matrix _{M rot,} as converted coordinate is the upper w axis It is a set matrix. That is, the rotation matrix _Mrot satisfies the condition of the following equation (4).
The reference signal value with respect to the reference color _{(R A, G A, B} A) , in the parameter setting processing to be described later, a value set, previously stored in the storage unit 154.

Then, the signal conversion unit 155D further converts the signal values (u, v, w) into spherical coordinates as in the following equation (5) to obtain the spherical coordinate values (θ, φ, r).

After that, the segmentation process is performed.
To this end, the classification unit 155G first performs a determination process of whether or not each pixel has been classified (step S5). That is, in step S3, the pixels whose classification ID is set to "01" or "02" are classified and are determined to be "Yes" and excluded.

On the other hand, the pixel to which the classification ID is not attached in step S3 is determined as "No". In this case, the index calculation unit 155F calculates the index value D for the pixel for each of the plurality of classification categories (step S6).
Specifically, the index calculation unit 155F calculates the index value D for the classification category m based on the following formula (1).

Here, the subscript m of the index value D is a variable for identifying the classification category, and the index value D _m indicates the index value with respect to the classification category m. Further, the classification color classification reference value for segment _{m (θ m, φ m,} r m), parameter a for classification category _{m (σ θm, σ φm,} σ rm) are expressed as. Color classification reference value (theta _m, phi _{m, r m)} is classification categories is a central parameter in m, converts the RGB values of the representative colors (center color) corresponding to the classification category m to uvw value by a rotation matrix _{M rot} After that, it is a spherical coordinate value converted into spherical coordinates. Moreover, parameters _{(σ θm, σ φm, σ} rm) is a value indicating the allowable range for classification category m, the range of _{(θ m ± σ θm, φ} m ± σ φm, r m ± σ rm), The color is classified into the classification category m. The parameters (σ _{θ m} , σ _{φ m} , σ _rm ) also have a function of normalizing the difference in the unit system of angle and distance.
It should be noted that these color classification reference value _{(θ m, φ m, r} m), and parameters _{(σ θm, σ φm, σ} rm) is set in the parameter setting processing to be described later, is stored in advance in the storage unit 154 There is.

The index value D shown in the equation (1) is a value corresponding to the distance in consideration of the spread of the distribution with respect to the classification target. _{With such an index value D, the parameters (σ θ m} , σ _{φ m} , σ _rm ) can be easily changed according to the variation factors depending on the shooting environment and the like.
That is, in the spherical coordinate value, r indicates the distance from the dark value in the uvw coordinate system, and _{by adjusting the parameter σ rm} with respect to r, the allowable range for the change in brightness due to shading or the like is set. be able to.
Further, φ indicates the direction formed by the dark value toward the coordinates corresponding to the color of each pixel and the w-axis in the uvw coordinate system. φ indicates, for example, the change in chromaticity with the work W due to the coating thickness of the work W, and by _{adjusting the parameter σ φm} with respect to φ, the allowable range for the change in chromaticity due to the amount of coloring material used. Can be set.
Further, θ indicates the direction from the dark value toward the coordinates corresponding to the color of each pixel and the angle formed by the v-axis when the coordinates of the uvw coordinate system are projected onto the uv plane. This θ indicates the change in chromaticity due to other factors such as the change in color material for each type of work W. _{By adjusting the parameter σ θm} with respect to θ, for example, the chromaticity due to the difference in color material You can set the tolerance for changes in.

In step S6, the index calculation unit 155F calculates the index value D based on the equation (1) for each of the plurality of preset classification categories.
_{Then, the classification unit 155G determines whether or not the minimum index value Min {D m} } among the plurality of index values D calculated for the target pixel is equal to or less than a predetermined threshold value D _{t (step).} S7).
When it is determined as No in step S7, it indicates that the colors of the target pixels are not similar to each classification. In this case, the classification unit 155G determines that the target pixel does not correspond to any of the classification categories, and sets the classification ID to, for example, "99" (step S8).
On the other hand, when it is determined as "Yes" in step S6, the target pixel is _{classified into the classification category corresponding to the minimum index value Min {D m} }, and the classification ID corresponding to the classification category is set (step S9). ).

After the above, the inspection unit 155H detects the work W to be inspected included in the image information based on the classification result of the segmentation process, and executes the inspection process (step S10). For this purpose, for example, the inspection unit 155H performs pattern detection based on the template data stored in the storage unit 154, and detects the work W included in the image information. Then, the type and number of detected work W are determined.

[Parameter setting process]
Then, the reference color _{(R A, G A, B} A), the color classification reference value _{(θ m, φ m, r} m), and parameters _{(σ θm, σ φm, σ} rm) for parameter setting processing for setting explain.
In the parameter setting process, the work W colored with the color sample corresponding to each classification is placed on the tray 21 of the photographing table 2. A plurality of work Ws corresponding to a plurality of color samples may be placed. Further, when setting the color of the uncolored base material as the reference color, for example, the uncolored base material is placed on the tray 21 as the work W.

FIG. 4 is a flowchart showing a parameter setting process in the image inspection method.
In the parameter setting process, first, the image information acquisition unit 155A of the image inspection device 1 controls the image pickup device 12 to perform imaging of the work W, and reads out and acquires the captured image information (step S11).
As for the acquisition of the image information, the imaging device 12 may perform one imaging process to acquire the captured image information, or may perform a plurality of imaging processes. Parameter setting process, the reference signal value corresponding to a reference color _{(R A, G A, B} A) other, based on a plurality of sample data, obtains the center value and variance for the classification target. Therefore, when acquiring a plurality of sample data for the classification category by one imaging process, the measurement area of the classification category is set in advance, and the signal values for the plurality of pixels included in the measurement area are combined with the plurality of sample data. do. Further, when the imaging process is performed a plurality of times, a plurality of signal values acquired for a predetermined pixel corresponding to the classification classification can be used as a plurality of sample data.
Further, the acquired image may be subjected to shading correction as in step S2.

Next, the signal acquisition unit 155C acquires the signal value (reference signal value) of the reference pixel having the reference color from each pixel of the captured image information (step S12).
For example, when inspecting a manufactured product (work W) such as a semiconductor chip on a factory production line or the like, the image pickup device 12 images the work W conveyed to the photographing table 2 by a transfer means such as a belt conveyor. In this case, the imaging range of the imaging device 12 is a preset range, and a specific pixel in the image information can be set as a reference pixel. Examples of the specific pixel position include background pixels (pixels corresponding to the photographing table 2 and pixels corresponding to the tray 21) in the vicinity of the outer peripheral edge of the image.
Further, as described above, an uncolored base material (base material of the work W) can be set as the reference color. In this case, if the placement position of the uncolored base material is set in advance on the tray 21, the reference pixel can be easily specified.

Further, the reference pixel may be appropriately selected by the operation of the input operation unit 14 of the user. In this case, the display control unit 155I causes the monitor 13 to display the captured image information. When a plurality of times of imaging are performed, a representative value may be used as the signal value of each pixel, and for example, the image captured for the first time may be displayed as the representative image. As a result, the user can select the reference pixel based on the image displayed on the monitor 13.
Signal acquiring unit 155C acquires the reference signal value at the set reference pixels as described above _{(R A, G A, B} A) a. In the case where a plurality of image information was performed a plurality of times of imaging processing is obtained, the representative value of the signal value of each reference pixel, the reference signal value can be _{(R A, G A, B} A) and. As a representative value, for example, an average value, a median value, a mode value, a root mean square, and the like can be exemplified.

Further, the signal acquisition unit 155C further specifies the pixel position corresponding to each classification and acquires the signal value (sample data) (step S13).
As described above, when the image pickup range by the image pickup apparatus 12 is constant, the position of the work W in which the color sample is colored is placed at a preset position, so that the pixels corresponding to each classification in the image information. Can be identified. Further, the pixels corresponding to each classification may be designated by the operation of the input operation unit 14 of the user. As described above, when acquiring a plurality of sample data based on one image information, an area corresponding to one classification category is set, and the signal value in each pixel in the area is set. Obtain as sample data. Further, in the case of performing the imaging process a plurality of times, it is sufficient to acquire the signal value (sample data) obtained by the a plurality of imagings for the set pixels.
Then, the signal acquisition unit 155C calculates representative values of a plurality of signal values (sample data) for each classification category and sets them as color classification reference values (R _m0 , G _m0 , B _m0 ) (step S14). As a representative value, for example, an average value, a median value, a mode value, a root mean square, and the like can be exemplified.

Thereafter, the signal conversion unit 155D is based on the reference signal value _{(R A, G A, B} A), and calculates the rotation matrix _{M rot} satisfying the expression (4) (step S15). Calculated here a rotation matrix _{M rot} the reference signal value _{(R A, G A, B} A) together is stored in the storage unit 154 is used in the step S4 described above.

Then, the signal conversion unit 155D as in step S4, using a rotation matrix M _rot, a plurality of sample data and for the classification categories, converting each color segment reference value uvw coordinate system, further, the spherical coordinate values Convert (step S16).
The spherical coordinate value obtained by converting the color classification reference value (R _m0 , G _m0 , B _m0 ) into spherical coordinates becomes the initial color classification reference value (θ _m0 , φ _m0 , r _m0 ). Further, the signal conversion unit 155D calculates the respective variances for θ, φ, and r from each spherical coordinate value when each sample data is converted into spherical coordinates, and the initial parameters (σ _θm0 , σ _φm0 , σ _rm0). ) (Step S17).

Thereafter, the display control unit 155I are various parameters, i.e., the reference signal value _{(R A, G A, B} A), initial color classification reference value _{(θ m0, φ m0, r} m0), and the initial parameter _(σ θm0 , Σ _φm0 , σ _rm0 ) is displayed on the monitor 13 (step S18). At this time, the reference signal value _{(R A, G A, B} A) together, to display the reference color, and further, the initial color classification reference value _{(θ m0, φ m0, r} m0) and initial parameter _(σ θm0, σ φm0 , Σ _rm0 ), the colors corresponding to each classification, and the permissible range thereof may be displayed by gradation or the like.

After that, the parameter acquisition unit 155E determines whether or not an operation signal for changing the correction parameter has been input by the operation of the input operation unit 14 (step S19). The correction parameters include color classification reference values and parameters that are within the permissible range of each classification classification.
In step S19, if it is determined that Yes, based on the input correction parameters, initial color classification reference value _{(θ m0, φ m0, r} m0) and initial parameter _{(σ θm0, σ φm0, σ} rm0) Fix and, the color classification reference value _{(θ m, φ m, r} m) and parameters _{(σ θm, σ φm, σ} rm) is stored as (step S20).
On the other hand, in step S19, if it is determined No, the initial color classification reference value _{(θ m0, φ m0, r} m0) and initial parameter _{(σ θm0, σ φm0, σ} rm0) color classification reference value (theta _m , φ _{m, r m)} and parameters _(σ θm, σ φm, stored as sigma _rm) (step S21).

In the present embodiment, as described above, when changing the allowable range for brightness change due to shading, etc., may be changed color classification reference value r _m and parameter sigma _rm. Also, when changing the allowable range for the chromaticity change due to the amount of the coloring material may be changed color classification reference value phi _m and parameter sigma _{[phi] m.} Furthermore, when changing the allowable range for the chromaticity change due to the difference of the coloring material may be changed color classification reference value theta _m and parameter sigma _.theta.m. That is, in step S20, the user does not need to change a plurality of items when, for example, changing a parameter for one factor, and can easily make a correction for a target item.

[Action and effect of this embodiment]
In the present embodiment, the signal value of each pixel of the image information acquired by the image information acquisition section 155A (R, G, B) and, using a rotation matrix _{M rot,} the signal value of the uvw coordinate system (u, v, It is converted to w), and further converted to spherical coordinate values (θ, φ, r) by spherical coordinate conversion. In this case, the signal conversion unit 155D includes the reference signal value corresponding to the reference color when converting _{(R A, G A, B} A) to uvw coordinate system, (u, v, w) = (0,0, k _WA ), that is, the rotation matrix _Mrot whose transformed coordinates are the w-axis is used. The index calculation unit 155F includes this spherical coordinate values (θ, φ, r), classification category set color classification reference value _{m (θ m, φ m,} r m) and, adjustable by parameter acquiring unit The index value D _m for classifying colors is calculated based on various parameters (σ _{θ m} , σ _{φ m} , σ _rm).
In such an image inspecting apparatus, the dark value _{(R D, G D, B} D) dark value obtained by converting the uvw coordinate system _{(u D, v D, w} D) as the origin of the dark values _(u D, v The color difference from the reference color is represented by the angle formed by the direction from _D , w _{D) to the signal value (u, v, w) and the w axis.} This color difference is converted into θ and φ by spherical coordinate conversion. Therefore, when adjusting the permissible range of chromaticity, it is not necessary to change the _{parameter σ rm with respect to r, and only σ θm} or σ _φm needs to be changed. Further, when changing the permissible range of brightness, it is not necessary to change _{σ θm} or σ _{φm, and only σ rm} needs to be changed. As described above, in the present embodiment, the user does not need to search for the optimum value while fine-tuning a plurality of parameters when setting the parameters, and the variation factors (for example, brightness, coating thickness of the work W) are not required. , The type of coloring material, etc.), and only the parameters need to be adjusted. As a result, parameter adjustment for performing image segmentation can be easily performed.

In the present embodiment, the index value D _m is calculated based on the equation (1). With such an index value D _m , the smaller the value, the closer the pixel color is to the color of the classification category, and it becomes possible to easily perform image segmentation. _{In addition, by setting parameters (σ θm} , σ _φm , σ _rm ) corresponding to each of the spherical coordinate values θ, φ, and r, it is possible to calculate an index value in consideration of the spread of the distribution for each classification category.

In the present embodiment, the signal acquisition unit 155C acquires a signal value for a part of pixels in the image information as a reference signal value. As a result, the color set by the user can be used as the reference color.

In the present embodiment, the classification unit 155G classifies the pixels into the classification classification corresponding to the minimum index value D. As a result, one of the classification categories is assigned to each pixel of the image information, and image segmentation can be preferably performed.

[Modification example]
The present invention is not limited to the above-described embodiment, and modifications, improvements, and the like within the range in which the object of the present invention can be achieved are included in the present invention.

In step S12 of the parameter setting process, an example of acquiring the reference pixel and the reference signal value from the image information has been shown, but the present invention is not limited to this.
For example, an image for setting a reference color may be imaged separately from the image of the work W in which the color sample is colored. Further, the reference signal value may be input by the user's input operation. Alternatively, a reference color may be set in advance and stored in the storage unit 154. In this case, the rotation matrix _Mrot is also a known value.

An example in which the image information acquisition unit 155A reads out the image information imaged by the image pickup device 12 and stored in the storage unit 154 is shown, but the present invention is not limited thereto. For example, the image information acquired via the Internet or the like may be acquired.
In the above embodiment, as an example, an example in which the image inspection device 1 is provided in a production line such as a factory is shown, but the present invention is not limited to this, and for example, the image inspection of the present invention is performed on a printing device such as a printer. The device may be used. In this case, the print correction can be suitably performed by inspecting the color patch (pattern image) printed by the printer with the image inspection device 1.

In the above embodiment, an example in which the image inspection process and the parameter setting process are separated has been shown, but the present invention is not limited to this, and the parameters (σ _{θ m} , σ _{φ m} , σ _rm ) are set in the middle of the image inspection process. It may be changeable. For example, when No is determined in step S7, the parameters (σ _θm , σ _φm , σ _rm ) may be changed.

In addition, the specific structure for carrying out the present invention can be appropriately changed to another structure or the like within the range in which the object of the present invention can be achieved.

1 ... Image inspection device, 2 ... Shooting table, 11 ... Light source unit, 12 ... Imaging device, 13 ... Monitor, 14 ... Input operation unit, 15 ... Control device, 21 ... Tray, 154 ... Storage unit, 155 ... Arithmetic circuit unit , 155A ... Image information acquisition unit, 155B ... Image processing unit, 155C ... Signal acquisition unit, 155D ... Signal conversion unit, 155E ... Parameter acquisition unit, 155F ... Index calculation unit, 155G ... Classification unit, 155H ... Inspection unit, 155I ... Display control unit, W ... work.

Claims (4)

An image information acquisition unit having a plurality of pixels and acquiring image information in which signal values are recorded for each pixel, and an image information acquisition unit.
A reference signal acquisition unit that acquires a reference signal value that is the signal value for the reference color, and
A signal conversion unit that converts the signal value of the pixel of the image information into an uvw coordinate system using a rotation matrix, and further converts the converted coordinate value of the uvw coordinate system into a spherical coordinate system to acquire a spherical coordinate value. When,
A parameter acquisition unit that is set for each of a plurality of classification categories for classifying colors and acquires parameters indicating the distribution of colors belonging to the classification categories.
An index calculation unit that calculates an index value for classifying the pixels based on the color classification reference value for each classification category set in the spherical coordinate space, the parameter, and the spherical coordinate value.
A classification unit that classifies the color of the pixel based on the index value is provided.
The rotation matrix, Ri matrix der coordinate values at the time of converting the reference signal value in uvw coordinate system is a coordinate value on the w axis,
The classification unit is an image inspection apparatus characterized in that the pixels are classified into the classification classification in which the index value is the minimum. In the image inspection apparatus according to claim 1,
The index calculation unit sets the spherical coordinate values as (θ, φ, r) and the color classification reference values as (θm, φm).
, Rm), with the parameters as (σθm, σφm, σrm), and the index value Dm as the following equation (
Calculated by 1)

An image inspection device characterized in that. In the image inspection apparatus according to claim 1 or 2.
The reference signal acquisition unit is an image inspection device that acquires the signal value for a predetermined pixel in the image information acquired by the image information acquisition unit as the reference signal value. An image inspection method that inspects images with a computer.
The computer
An image information acquisition step of having a plurality of pixels and acquiring image information in which a signal value is recorded for each of the pixels,
A reference signal acquisition step for acquiring a reference signal value which is the signal value with respect to the reference color, and
A signal conversion step of converting the signal value of the pixel of the image information into an uvw coordinate system using a rotation matrix, and further converting the converted coordinate value of the uvw coordinate system into a spherical coordinate system to acquire a spherical coordinate value. When,
A parameter acquisition step for acquiring a parameter indicating the distribution of colors belonging to the classification category, which is set for each of a plurality of classification categories for classifying colors.
An index calculation step for calculating an index value for classifying the pixels based on the color classification reference value for each classification category set in the spherical coordinate space, the parameter, and the spherical coordinate value.
A classification step of classifying the color of the pixel based on the index value is performed.
The rotation matrix used in the signal conversion step is a matrix in which the coordinate values when the reference signal value is converted into the uvw coordinate system become the coordinate values on the w axis.
In the classification step, the pixels are classified into the classification category having the minimum index value.
including,
An image inspection method characterized by that.

Images