JP2001324385A - Image processing device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing device capable of measuring and processing image data relative to a small object image. SOLUTION: In this image processing device, a reflectance measuring part measures a spectral reflectance in pixel unit relative to a measuring object, and a display means displays an image formed from spectral reflectance data measured by the reflectance measuring part. A designation means designates an image part to be extracted by a user relative to the image displayed by the display means, and an extraction means determines an average spectral reflectance (color) on the image part designated by the designation means from the spectral reflectance data measured by the reflectance measuring part.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to image measurement and image data processing.

[0002]

2. Description of the Related Art The color of an object depends on the color reflected from the surface of the object and is represented by tristimulus values, spectral characteristics, and the like. When a plurality of colorants such as dyes are mixed in a woven fabric or the like to reproduce a target color, measurement of the target color or the color of the colorant, prediction of a mixing ratio, color prediction of a mixture, and the like are performed.
In computer color matching, data processing is performed using a computer algorithm based on color data of a colorant or the like to predict a mixing ratio of a colorant such as a dye and a color of the mixture.

[0003]

As a conventional color inspection apparatus, there is a conventional color inspection apparatus that measures R, G, and B for each pixel on a two-dimensional measurement plane including a large number of pixels using a camera measurement apparatus. A color inspection is performed by extracting RGB data of a necessary area from a large number of measured RGB data. But,
There is a problem that RGB data is affected by a light source or the like. Some conventional color matching apparatuses perform color matching and color inspection using one piece of reflectance data measured by a spectral colorimeter. Spectral reflectance is a material-specific quantity,
Not affected by light sources. The color inspection and color matching are performed by performing data processing on one reflectance. However, with this apparatus, only the data of the entire measurement surface is obtained, and therefore, it is not possible to perform processing for a required specific color in an image. Therefore, it cannot be used effectively for a material having a picture.

An object of the present invention is to provide an image processing apparatus and method capable of measuring and processing image data on a small target image.

[0005]

According to the present invention, there is provided an image processing apparatus comprising: a reflectance measuring unit for measuring a spectral reflectance of an object to be measured on a pixel basis; a spectral reflectance measured by the reflectance measuring unit; Display means for displaying an image composed of data;
For the image displayed on the display means, the image part to be extracted is specified by the user and the spectral reflectance data measured by the reflectance measuring unit is used to calculate the average spectral reflection in the image part specified by the specification means. Extraction means for determining the rate. Since the image portion can be specified in pixel units, it is possible to extract a necessary image portion in the image and extract the average spectral reflectance for a material having a picture or the like. The spectral reflectance is an amount specific to a substance, and the obtained average spectral reflectance is not affected by a light source or the like.

In the above image processing apparatus, for example,
The display unit includes a conversion unit that converts the reflectance data into a color specification value, and also displays the color specification value data for the image portion specified by the specification unit on the screen.

In the above image processing apparatus, for example,
In the specification of the partial image to be extracted, the specification means specifies a point or an inside or outside of a rectangle or a polygon.

In the above image processing apparatus, for example,
Further, there is provided extraction range designating means for designating an extraction range approximating the image portion designated by the designating means, wherein the extraction means calculates an average spectral reflectance in the designated extraction range. Preferably, the extraction range designating means includes a pixel whose color difference is within a range designated by the user with respect to the image part designated by the designating means, and a color difference within a predetermined range with respect to the image part designated by the designating means. The extraction range is specified by any of the pixels in the image data and the pixels binarized using the color specification value of the image portion specified by the specifying unit.

In the above image processing apparatus, for example,
The extraction means, from the spectral reflectance data measured by the reflectance measurement unit, the average of all the extraction range specified by the extraction range specifying means, the peak of the histogram of the extraction range specified by the extraction range specifying means, and The average spectral reflectance is determined as one of the average values of the pixels within a certain range of the histogram of the extraction range designated by the extraction range designation means.

An image processing method according to the present invention measures a spectral reflectance of an object to be measured on a pixel-by-pixel basis, displays an image composed of the measured spectral reflectance data, An image portion to be extracted is specified by the user, and an average spectral reflectance in a specified range is obtained from the measured spectral reflectance data. Preferably, further, an extraction range approximating the specified image portion is specified, and the average spectral reflectance is determined in the specified extraction range.

In the image processing method, spectral reflectance data is obtained for a plurality of substances, and the plurality of spectral reflectances are combined to approximate the average combined reflectance of the image portion. Thus, color matching can be performed using the combined spectral reflectance.

In the above image processing method, the area ratio of the pixels of the image portion to the entire image is obtained. Thus, color inspection can be performed.

[0013]

Embodiments of the present invention will be described below with reference to the accompanying drawings. In the drawings, the same reference symbols indicate the same or equivalent ones. FIG. 1 shows a configuration of a color matching system. This system is 2
-Dimensional spectrophotometer 10 and data processing device (computer)
22. The two-dimensional spectral colorimeter 10 irradiates a measurement target (sample) 14 with a light source 12 and
The reflected light from 4 is split by a spectroscope 16 and detected by a two-dimensional CCD sensor 18 of 5000 pixels. The spectral reflectance data is measured at intervals of 5 nm in a wavelength range of 380 to 760 nm. The area to be measured is 10 mm × 20 mm, and the area of one pixel is 0.2 mm × 0.2 mm. The spectral reflectance data of 5000 pixels thus measured is processed by the data processing unit 22 and sent to the data processing device 22 as digital data. The total number of obtained images is 5
000 pixels. The data processing device 22 is a computer including a display device and input devices such as a keyboard and a mouse.
The data processing device 22 extracts the reflectance of a pixel in a necessary area from various spectral reflectance data from the two-dimensional spectral colorimeter 10 by various methods according to a user's instruction from the input device, and The reflectance of the color specified in can be extracted. The result is displayed on a display device.
Further, data processing such as color matching (color matching), color inspection, and color composition identification (analysis) is performed using the extracted reflectance.

FIG. 2 shows one screen on the display device of the data processing device 22. On this screen, the measured image is displayed on the measured image display unit 30. This image is, for example, a multi-valued color image of a design. The measurement sample display section 32 displays a sample in the measurement image.
In this screen, the sample has not been set yet. The extracted color display unit 34 displays the color of the sample (extracted color), and the color distribution display unit 36 shows the extracted color and the histogram of the average L value, a value, and b value of the entire measured image. In this example, since the sample has not been specified yet, the color distribution display unit 36 displays only the average color distribution (histogram) of the entire measurement image. On the upper left portion of the screen, a masking condition button 40, a color designation condition button 42, a range designation condition button 44, an extraction method button 46, and a reflectance graph button 48 are displayed.

FIG. 3 shows a screen in a state where a sample is designated in the measurement sample display section 32. The measurement sample display section 32 displays a sample 33. The extracted color corresponding to the sample 33 is displayed in the extracted color display section 34
And the extraction value (L value, a value, b value) and the extraction range are displayed on the display unit 38. The sample 33 is a portion shown as a white region in the color image on the measurement sample display unit 32. The user operates the measurement sample display section 3
In step 2, the user operates the mouse to specify a portion to be sampled (points (multiple selections possible), rectangle, etc.). Designation is possible in pixel units. A pixel range with a similar color (reflectance) can be added to this designated portion by any of the methods described later. Sample 33 is the extraction range thus obtained. (That is, as can be seen by comparing the measurement sample display section 32 of FIGS. 2 and 3, the sample 3
Reference numeral 3 denotes a plurality of white portions newly appearing in the measurement sample display unit 32 in FIG. The color distribution display section 36 also shows a histogram 37 of the L value, a value, and b value of the extracted color.

FIG. 4 shows an example of the reflectance graph of the extracted color displayed by the reflectance graph button 48. In this graph, the vertical axis represents the reflectance, and the horizontal axis represents the wavelength of visible light.

Next, an image processing program of the data processing device 22 will be described. FIG. 5 is a flowchart of the image processing program. First, the data processing device 22 is caused to perform image color measurement, and receives image color measurement data (S1).
0). Here, a pixel of 10 mm × 20 mm area is used.
The reflection is measured in (0.2 mm × 0.2 mm) units. The image in this region is composed of spectral reflectance data of 5000 pixels in total. The obtained image data is displayed on the measurement image display unit 30 and the sample image display unit 32. Note that the spectral reflectance is calculated by a known formula using the spectral distribution of the illumination light.
Is converted to a color specification value. When converting to Lab value,
First, the reflectance is converted into a tristimulus value (XYZ) under a certain light source. Here, a D65 light source is used. Next, the tristimulus values are converted to Lab values by a color difference formula (Lab formula of CIE1976).

Next, masking is performed (S1).
2) In the sample image display section 32, an unnecessary image or a necessary image (pixel) is selected. That is, a target portion for performing the following processing is specified. Here, masking condition button 4
On the screen displayed by 0, select one of four types of masking designated ranges of point (multiple selections possible), rectangle, polygon, whole surface (no masking), and further, as conditions of designated range, inside or outside Is specified. When the inside is selected as the condition of the designated range, the masking designated range is left, and when the outside is selected, the masked designated range is deleted. Subsequent processing is performed only on the extraction target portion.

Next, for the image displayed on the sample image display section 32, an image portion of a color (reflectance) to be extracted is designated (S14). Here, when there are a plurality of colors (reflectance) to be extracted, an initial value of the color (reflectance) to be extracted is specified. The image (pixel) designation method is described in the color designation condition button 42
On the screen displayed by, any one of four types of points (a plurality can be selected), a rectangle, a polygon, and the entire surface can be selected. Furthermore, the inside or outside is designated as the condition of the designated range. The specified color (reflectance) becomes an initial value when the next extraction range is specified. Since the image is measured in pixel units, a small sample in the picture image can be specified.

Next, an extraction range is designated (S16). That is, a pixel range whose color (reflectance) is similar to the image (pixel) specified in step S14 is added. The range specification condition is displayed on a screen displayed by the range specification condition button 44, in which the color specification, the color difference range, the automatic binarization method are used.
You can select one of the types. If the designation condition is "color designation", only the part designated in step S14 is used. Therefore, step 16 ends without doing anything.

If the designated condition is "color difference range", the Lab value or L
The maximum value and the minimum value of the CH value are set as an extraction range, and pixels that fall within the range are added. The extraction range can be adjusted, and the user performs adjustment in one of the following three methods. (1) Direct input of Lab value or LCH value range (maximum value and minimum value). (2) On the Lab histogram or LCH histogram, specify the extraction range by moving the pick with the mouse. (3) Use the color difference range designation slider 50 to increase or decrease the extraction range and specify it.

If the designated condition is "automatic setting", an approximate color is automatically added to the image portion designated in step 14. Here, for a pixel whose color designation condition is designated by a point, rectangle, or polygon, the average L of designated pixels
The ab value (or the average LCH value) is converted to the reference Lab value (or
CH value). Pixels that fall from the reference Lab value (or the reference LCH value) to the search range ΔE or less set by the automatic color range setting condition are set as pixels to be automatically searched. However, the reference La
Pixels larger than the search range ΔE from the b value (or the reference LCH value) are excluded from the automatic search.

In the automatic search, the color difference between the target pixel and each selected pixel is calculated, and the pixels having the same color judgment ΔE or less set under the automatic color range setting condition are selected under the following conditions. Condition 1: When the same color area in the automatic color range setting condition is continuous, the same color determination is performed on the pixels (eight surrounding pixels) in contact with each of the already selected pixels, and the same color range is selected. The same operation is performed for each newly added pixel. This operation is repeated until no new pixels are added. Condition 2: When the same color range of the color range automatic setting condition is non-consecutive, the same color determination is performed on each of the already selected pixels and the remaining target pixels of the automatic search, and the same color range is selected. The same operation is performed for each newly added pixel. This operation is repeated until no new pixels are added. When the color range automatic setting condition is changed, the search range and the same color judgment are recalculated.

If the designation method is the "binarization method", the L value, a value, b
Value using one or more of the following values:
The range is selected by the value processing. Here, for a pixel whose color designation condition is designated by a point, a rectangle, or a polygon, the average Lab value (or the average LCH value) of the designated pixel is used as a reference L.
ab value (or reference LCH value), and 2
Perform value conversion. (1) In the case of binarization using the L value For the histogram, the L value at which σ _B ² = (M * ω _K −μ _K ) ² / (ω _K * (1−ω _K )) is the maximum It is calculated from the following equation. Here, K is increased in one step from 1 to (200 / total width of histogram). ω _K = ω _{K -1} + P _K , ω ₀ = 0 μ _K = μ _{K -1} + L _K * P _K , μ ₀ = 0 L _K = K * (total width of histogram) n _K = L _K the number of pixels n = total number of pixels _{P K = n K / n M} = Σ (L K * P K), K = 1, ( aggregate width of 200 / histogram) from Motoma' was K, L _K follows larger moiety Is divided into two pixels. Then, the pixels in the area including the reference L value are extracted. However, if η is smaller than the binarization determination index, all regions are extracted without being divided into two. ^{_{η = σ B 2 / σ T}} 2 σ T 2 = Σ (L K -M) 2 * P K, K = 1, (200 / histogram aggregate width)

(2) Binarization using a and b values K and L _K are replaced as follows, and pixels are similarly extracted. K is increased in one step from 1 to (200 / total width of histogram). L _K = K * (total width of histogram) −100

(3) Binarization using C value K and L _K are replaced as follows, and pixels are similarly extracted. K is increased in one step from 1 to (100 / total width of histogram). L _K = K * (histogram aggregation width)

(4) Case of binarization using H value K and L _K are replaced as follows, and pixels are similarly extracted. K is increased in one step from 1 to (360 / total width of histogram). L _K = K * (histogram aggregation width)

When performing the binarization processing on a plurality of L values, a values, b values, C values, or H values, the extraction pixel is determined based on the AND condition of each result.

Next, an extraction process is performed (S18). Here, the representative value of the spectral reflectance is determined from the extraction range of the color (reflectance) in the color range specified in step S16. Here, the average reflectance is determined. The method of averaging is the total average, La
b or average of the peak area of the LCH histogram, within the range of the histogram (± 0.5σ, ± σ, ± 2σ, ± 3
σ, the average of the peak half values). However, if the extraction range is smaller than a predetermined value (n pixels or less), only the total average is used.

There are three types of determination of the representative value of the spectral reflectance. However, when the number of designated pixels is n or less, the average of the entire extraction range is used as the representative value. (1) Average of all extraction ranges The average value of the color (reflectance) specified in step S16 is set as a representative value. (2) Histogram peak L value, a value of the color (reflectance) specified in step S16,
The average value of the pixels in the region where the number of pixels is the largest on the histogram of the b value, the C value, or the H value is set as the representative value. Here, the selection is performed using one or more of the histograms of the L value, the a value, the b value, the C value, and the H value. However, L value, a value, b
When a plurality of values, C values, or H values are selected, a pixel is selected based on an AND condition of each result.

(3) Range of Histogram The L value, a value, and L value of the color (reflectance) designated in step S16
In the histogram of the b value, the C value, or the H value, an average value of pixels included in a certain range is set as a representative value.

There are the following two types of range selection methods. (1) Selection using average value and standard deviation σ The average value and standard deviation σ are calculated from the histogram of L value, a value, b value, C value or H value, and the average value ± 0.5σ, σ,
The average value of pixels in a region falling within 2σ or 3σ is set as a representative value. Here, the selection is performed using one or more of the histograms of the L value, the a value, the b value, the C value, and the H value. However,
When a plurality of L values, a values, b values, C values, or H values are selected, a pixel is selected based on an AND condition of each result. (2) Selection using half width First, the L value, a value, b value, C value, or H value and the corresponding pixel that become the maximum number of pixels from the histogram of L value, a value, b value, C value, or H value Find the number. Next, the L value, a value, b value, C value, or H value that is half the number of pixels (half width) of the maximum number of pixels is changed to the L value, a value, b value, C value, or the maximum number of pixels. It is obtained by increasing or decreasing the value from the H value. Here, the selection is performed using one or more of the histograms of the L value, the a value, the b value, the C value, and the H value. However, L value, a value, b
When a plurality of values, C values, or H values are selected, a pixel is selected based on an AND condition of each result.

Next, the reflectance is checked (S20).
That is, with respect to the reflectance of each pixel obtained in step S18, the reflectance which is ± m% or more away from the average reflectance is cut as an abnormal value, and the average is calculated again. All wavelengths are checked, and pixels exceeding even one wavelength are cut off.
However, if all the pixels are cut, the original average reflectance is output. Then, the extracted spectral reflectance is output (S22).

Further, data processing such as color inspection, color composition identification (analysis), and color matching (color matching) is performed using the spectral reflectance (color) extracted in the above-described processing. For example, the color inspection is performed using the area ratio of the extracted pixels to the entire pixels.

The general procedure of color inspection is as follows. First, the spectral reflectance of the reference sample is measured by a colorimeter and registered. Also, the management frame of the reference sample is registered. Next, the spectral reflectance of the sample to be inspected is measured by a colorimeter. Next, the result is compared with the set value of the management frame to determine the color based on the area ratio. Then, it outputs a determination result such as good or bad.

When a desired color is reproduced by mixing a plurality of coloring agents such as dyes, color matching is performed. Here, one or two or more spectral reflectances of a colorant or the like are measured, and they are combined to obtain a combined reflectance. Is used to predict the mixing ratio of a coloring agent such as a dye, or to predict the color of a mixture.

The general procedure for color matching is as follows. First, the reflectance of a pigment or dye is measured by a colorimeter, and basic data of the pigment or dye is created. Then, the basic data of the created pigment or dye is registered. Next, the spectral reflectance of the target sample is measured by a colorimeter. Next, computer color matching calculation is performed using the spectral reflectance. Then, the obtained resipe (mixing ratio of pigment and dye) is output. Next, toning is performed according to the obtained resipe.

[0038]

Since the spectral reflectance can be measured for each pixel, the color of a minute sample can be extracted. Therefore, it is possible to specify and extract the color of the specific portion of the color pattern sample. In addition, since the spectral reflectance can be measured for each pixel, it is possible to specify a color in consideration of color unevenness. For example, the reflectance excluding the uneven portion is used. In addition, the reflectance excluding the shadow of the twill is used. Furthermore, a desired color (reflectance) can be extracted more accurately by adding an extraction range in which the color (reflectance) approximates. Since spectral reflectance is used, color matching, color inspection, and the like can be performed more accurately.

[Brief description of the drawings]

FIG. 1 is a block diagram of a color matching system.

FIG. 2 is a diagram of a colorimetric screen in the data processing device.

FIG. 3 is a diagram of a color extraction screen in the data processing device.

FIG. 4 is a diagram of reflectance data.

FIG. 5 is a flowchart of an image processing program.

[Explanation of symbols]

Reference Signs List 10 2D spectral colorimeter, 14 Measurement object (sample), 16 Spectroscope 16, 18 2D CC
D sensor 18, 22 Data processing unit 22, 3
0 measurement image display section, 32 measurement sample display section,
34 extraction color display section, 36 color distribution display section,
38 display part, 40 masking condition button,
42 color designation condition button, 44 range designation condition button, 46 extraction method button, 48 reflectance graph button, 50 input section.

Claims (10)

[Claims] 1. A reflectance measuring unit for measuring a spectral reflectance of an object to be measured on a pixel-by-pixel basis; a display unit for displaying an image composed of spectral reflectance data measured by the reflectance measuring unit; In the image displayed on the display unit, an image portion to be extracted is specified by a user, and an average spectral reflection in the image portion specified by the specifying unit is determined from spectral reflectance data measured by the reflectance measuring unit. An image processing apparatus comprising: an extraction unit for obtaining a rate. 2. The image processing apparatus according to claim 1, wherein the display unit includes a conversion unit that converts the reflectance data into a color specification value, and also displays the color specification value data for the image portion specified by the specification unit. The image processing device according to claim 1. 3. The method according to claim 1, wherein said specifying means specifies an inside or outside of any of a point, a rectangle, and a polygon in specifying a partial image to be extracted. Image processing device. 4. An image forming apparatus according to claim 1, further comprising an extraction range designating unit for designating an extraction range approximating the image portion designated by the designating unit, wherein the extraction unit calculates an average spectral reflectance in the designated extraction range. The image processing apparatus according to claim 1 or 2, wherein 5. The extraction range designating means includes: a pixel whose color difference is within a range designated by a user with respect to an image portion designated by the designation means; 2. The image processing apparatus according to claim 1, wherein the extraction range is specified by one of a pixel within the range and a pixel binarized using the color value of the image portion specified by the specifying unit. . 6. The extraction means, based on the spectral reflectance data measured by the reflectance measuring unit, an average value of all the extraction ranges designated by the extraction range designation means, and the extraction range designated by the extraction range designation means. 2. The average spectral reflectance is obtained as one of the peak of the histogram and the average value of the pixels within a certain range of the histogram of the extraction range designated by the extraction range designating means. Image processing device. 7. A method for measuring a spectral reflectance of an object to be measured on a pixel basis, displaying an image composed of the measured spectral reflectance data, and determining an image portion to be extracted from the displayed image by a user. An image processing method for calculating the average spectral reflectance in a specified image portion from the measured spectral reflectance data specified by. 8. The image processing method according to claim 7, further comprising specifying an extraction range approximating the specified image portion, and calculating an average spectral reflectance in the specified extraction range. 9. The image according to claim 7, wherein spectral reflectance data is obtained for a plurality of substances, and the plurality of spectral reflectances are combined to approximate the average combined reflectance. Processing method. 10. The method according to claim 7, wherein an area ratio of pixels of the image portion with respect to the entire image is obtained.
The image processing method described in 1.