JP2011002287A - Method for obtaining color value from spectral data and colorimeter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a colorimeter for combining advantages of a spectral colorimeter and a stimulus-value directly reading colorimeter.SOLUTION: A light source for singularly changing a spectral characteristic and a color patch having a known color value are prepared (S1). While the spectral characteristic of the light source is gradually changed, the color patch is irradiated with light from the light source (S2). An optical sensor is provided with one, two, three or four filters, and detects a reflection light from the color patch. A spectral data are sampled at each step of a change in the spectral characteristic (S3). A transform matrix for deriving the corresponding color value is obtained from the obtained spectral data by using an optimal solution estimation method (S4). An object to be measured is irradiated with light from the same type of the light source as the light source used for obtaining a transform matrix (S5). An optical sensor having the same spectral characteristic as the optical sensor used for obtaining the transform matrix detects reflection light from the object to be measured(S6). The color value of the object to be measured is obtained from the obtained spectral data by using the transform matrix (S7).

Description

  The present invention relates to a method for obtaining a chromaticity value from spectral data and a colorimeter for carrying out the method.

There are two types of colorimeters: a stimulus value direct-reading colorimeter and a spectral colorimeter (see, for example, Patent Documents 1 and 2).
The stimulus value direct reading method is a method in which the reflected light from the measurement object is dispersed through a filter corresponding to red, green, and blue, and each light quantity is measured by a sensor to directly measure tristimulus values X, Y, and Z. It is. On the other hand, in the spectroscopic method, the reflected light from the measurement object is dispersed through a spectral filter set at a predetermined wavelength interval, for example, a wavelength of 10 nm, and the reflectance for each wavelength is measured by a sensor. This is a method of calculating the tristimulus values X, Y, and Z by calculating with a microcomputer based on the above.

  Stimulus value direct-reading colorimeters are relatively inexpensive, have the advantages of being compact, excellent in mobility, and capable of easily measuring tristimulus values, but have low measurement accuracy, color matching conditions, etc. It is not suitable for advanced color analysis. On the other hand, spectral colorimeters require spectral reflectance at each wavelength, so they can be measured with high accuracy, but are equipped with spectral filters, etc., compared to stimulus value direct-reading colorimeters. Is complex, large, and expensive.

Japanese Patent Laid-Open No. 9-72786 JP 2004-163314 A

  An object of the present invention is to provide a colorimeter having the advantages of both a spectroscopic colorimeter and a stimulus value direct reading colorimeter.

  In order to solve the above problems, according to the present invention, (1) a light source capable of changing spectral characteristics by itself and at least one color patch to which a predetermined chromaticity value is assigned are prepared, (2) Light sensor irradiates light to the color patch from the light source while gradually changing the spectral characteristics of the light source, and (3) an optical sensor provided with 1 to 4 types of filters for the reflected light from the color patch And spectral data for each filter is sampled at each stage of change of the spectral characteristics, and (4) the corresponding chromaticity value is derived from the sampled spectral data using the optimum solution estimation method. A method for obtaining a chromaticity value from spectral data characterized by obtaining a transformation matrix is provided.

According to a preferred embodiment of the present invention, further, (5) the measurement object is irradiated with light from the same type of light source as the light source, and (6) the reflected light from the measurement object is reflected in the same spectrum as the optical sensor. (7) A chromaticity value of the measurement object is obtained from the detected spectral data for each filter using the conversion matrix.
According to another preferred embodiment of the present invention, the optimal solution estimation method is a multiple regression analysis method or a neural network or a genetic algorithm.

  Further, according to the present invention, a colorimeter for carrying out the above method according to the present invention, which irradiates a measurement object with light, the same type of light source as that used when obtaining the conversion matrix, Detecting reflected light from the measurement object, an optical sensor having the same spectral characteristics as the optical sensor used when obtaining the conversion matrix, a data processing unit including the conversion matrix, and a data display unit. The data processing unit obtains the chromaticity value of the measurement object from the spectral data for each filter detected by the optical sensor having the same characteristic, using the conversion matrix, and displays the chromaticity value in the data display Provided is a colorimeter characterized by being displayed on a screen.

  According to the present invention, while changing the spectral characteristics of the light source stepwise, the color patch is irradiated with light, the reflected light is detected by the optical sensor, the spectral data is sampled, and the chromaticity based on the acquired spectral data A transformation matrix for deriving values is obtained. In this case, the output of the light sensor is the integrated value of the spectrum of the light emitted from the light source, the spectral reflectance of the color patch (measurement target), and the spectral sensitivity of the light sensor. By changing it, a plurality of different outputs can be obtained from the optical sensor.

According to the present invention, a large number of spectral data can be sampled only by using an optical sensor having four types of filters, and a highly accurate conversion matrix can be obtained using these spectral data. By using this conversion matrix, the chromaticity value can be obtained from the spectral data with higher accuracy than the conventional stimulus value direct-reading colorimeter.
In addition, according to the present invention, the colorimeter is only equipped with an optical sensor having four types of filters, so that it is as small and small as a conventional colorimetric colorimetric device that directly reads a stimulus value. It has a simple configuration and can be manufactured at a cost comparable to that of a conventional colorimetry that directly reads the stimulus value.

It is a flowchart of the method of calculating | requiring chromaticity value from the spectral data by this invention. It is the graph which showed the relationship between the average value of the estimation error of the chromaticity value estimated using the conversion matrix, and the number of patterns of the applied voltage at the time of calculating | requiring the conversion matrix. It is a block diagram which shows roughly the structure of the colorimeter by this invention.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is a flowchart of a method for obtaining a chromaticity value from spectral data according to the present invention. Referring to FIG. 1, according to the present invention, first, a light source capable of changing spectral characteristics by itself and at least one color patch provided with a predetermined chromaticity value are prepared (S1 in FIG. 1). ). A light source that can change the spectral characteristics by itself means a light source that can change the spectral characteristics only by changing the supply voltage to the light source without using additional means such as a filter. Examples of such light sources LED (regardless of emission color) is mentioned.

  According to the present invention, next, while changing the spectral characteristics of the light source stepwise, light is emitted from the light source to the color patch (S2 in FIG. 1), and 1 to 4 types of reflected light from the color patch are used. The spectral data for each detected filter is sampled at each stage of change in spectral characteristics (S3 in FIG. 1). As an optical sensor provided with 1 to 4 types of filters, for example, an optical sensor provided with 3 types of filters of red (R), green (G), and blue (B), cyan (C), and magenta (M). An optical sensor including three types of yellow (Y) filters can be used. Alternatively, an optical sensor provided with only one type of filter of red (R), green (G), and blue (B) can be used. In addition, an optical sensor comprising a combination of four sensors: a sensor with a cyan (C) filter, a sensor with a magenta (M) filter, a sensor with a yellow (Y) filter, and a sensor without a filter. It can also be used.

  Then, a conversion matrix for deriving the corresponding chromaticity value is obtained from the sampled spectral data using the optimum solution estimation method (S4 in FIG. 1). As the optimal solution estimation method, a multiple regression analysis method, a neural network, or a genetic algorithm can be used.

  According to the present invention, the measurement object is further irradiated with light from the same type of light source as that used for obtaining the conversion matrix (S5 in FIG. 1), and the reflected light from the measurement object is converted into the conversion matrix. Is detected by an optical sensor having the same spectral characteristics as the optical sensor used to determine the value (S6 in FIG. 1), and the chromaticity value of the measurement object is determined using the conversion matrix from the detected spectral data for each filter (FIG. 1 S7). In this case, the light source of the same type as the light source used when obtaining the conversion matrix means a light source having the same model number as the light source used when obtaining the conversion matrix.

[Example]
Next, a conversion matrix was actually obtained according to the method of the present invention, and the measurement accuracy by the obtained conversion matrix was examined. In this case, a commercially available white LED is used as a light source, and a set of three color sensors each including a red (R) filter, a green (G) filter, and a blue (B) filter are used as light sensors, As the color patches, 24 types of color patches having known chromaticity values (X, Y, Z) as shown in Table 1 were used.

  Then, by changing the voltage applied to the white LED in three stages, the spectral characteristics of the white LED are changed in three stages, and for each color patch, the sensor output (spectrum) for each color patch is changed. Data). The acquired spectral data is shown in Table 2. In Table 2, R0, G0, B0 are spectroscopic data when applied voltage at the first stage, and R1, G1, B1 are spectroscopic data when applied voltage at the second stage are R2, G2, B2 represents spectroscopic data when applied voltage at the third stage.

First, a transformation matrix for deriving chromaticity values from spectral data was obtained using multiple regression analysis based only on the spectral data acquired when the applied voltage was the first stage.
In the multiple regression analysis method,
φ = m · ω
This results in the problem of obtaining the transformation matrix m.
At this time, the transformation matrix m is
m = φ · ω ^t · (ω · ω ^t ) ⁻¹
And the transformation matrix is

と表される。

It is expressed.

  Furthermore, using the same white LED and the same light sensor as before, the same color patch as before is irradiated from the white LED (the applied voltage is arbitrary), and the reflected light from each color patch is detected by the light sensor and acquired The chromaticity value of each color patch was estimated using the conversion matrix obtained from the spectral data obtained above. Then, an error (mean square error (RMS)) between the estimated value and the known chromaticity value of each color patch was examined. The results are shown in Table 3.

  In Table 3, the average value of the estimation error is 1.7692, which is almost the same as the measurement error of the conventional stimulus value direct-reading colorimeter.

を満たす変換行列Ａを、 Next, using the spectral data acquired in all three patterns of applied voltage in the first stage to the third stage, similarly to the above, using the multiple regression analysis method,

A transformation matrix A satisfying

として求めた。

As sought.

  Then, using the same white LED and the same light sensor as before, irradiate light from the white LED to the same color patch as before (applied voltage is arbitrary), and the reflected light from each color patch is detected by the light sensor and acquired The chromaticity value of each color patch was estimated using the conversion matrix obtained from the spectral data obtained above. Further, the error (mean square error (RMS)) between the estimated value and the known chromaticity value of each color patch was examined. The results are shown in Table 4.

  In Table 4, the average value of the estimation error is 1.464284, which is smaller than the average value of the estimation error shown in Table 2. Therefore, the measurement error is larger than that of the conventional stimulus value direct-reading colorimeter. You can see that it is getting smaller.

  Further, in the same manner as described above, after obtaining the conversion matrix using the spectral data acquired in the two patterns of applied voltage in the first stage and the second stage, the same white LED and the same light sensor are used, Light is emitted from the LED to the same color patch (the applied voltage is arbitrary), the reflected light from each color patch is detected by an optical sensor, and the conversion matrix obtained above from the acquired spectral data is used to determine the color patch. The chromaticity value was estimated and the error (mean square error (RMS)) between the estimated value and the known chromaticity value of each color patch was examined.

And, for the three different transformation matrices obtained above, plot the average value of the estimation error of the chromaticity values estimated using the transformation matrix against the number of applied voltage patterns when obtaining the transformation matrix. Thus, a graph as shown in FIG. 2 was obtained.
From the graph of FIG. 2, the accuracy of the conversion matrix improves as the number of applied voltage patterns and thus the number of steps for changing the spectral characteristics of the white LED increases and the number of samplings (spectral data) increases. I understand that

  According to the present invention, while changing the spectral characteristics of the white LED stepwise, the color patch is irradiated with light, the reflected light is detected by the optical sensor, the spectral data is sampled, and the color is determined based on the acquired spectral data. Since the conversion matrix for deriving the degree value is obtained, it is possible to sample a large number of spectral data by using an optical sensor equipped with four types of filters, and a highly accurate conversion matrix based on the spectral data. Can be obtained. By using this conversion matrix, the chromaticity value can be obtained from the spectral data with higher accuracy than the conventional stimulus value direct-reading colorimeter.

  FIG. 3 is a block diagram showing a schematic configuration of a colorimeter for carrying out a method for obtaining a chromaticity value from spectral data according to the present invention. Referring to FIG. 3, the colorimeter according to the present invention irradiates light to be measured 6 with light source 1 of the same type as the light source used when obtaining the conversion matrix and reflected light from measurement object 6. An optical sensor 2 having the same spectral characteristic as that of the optical sensor used for obtaining the conversion matrix to be detected, a data processing unit 3 including the conversion matrix, and a data display unit 4 are provided.

In this case, the light source of the same type as the light source used when obtaining the conversion matrix means a light source having the same model number as the light source used when obtaining the conversion matrix.
The optical sensor 2 may be an optical sensor having the same spectral characteristics as the optical sensor used when obtaining the conversion matrix, and need not be the same optical sensor as used when obtaining the conversion matrix. However, since the optical sensor used for obtaining the conversion matrix is an optical sensor having 1 to 4 types of filters, the optical sensor of this colorimeter is also an optical sensor having 4 types of filters at most. Composed. In this embodiment, the optical sensor 2 includes color sensors 4a to 4c each including a red filter 3a, a green filter 3b, and a blue filter 3c.

  Then, the data processing unit 3 obtains the chromaticity value of the measurement object 5 from the spectral data for each of the filters 3 a to 3 c detected by the optical sensor 2 using the conversion matrix, and the chromaticity value is displayed on the data display unit 4. It is supposed to be displayed.

  The colorimeter according to the present invention is equipped with at least an optical sensor having four types of filters, and thus has a small and simple configuration comparable to that of a conventional stimulus value direct-reading colorimeter. In addition, it can be provided at a low price comparable to that of a conventional colorimetry that directly reads stimulus values.

DESCRIPTION OF SYMBOLS 1 Light source 2 Optical sensor 3a-3c Filter 4a-4c Color sensor 5 Data processing part 6 Data display part 7 Measurement object

Claims (4)

(1) preparing a light source capable of changing spectral characteristics by itself and at least one color patch provided with a predetermined chromaticity value;
(2) irradiating the color patch from the light source while gradually changing the spectral characteristics of the light source;
(3) The reflected light from the color patch is detected by an optical sensor having 1 to 4 types of filters, and spectral data for each filter is sampled at each stage of change in the spectral characteristics,
(4) A method for obtaining a chromaticity value from spectral data characterized by obtaining a transformation matrix for deriving the corresponding chromaticity value from the sampled spectral data using an optimal solution estimation method. (5) irradiating the measurement object with light from the same type of light source as the light source,
(6) The reflected light from the measurement object is detected by an optical sensor having the same spectral characteristics as the optical sensor,
(7) The chromaticity value of the measurement target is obtained from the detected spectral data for each filter using the conversion matrix.   3. The method according to claim 1, wherein the optimum solution estimation method is a multiple regression analysis method, a neural network, or a genetic algorithm. A colorimeter for carrying out the method according to claim 2, comprising:
A light source of the same type as the light source used to determine the transformation matrix, which irradiates the measurement object with light;
A photosensor having the same spectral characteristics as the photosensor used to determine the conversion matrix, detecting reflected light from the measurement object;
A data processing unit comprising the transformation matrix;
A data display unit,
The data processing unit obtains the chromaticity value of the measurement object from the spectral data for each filter detected by the optical sensor having the same characteristics, using the conversion matrix, and the chromaticity value is stored in the data display unit. A colorimeter characterized by displaying.

Images