JPH07113689A - Colorimeter - Google Patents

Abstract

PURPOSE:To provide a colorimeter which is practically useful, allows light intensity to enter an object whose reflection light spectrum change according to the ratio of contained constituent with a high S/N ratio and can classify color efficiently and quantitatively. CONSTITUTION:A colorimeter is provided with a light source 101 for generating illumination light for lighting an object to be measured, a light reception element for converting the reflection light from the object to an electrical signal, and a liquid crystal filler 206 which is installed on an optical path between the light source 101 and the object and has, as transmission characteristics, a differential vector between average spectra of the class with most diffferent color out of a plurality of classes determined based on the difference in the content of the spectrum constituent of reflection light.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is a color measuring device for measuring the difference in color caused by the ratio of contained components in the color measurement of products in various industries or the color measurement of a subject in the medical academic field. Regarding

[0002]

2. Description of the Related Art As a conventional color measuring apparatus, mainly the intensity of reflected light in the wavelength region representing the three primary colors of red (Red), green (Green) and blue (Blue) is measured, and the International Commission on Illumination ( A color coordinate system standardized by CIE, for example, an XYZ color system, an L ^* a ^* b ^* color system, or the like is used to display colors as numerical values to identify colors.

Such a color measuring device can express an object color as an absolute value by a color system which is said to be close to human sense. However, the color discrimination ability for detecting a slight difference in object color is due to the characteristics of the color system, and is ultimately determined by the R, G, B values. Therefore, such a color measuring device cannot clearly discriminate a difference in color to be discriminated, that is, a difference in reflection spectral spectrum for classification. For example, colors having different spectra within the wavelength region of the G color matching function cannot be classified by the G measurement value.

Further, as this type of apparatus, there is a multi-channel spectrophotometer which measures a reflection spectrum of an object and discriminates a color difference from the spectrum difference.
However, this device requires an expensive device such as a diffraction grating and a highly sensitive detector array, and when measuring a spectrum, the number of dimensions per data increases, so the size of the device for processing and analyzing it. Is large, the configuration of the device is complicated, and the cost is high.

On the other hand, a method of estimating the contained component ratio by measuring the reflected light intensity at a specific wavelength when the reflection spectrum changes depending on the contained component ratio in the object has been attempted. For example, in the medical field, obtaining the characteristic distribution of tissue hemoglobin on the surface of internal organs is considered to be important for pathological analysis of various diseases. Therefore, in the document “Sato,“ Optical Properties of Living Body ”, Medical Electronics and Biotechnology, 24, 22-27, (1986)”, the reflected light of a few specific wavelengths is analyzed from the analysis of the scattered reflection spectrum of organs such as stomach and duodenum. A method of estimating the oxygen saturation distribution of hemoglobin by measuring the intensity distribution has been proposed. That is, the absorption spectrum of oxygen-bound hemoglobin has a bimodal absorption peak at wavelengths 577 and 542 nm, and deoxygenated hemoglobin has a wavelength of 5
Since there is one absorption band at 55 nm, the oxygen saturation level of tissue hemoglobin is estimated by these peak wavelengths or a combination of these peak wavelengths and a wavelength of 569 nm that is not affected by oxygen saturation level.

However, in order to put such a method into practical use, it is necessary to measure the reflected light intensity of a specific wavelength by using an interference filter having a narrow bandwidth or a monochromator, but in that case, the input light energy is very weak. Therefore, the S / N of the measurement data is poor and quantitative measurement may be difficult.

In order to solve such a problem, in the document "Sato et al.," Spectral Diagnosis by Laser ", Artificial Organs, 16 , 1806-1810, (1987)", a laser having high wavelength selectivity and high brightness is used. A device using a light source has been proposed. However, in order to generate laser light tuned to a plurality of predetermined wavelengths, an expensive light source device such as a dye laser must be prepared, and it is difficult to construct a practically simple device.

[0008]

The conventional color measuring device described above measures the light intensity in the R, G, and B wavelength regions, and is therefore an optimum input method for the object to be classified. However, inconvenience may occur. Further, the multi-channel spectrophotometer is large in device size and expensive, and is not suitable for constructing a practical and simple device. On the other hand, a method of estimating the contained component ratio of the target object from the reflected light intensity at a characteristic wavelength of 2 to 3 requires measuring the light intensity in a very narrow wavelength region, and therefore, in order to realize it as a device. Has an S / N problem.

The color measuring apparatus of the present invention has been made in view of such a problem, and its purpose is to obtain a high S for an object whose reflected light spectrum changes depending on the ratio of contained components. Another object of the present invention is to provide a practically useful color measuring device capable of inputting the light intensity with / N and efficiently and quantitatively performing color classification.

[0010]

In order to achieve the above object, a color measuring apparatus of the present invention is provided with a light source for generating illumination light for illuminating an object to be measured, and reflected light from the object to be measured. Is installed on the optical path between the light source and the object to be measured and a light receiving element for converting into an electrical signal, of a plurality of classes determined based on the difference in the content rate of the spectral component of the reflected light. Among them, a color classification filter having a difference vector between average vectors of classes having the most different colors as a transmission characteristic is provided.

[0011]

That is, the color measuring apparatus of the present invention is arranged such that among the plurality of classes determined based on the difference in the content rate of the spectral component of the reflected light from the object to be measured, the average vectors of the classes having the most different colors are calculated. A color classification filter having a difference vector as a transmission characteristic is arranged on an optical path between a light source that generates illumination light for illuminating a measurement target and the measurement target to perform color measurement.

[0012]

Embodiments of the present invention will now be described in detail with reference to the drawings. First, the outline of the present embodiment will be described. The color measuring device of the present invention is a light source that generates illumination light for illuminating an object to be measured, a light receiving element that converts reflected light from the object to be measured into an electrical signal, the light source and the object to be measured. And a color classification filter installed on the optical path between the object and the object.

The transmission characteristic of this color classification filter is an example of spectral data of a plurality of samples prepared for each of a plurality of classes determined based on the difference in the content rate of the spectral component of the reflected light from the measurement object. Designed by a statistical method from the training set. That is, among the plurality of classes determined based on the difference in the content rate of the spectral component of the reflected light, the difference vector between the average vectors of the classes having the most different colors is used as the transmission characteristic.

That is, in the present embodiment, when the objects for color classification are limited, the statistical data as described above is applied to the spectral data examples (training set) of the samples belonging to each class for which the color classification is performed in advance. The optimum filter characteristic for color classification is derived by the method. Further, the difference in color handled in this embodiment is derived from the difference in the content rate of the spectral component, and the training set is prepared as a class roughly classified based on the difference in the content rate.

When color classification is performed, the filter control means arranges a color classification filter on the optical path,
With this color classification filter, wavelength characteristics suitable for classifying the reflected spectrum light of the measurement target are extracted. Therefore, according to the reflected spectrum light from the measurement object illuminated by the illumination light having such a wavelength characteristic, it is possible to perform quantitative color classification based on the content rate of the spectrum component.

According to such a method, the optimum filter characteristics are statistically derived with respect to the target of a specific color classification, so that efficient color classification can be performed. Further, since the characteristics of the color classification filter are practically realized by digitally calculating the measured values of the light intensity by several band filters, a device having a high S / N ratio can be constructed.

FIG. 1 is a block diagram of a color measuring apparatus according to a first embodiment of the present invention. This apparatus includes a light source box 100 for generating white light, a light source color creating apparatus 200 for converting white light generated by the light source box 100 into illumination light having a predetermined spectrum, and outputting the illumination light. Light guide 4 that guides the illuminated light to a specified location
00, a detector head 500 that illuminates an object with reflected light from the object and outputs the converted electrical signal, and an electric signal line 600 that transmits the output electric signal from the detector head 500. Light source color creation device 2
00 to send a command signal so that a predetermined light source color is created, and from the detector head 500 to the electric signal line 60.
A processor 3 for processing an electrical signal input via 0
00 and a user interface 700 for displaying a processing result at the same time when an operator's instruction is input to the processor 300.

FIG. 2 is a diagram showing the configurations of the light source box 100 and the light source color creating apparatus 200. Light source box 1
A light source 101 for a light bulb and a white lamp 102 are provided in 00, and white light generated by the white lamp 102 is collimated by a lens 103 and guided to a light source color creating apparatus 200.

In the light source color creating apparatus 200, the white light incident from the light source box 100 is condensed on the slit 202 by the lens 201, and the white light passing through the slit 202 is reflected by the concave mirror 203 and the diffraction grating 204.
Be led to. The diffraction grating 204 diffracts the white light in a specific direction with respect to the wavelength, and the diffracted light is reflected by the concave mirror 205 and condensed on the liquid crystal filter 206. The liquid crystal filter 206 has a configuration in which the liquid crystal filter driver 207 controlled by the processor 300 sets a predetermined position to a predetermined transmissivity so that the transmissivity can be set at an imaging position of each wavelength.

The illumination light which has passed through the liquid crystal filter 206 and converted into a predetermined light source color is condensed by the lens 208,
The light is guided to the light guide 400 composed of an optical fiber bundle by the optical splicer 209.

The detector head 500 is constructed as shown in FIG. The illumination light guided by the light guide 400 is emitted by the lens 501 at a wide angle. The reflected light from the target object is condensed on the detector 503 by the lens 502. The detector 503 has a photoelectric conversion element such as a photodiode, generates a current signal according to the intensity of incident light, and the current signal is amplified by the amplifier 504 and sent to the electric signal line 600.

FIG. 4 is a diagram showing the internal configuration of the processor 300. In the figure, the CPU 301 includes a ROM
303, CPU memory 302, A / D converter (hereinafter referred to as A
/ D) 304, statistical analysis processor 306, liquid crystal filter driver interface (I / F) 307,
The user interface (I / F) 308 and the external storage device interface (I / F) 305 are the internal bus 309.
Are connected to each other. The CPU 301 controls each component and controls the entire apparatus.

The ROM 303 stores an operating system (OS) and an execution program under the control of this OS, and the execution program is read and executed as necessary. CPU memory 3
Reference numeral 02 is a RAM, and programs and data are recorded when processing is executed. A / D 304 is the detector head 5
The output signal from 00 is digitally input.
The statistical analysis processor 306 is composed of a product-sum calculator, a memory, etc., and is a dedicated processor for performing matrix calculation at high speed by pipeline processing or the like.

A liquid crystal filter driver interface (I / F) 307 is an interface for sending a command signal for determining an address and a transmittance to the liquid crystal filter driver 207. A user interface interface (I / F) 308 is an interface for outputting to a display device for displaying a program menu and measurement results and for inputting a signal from a keyboard. An external storage device interface (I / F) 305 is an external storage device 310 that is used when recording programs and measurement results on a recording medium such as a hard disk or a floppy disk and reading the programs and data from them, that is, a hard disk driver or a floppy disk. It is an interface with the disk driver.

The operation of the above configuration will be described below. In this embodiment, there are two phases: statistical analysis for determining the characteristics of the classification filter and actual classification processing. First, a color classification filter suitable for color classification of a specific color can be designed as follows. A plurality of samples having each color (reflection spectrum light) to be classified are prepared for each class and statistical analysis is performed. First, create a training set for statistical analysis. This is designed so that the transmittance distribution of the liquid crystal filter 206 is maximized only in a specific wavelength region R ₁ as shown in FIG. Then, the sample is illuminated with illumination light that is close to monochromatic light from the light source color creation device 200 that has the liquid crystal filter 206 set in this way.

The intensity of reflected light from the sample illuminated by such illumination light is measured and recorded in the CPU memory 302 in the processor 300. Next, the above wavelength region R ₁
Move and repeat the same operation as above. The reflected light intensity of each wavelength region divided into n regions (n is about several tens) with respect to the set wavelength range is recorded. An example of spectrum data measured in this way is shown in FIG. The data example shown here is spectrum data obtained by measuring the reflected light intensity for a total of 11 wavelength regions at intervals of 25 nm over a wavelength range of 450 to 700 nm. Spectral data of each class is measured for all prepared samples to create a training set.

If the total amount of data becomes large, the data is temporarily recorded in the hard disk or floppy disk by the external storage device 310. A first method for designing a classification filter having spectral characteristics suitable for classifying a class from the training set created as described above will be described. It is assumed that a training set includes a series of samples in which the color changes little by little depending on the difference in the component ratio of the two color components.

The target color spectrum is the n-dimensional spectrum x =
The color change is represented by [x ₁ , x ₂ , ..., X _n ] ^t , and it is considered that the color change is observed by adding another color b, which is a factor of the change, to the reference color a.

X = αa + βb (1) where a and b are unit vectors. It is considered that the color change appears due to the change in the component ratio of the two color components. In the prepared training set, if the absolute intensity of the reference color a changes little between classes, b is found by finding the difference vector of the average vectors of the two classes with the most different colors, and this is approximated. It can be a classification standard.

[0030]

[Equation 1]

A conceptual diagram of the difference vector d is shown in FIG. The color component ratio of the vector b can be estimated by the inner product of the difference vector d and the reflected light vector x of the target object.

[0032] ^{x t · d = (αa +} βb) t · c · b = αca t b + βcb t b = α'a t b + β '... (5) where, d = cb, α' placed = αc, and β '= βc .

Since the value of α'a ^t b is almost the same for all data, the value of β'can be estimated as the value of β '.
The inner product of the equation (5) can be obtained by measuring the intensity of the reflected light that has passed through the color classification filter representing d.

FIG. 8 is a conceptual diagram showing the effect of the above first method. Note that in FIG. 8, the number of dimensions is 2 for simplicity,
The respective main axes are c ₁ and c ₂ . Next, a second technique for designing the color classification filter will be described. This method is a method that can be applied even when the absolute intensity of the reference color a is not constant due to a change in illumination light intensity. This method basically projects the spectrum vector x onto an arbitrary orthogonal complementary space of a and performs classification on the projection space.

[0035]

[Equation 2]

That is, in the prepared training set, a plurality of sample vectors {x ₁ } are measured for the spectral data example that is closest to a, that is, does not include the component of b most, by changing the illumination light intensity. Is subjected to principal component analysis (KL conversion). Then, the first principal component vector u ₁ becomes a vector substantially close to the vector a, and its eigenvalue becomes a variance value representing the variation in intensity.

[0037]

[Equation 3] It should be noted that r ≦ n is a rank when the sample spectrum {x ₁ } is subjected to KL conversion, and is a number smaller than the principal component number at which the eigenvalue is almost zero. If the spectrum vector x is defined as in (1), the projection onto the space A can be expressed as follows.

[0038]

[Equation 4]

That is, the component of the reference color a can be removed by the projection onto the space A, and the component intensity β of b can be estimated. Since the projection to the space A is basically an inner product calculation, it can be obtained by measuring the light intensity that has passed through a plurality of color classification filters representing each component vector of A as each component value. The effect of this method is shown in FIG. In FIG. 9, the spectral vector x expressed in the three-dimensional space is orthogonal to a
The figure shows the case of projection onto the dimensional space A.

[0040]

[Equation 5]

Next, a third method for designing a color classification filter will be described. In this method, the difference vector d described in the first method is rotated to make it an effective vector by color classification. Specifically, the Monte Carlo algorithm is used as shown below.

[0042]

[Equation 6]

(3) The d is rotated by a small angle α along the plane P in which d is arbitrarily set, and the variance v is newly calculated by the equation (7). (4) If v> v ₀ , the vector rotated by α is set again as d, and v ₀ = v. If v ≦ v ₀ , d is rotated along the plane P to the opposite side by a small angle α, and the variance v is calculated.

(5) Repeat (3) or (4) until the variance v becomes maximum. (6) Regarding the plane Q orthogonal to the plane P (3) to (5)
Do. (7) The classification vector d _p obtained for the plane P and the vector d _q obtained for the plane Q are vector-added to obtain a classification vector d _b .

D _b = d _p + d _q (8) The effect of the method described above is shown in FIG. Next, the phase in which the classification process is actually performed using the characteristics of the classification filter obtained by the method described above will be described. For example, in the case of performing color classification by the difference vector d, the transmittance distribution of the liquid crystal filter 206 is set to 2 of positive and negative components of the difference vector d.
It is sequentially set so that each type of spectral transmission characteristic is realized. The object is illuminated by the illumination light of the specific wavelength extracted by the two color classification filters set in this way, the reflected light at that time is converted into an electric signal, and the positive and negative of each reflected spectrum light is converted. Vector component z
^{+ And} z ^- are measured. The processor 300 subtracts the measured values of the positive and negative components of each color classification vector as described above. ^{z = z + -z - ... (} 9)

[0046]

[Equation 7] Further, it may be possible to quantitatively evaluate how much the b component is included in the vector based on the z value itself.

According to the first embodiment described above, the measurement of the training set, the design of the classification vector, and the actual color classification can be consistently performed by the same device having a relatively simple structure.

The second embodiment will be described below. In the second embodiment of the present invention, the light source color creating apparatus 200 is provided with a rotary filter in the main processing for color classification to make the configuration simpler. In this embodiment, the configuration related to the statistical processing phase for determining the characteristics of the classification filter is the same as in the first embodiment. FIG. 12 shows the internal configuration of the light source color creation device 200 in this processing for performing color classification. The white light emitted from the white lamp 102 in the light source box 100 is collimated by the lens 103, and the light source color creating apparatus 20
It is guided to the inside of 0, reflected by the mirror 210 and the mirror 211, and passes through the rotary filter 212.

The rotation filter 212 is controlled in rotation by a motor 213, and the motor 213 is controlled by the processor 300.
214 controlled by a command signal from the motor driver
Controlled by. In addition, the rotary filter 212 is shown in FIG.
As shown in, several color filters having a predetermined frequency transmission characteristic in the rotation direction are provided. These color filters are composed of colored glass filters dyed by adjusting the contained components and multilayer filters with metal coating on the surface.The transmission characteristics are similar to those of the color classification filters determined by pretreatment. Has been realized. For example, when the characteristic of the color classification filter is a difference vector d as shown in equation (3), the difference vector d has positive and negative characteristics as shown in FIG. It is composed of multiple band filters that approximately realize the characteristics. The light that has passed through the rotary filter 212 and converted into a predetermined frequency characteristic is condensed by the lens 215 and guided to the light guide 400 via the optical coupler 209.

The operation of the above configuration will be described below. In the present embodiment, the light source color creation device 200 using the diffraction grating 204 and the liquid crystal filter 206 as shown in FIG. 2 is used in the preprocessing for determining the characteristics of the color classification filter, but in the classification processing, as shown in FIG. Simple light source color creation device 20
0 is used. In the pre-processing, the liquid crystal filter 206 is used instead of the light source color creation device 200 as shown in FIG.
A normal spectroscope (monochromator) in which a slit is placed at the position and the installation angle of the diffraction grating 204 can be rotationally controlled may be used. In the color classification process, the reflected light from the object illuminated by the illumination light that has passed through the color filters provided in the rotary filter 212 is detected by the detector head 500 and digitally recorded by the processor 300.

When one object is illuminated with the illumination light from all the color filters and the reflected light intensity is detected, the processor 300 executes predetermined addition and subtraction between the detected light intensity values to perform classification evaluation. .

According to the second embodiment described above, when the target of color classification is limited in advance, the color classification process can be easily performed, the cost can be kept low, and a large number of objects can be processed. High speed processing is possible.

The third embodiment will be described below. This embodiment presents an endoscope apparatus as an example of configuring an image apparatus having a color classification function according to the principles of the present invention. The configuration of this embodiment is shown in FIG. The configuration is the outline, and the endoscope 10
00, a light guide 1100, a light source device and a processor 1200 including an image processing unit, and a TV monitor 1300. An illumination lens 1001 for illuminating the illumination light guided by the light guide 1100 is provided at the tip of the endoscope 1000, and an image of the illuminated object is also installed at the tip of the endoscope 1000. The imaging lens 1002 forms an image, and the CCD image pickup device 100 is formed.
3 is imaged.

The CCD image pickup device 1003 converts the image of the target object into a monochrome video signal and sends it to the processor 1200. Although a light source box 1201 and a light source color creating device 1202 are provided in the processor 1200, these are configured similarly to the light source box 100 and the light source color creating device 200 in the first and second embodiments. When the illumination light having a predetermined spectrum is created by the light source box 1201 and the light source color creation device 1202, and the image of the object is picked up by the illumination light, the video signal is converted into the A / D converter 1.
It is converted into a digital signal by 203, and the selector 12
04 through the frame memories 1205a to 1205m
It is recorded in a predetermined area of (m ≧ 3).

The above operation is repeated, and at most three image signals among the images recorded in the frame memories 1205a to 1205m are selected by the selector 1206.
It is sent to the matrix circuit 1207. Matrix circuit 1
At 207, at most three images are converted into predetermined color images, which are further converted into predetermined analog video signals by the video D / A converter 1208, and then the TV monitor 1300.
Is projected on. CPU 120 in processor 1200
Reference numeral 7 controls the operation of each component, such as instructing the spectrum of the illumination light created by the light source color creating device 1202, controlling the selectors 1204 and 1206, and sending the matrix calculation coefficient to the matrix circuit 1207. To do.

The operation of the above configuration will be described below. In the present embodiment, in addition to the function of displaying a normal RGB color image in the endoscope apparatus, a function of imaging a subtle difference in color depending on a specific symptom, for example, a characteristic distribution of tissue hemoglobin is added. . With respect to a plurality of cases prepared in advance, the characteristics of the filter for classifying a specific color are determined by the method described in the first embodiment. Then, in the actual diagnosis and examination, the following two modes of processing are performed.

First, in the RGB mode, the spectrum of the illumination light is sequentially converted into the characteristics of R, G, B, and the image of each color is recorded. This operation ends within 1/30 second of the video frame rate, and is displayed as an RGB color image every 1/30 second. Next, in the color classification mode, spectral light that realizes the filter characteristics of the color classification determined in advance is illuminated, and the image is recorded. Since the color classification filter is realized by a plurality of bands as shown in the first embodiment, the operation of recording the image of the number of bands in the frame memory is performed within 1/30 seconds of the video frame rate. The matrix circuit 120 is provided between the plurality of images.
A predetermined calculation is performed by 7, and the color classification information is imaged and displayed on the TV monitor 1300 as a monochrome image or a pseudo color image.

According to the third embodiment, by applying the principle of the present invention to an image device, the color classification information can be displayed as distribution information for the area displayed as an image. Therefore, it is possible to provide a device capable of visually grasping the situation of an object in which different colors are mixed.

[0059]

As described above, according to the present invention, it is possible to input the light intensity with a high S / N to an object whose reflection spectrum changes depending on the content ratio, and efficiently and quantitatively. It is possible to provide a practically useful color measuring device that performs color classification.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a color measuring device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a light source box and a light source color creation device shown in FIG.

FIG. 3 is a diagram showing a configuration of the detector head shown in FIG.

FIG. 4 is a diagram showing an internal configuration of a processor shown in FIG.

FIG. 5 is a diagram showing a design example of a transmittance distribution of a liquid crystal filter.

FIG. 6 is a diagram showing an example of measured spectrum data.

FIG. 7 is a conceptual diagram of a difference vector d.

FIG. 8 is a conceptual diagram showing the effect of the first method for designing a classification filter.

FIG. 9 is a conceptual diagram showing the effect of the second method for designing a classification filter.

FIG. 10 is a conceptual diagram showing the effect of the third method for designing a classification filter.

FIG. 11 is a diagram showing a configuration of a rotary filter.

FIG. 12 is a diagram showing an internal configuration of a light source color creation device.

FIG. 13 is a diagram showing a configuration of a third exemplary embodiment.

[Explanation of symbols]

100 ... Light source box, 101 ... Power supply for light bulb, 102 ...
White lamp, 103 ... Lens, 200 ... Light source color creating device, 201 ... Lens, 202 ... Slit, 203, 20
5 ... concave mirror, 204 ... diffraction grating, 206 ... liquid crystal filter, 207 ... liquid crystal filter driver, 208 ... lens,
209 ... Optical junction device, 300 ... Processor, 400 ... Light guide, 500 ... Detector head, 600 ... Electrical signal line, 700 ... User interface.

Claims (1)

[Claims] 1. A light source for generating illumination light for illuminating an object to be measured, a light receiving element for converting reflected light from the object to be measured into an electrical signal, and the light source and the object to be measured. Installed on the optical path between
A color classification filter having, as a transmission characteristic, a difference vector between average vectors of classes having the most different colors among a plurality of classes determined based on the difference in the content rate of the spectral component of the reflected light; Characteristic color measurement device.