JPS59770B2 - Shikisaigazo Hanteisouchi - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention processes signals from a photoelectric converter, such as a color television camera, to determine specific color areas in an image in real time.

Traditionally, photoelectric converters such as those used in color television cameras can detect red, green,
The color signal separated into the three primary colors of blue was once converted into a format suitable for communication systems, such as the NTSC system, and then on the receiving side it was separated into the three primary color components again to reproduce a color image.

Such conversion is an effective method for the characteristics of the human eye. However, when determining the colors in an image from the color signals obtained from a color television camera, there is no need to perform such conversion, and the color can be determined by directly processing each signal separated into several color components. It is possible. When determining the colors of a color image, special attention must be paid to the fact that the colors must always be determined stably despite differences in the brightness of the light source and the spectral reflectance and transmittance of the object. . SUMMARY OF THE INVENTION An object of the present invention is to provide an apparatus that stably performs color determination even when a light source such as light or an object changes.

In order to achieve the above object, the present invention calculates the sum of signals separated into several color components, normalizes each signal with the sum signal, and compares it with a predetermined value. The present invention provides a circuit that determines a specific color.

FIG. 1 shows the basic configuration of the present invention, and FIG. 2 shows the spectral characteristics of each element in FIG. 1 in the visible wavelength range Λ. Reference numeral 1 in FIG. 1 denotes a light source having spectral characteristics as shown in FIG. 2 a(7)P(λ), which illuminates the object 2. Target 2 has 2-1, 2-2,・
・゛ Each of 2-m has a color part with a different spectral reflectance (or spectral transmittance in the case of transmitted illumination), and as shown in the figure, R(x, y) for a point (x, y) on the object y, λ). 3-1 and 3-2 in Figure 1
,..., 3-n is a taking lens, 4-1, 4, -
2,..., 4-n are the spectral transmittances T1(λ), T2(λ),..., T3(λ) as shown in FIG. 2b, for example.
The color separation filters 5-1, 5-2, . . . , 5-n are devices for photoelectrically converting optical images by scanning,
As shown in , all photoelectric converters have the same spectral characteristic S(λ
). A color television camera is an example of a device system that performs such color separation, but the present invention is also effective in a single-tube color camera in which a taking lens and a photoelectric converter are unified. Fl, F2,..., F in Figure 1
n is a color signal obtained as a time series signal. The color signals corresponding to the point (X, y) on the object now are F1 (X, y), F2
(X, y), ..., Fn (X, y), and if the spectral transmittance of the lens is 100 (fl) in the visible wavelength region A, it can be expressed as follows. However, SjOl'221
\n Here k is a constant and a value unique to the optical system.

In Equation (1), the brightness of the light source or the spectral characteristics of the object is uniformly multiplied by a constant in the visible wavelength range A, that is, multiplied by a and b, respectively.
If it is assumed to be doubled, then the left side 2 of equation (1) will be multiplied by Ab for all j.

Therefore, if we define the value J-1, both the denominator and the numerator are A.
Since it is multiplied by b, the value of Fj (X, y) remains constant despite the above changes.

In other words, for a point (X, y) on the object, Fj (X, y
) is a value specific to the color of that point. Therefore, when determining the color of each point (X, y) on the object, for example, the constant value αIj〈1 (1=
1, 2, ..., m; If j is also determined to be a color other than ゛i''゜.

FIG. 3 is an example of the above color determination in the case of m=n=3 in FIGS. 1 and 2.

In other words, blue, green, and red exist as colors in order of shortest wavelength in parts 2-1, 2-2, and 2-3 on object 2, and the spectral characteristics T1 (λ) and T2 (λ ), T3(λ)
If blue, green, and red have the maximum transmittance in order of shortest wavelength, then the product P(λ
)S(λ) has a peak in the center of the visible wavelength range (i.e. near green) as shown in Figure 2a, so F,
, F2F3. From FIG. 3, ΣFj is minimum in the black part j-1 minute 2-a and maximum in the white part 2-b.

Therefore, by calculation such as equation (2), (Fl, f2,
f3) is the blue part 2-1 (1,0,0) and the green part 2
-2 is (0, 1, 0), and the red part is (0, 0, 1), which is an unchanging value unique to each color as shown in the figure despite changes as shown by the broken line. In reality, when making a judgment like equations (3) and (4), if it is for all j instead of equations (3) and (4), it is judged that the color is ゛i゛, For a certain j, it may be determined that the color is not i.

FIG. 3 shows color determination for blue. By doing the above, it is possible to always perform stable color determination even when the brightness of the light source or object changes uniformly. Furthermore, according to this color determination, it is possible to ignore the influence of differences in the spectral characteristics of the light sources, and there is no need to define a standard white color. Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 4 shows an embodiment of the present invention. As shown in the figure, the signals from the photoelectric converters 5-1, 5-2, . 6 is an adder circuit ΣF1(t
), 7-1,71=! -2, . . . , 71n is a division circuit which calculates j-1 corresponding to equation (2) for all j as shown in the figure for each t.

8-1, 8-2, . ..., 8-n is a color component determination circuit,
If the color to be sought is "l", then output the logical value 61'', otherwise output the logical value ``O''.

Reference numeral 9 denotes an AND circuit, which outputs 1P only when the outputs of the color component determination circuits 8-1, 8-2, . . . 8-n are all 1.

By doing so, it is possible to identify the point corresponding to each time t as "1" if the color is "i", and as "60" otherwise. FIG. 5 shows the color component determination circuit 8-1 of FIG.
This is an example of 8-2, . . . , 8-n.

In the figure, 10 is an addition circuit, and 11 is a subtraction circuit, which calculates α1, +εj, and αIj−εj, respectively. 13,13
is a comparison circuit, and as shown in the figure, it outputs "1" from 12 if Fj (TKaij + εj), "1" from 13 if Fj (t)>αIj - εj, and 00゛ otherwise. .

14 is an AND circuit, and 12 and 13 are both T゛, that is αI
It is designed to output 61'' when j−εj<Fj(t) αIj+ε·.

FIG. 6 shows another embodiment of the invention.

15-1, 15-2, . . . 15-n are multiplication circuits for multiplying the signal Fi(t) by α1, (×1).

Since αIj is constant, there is no need to multiply it at each time,
Practically speaking, this can be realized using an attenuation circuit with a variable attenuation factor.
8-1, 8-2, ... 8-n is a color component determination circuit,
At each time t, "1" is output, and at J4, "O" is output.

Each color component determination circuit may be the same as that explained in FIG. 5, and the input signals Fj(t) and αIjlεj may be replaced with Fj(t), αIjFj(t), and δj, respectively. 'Here, δ・ may be a constant, but formula (3-1)
As is clear from Equation (3-2)', it is also possible to set n δ,=ε,/ΣF, (t).

Therefore, if the point corresponding to each time t from the AND circuit 9 has the color 61', it can be determined as "1", otherwise it can be determined as "O".

In the color component determination circuits in the first and second embodiments, a circuit that can be controlled by a computer can be constructed by using digital registers and digital-to-analog converters for externally applied inputs.

Further, any of the embodiments can be realized with a simple circuit configuration. Further, in the case of the second embodiment, since there is no division operation compared to the first embodiment, the color of each point of a color image can be determined at high speed in real time, and it has high practical value.

As described above, according to the present invention, only a portion of a specific color can be output as an image without being affected by changes in the brightness of the light source or object.

Therefore, it is possible to provide an effective signal for more advanced image processing such as measuring the area of a part of a certain specific color or determining its shape.

[Brief explanation of drawings]

FIG. 1 to FIG. 6 are explanatory diagrams of the present invention.

Claims (1)

[Claims] 1. Photoelectrically converting input light having a color to be determined by each of a plurality of photoelectric conversion means having mutually different spectral characteristics, and generating a second signal by adding the plurality of first signals obtained by the photoelectric conversion. and a plurality of third signals each having a value obtained by multiplying the second signal by a plurality of values given in advance.
detect whether each of the third signals and a corresponding one of the first signals match;
1. A color determination method, comprising determining whether or not the input light has a color defined by the plurality of predetermined values based on whether or not there is a match for all of the signals.