JPS6127955B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a color solid-state imaging device,
It is an object of the present invention to realize a wide band of luminance signals by eliminating the phenomenon in which the amplitude differs from horizontal line to horizontal line, that is, the brightness differs from horizontal line to horizontal line on a television screen, that is, line shading, with a simple configuration.

Image pickup tubes such as vidicon and plumbicon are generally used to convert an optical image of a subject into a television signal, but in recent years, solid-state image pickup devices have been developed and are finally being put into practical use. Examples of solid-state imaging plates include a combination of a photosensitive element such as a photodiode and a charge transfer element, and a photodiode array in which photosensitive elements are arranged in a grid pattern and sequentially read out. Regarding the charge transfer device, as explained in Japanese Patent Application Laid-open No. 1211/1983 as a charge coupled device (CCD), the charge transfer device is transferred to a 2-phase and 3-phase region through an insulating layer on a semiconductor substrate.
For example, if the semiconductor substrate is an N type, an electric wire wired in a phase configuration is provided, and by applying a negative charge to this electrode, an air space layer is created in the semiconductor, and the potential of this electrode is sequentially moved. As a result, the depletion layer also moves. If charges are injected into this depletion layer, the charges will also move as the depletion layer moves. Therefore, for example, a photodiode and this
All you have to do is combine a CCD, send the charge generated in the photodiode by light to the CCD, and then sequentially move the charge inside the CCD and take it out. An example of a solid-state imaging plate using this principle will be explained with reference to FIG.

In Figure 1, 11, 12, 13......, 1
N denotes a CCD that transfers charge in the vertical direction, and a transfer pulse for driving this is supplied from a vertical transfer pulse generator 2. 3 is a photosensitive element such as a photodiode, which is installed along the vertical transfer CCD, generates and accumulates charges according to the intensity of light, and after a certain period of time, simultaneously injects the accumulated charges into the vertical transfer CCD. do. The injected charge is transferred sequentially in the direction of the arrow, and then horizontally.
The signal is sent into the CCD 4, transferred in the direction of the arrow by the drive pulse from the horizontal transfer pulse generator 5, and taken out from the output terminal 6 as a continuous video signal. In this case, the transfer frequency of the horizontal transfer CCD
_H = _HD ×N (1 ₎ [Here, _HD is the horizontal scanning frequency, and N is the number of light-receiving elements constituting one horizontal scanning line. ] It is expressed as .

When obtaining a color television signal, three such image pickup plates are used, and a dichroic mirror separates the subject image into three primary colors, and each image pickup plate receives this light to generate electrical signals corresponding to the three colors. A common method is to obtain

The principle of this method is the same as that of a three-tube color television camera system using three image pickup tubes. However, unlike the case of a three-tube color television camera, extremely precise mechanical adjustment is required for registration adjustment, and in addition, adjustment is required in three locations: horizontal, vertical, and rotational. . As is well known, image pickup tubes obtain signals by deflecting and scanning an electron beam onto a photoconductive film using a magnetic field or an electric field. Therefore, registration adjustment can be performed using a direct current method based on a single image pickup tube. The objective is achieved by moving the position of the entire image by applying a strong magnetic or electric field, but the image pickup plate can only be adjusted mechanically, and furthermore, the amplitude cannot be adjusted as it is. In order to simplify or omit this adjustment location, a two-imaging plate system or a single imaging plate system can be considered.

FIGS. 2 and 3 show an example of a commonly considered single image pickup plate color camera system. Figure 2 shows a striped color filter 27,2
8 and 29 are color filter elements that pass red, blue, and green color lights, respectively. By placing this striped color filter overlappingly on the image pickup plate shown in FIG. 1, electrical signals corresponding to red, blue, and green are obtained point-sequentially as output signals. This output signal is converted into a color television signal by the processing circuit shown in the block diagram of FIG.

In FIG. 3, numeral 30 denotes a solid-state imaging plate on which the striped color filters shown in FIG. The output signal is transmitted to the amplifier 32
The signal is applied to the sampling circuit 33 via the . The pulses synchronized with the horizontal transfer pulses from the clock pulse generator 31 are shaped into the phase and width of each color signal through a waveform shaping circuit 34, and then sent to a sampling circuit where they are separated into each color signal. This separated signal is applied to a color encoder 35 to obtain a standard color television signal.

The bandwidth of the luminance signal according to this method will be explained using FIG. 4. When a monochrome subject is imaged using this method, if the output signals of the light-receiving elements through each color filter are made almost equal, the luminance signal bandwidth will be equal to the number of light-receiving elements, that is, a frequency equal to 1/2 of the horizontal clock frequency A. It has a bandwidth of up to B. However, it is necessary to insert a spatial frequency filter in the optical path to prevent false color signals and attenuate the repetition frequency of the striped color filter, that is, around 1/3 frequency D of the horizontal clock frequency, in terms of spatial frequency. The bandwidth is 1/3 of the horizontal clock frequency in order to remove frequency components spatially modulated by a striped color filter for use as a luminance signal using a low-pass filter, trap, etc.
narrower than the frequency. Furthermore, when capturing an image of a subject with high color purity, the effective number of elements is reduced to 1/3, so the luminance signal bandwidth becomes even narrower to 1/6 frequency C of horizontal clock frequency A, resulting in a significant drop in resolution. cause

Note that the dotted line in FIG. 4 is the overall luminance signal band characteristic when a monochrome object is imaged, and the solid line is the overall luminance signal band characteristic when an object with high color purity is imaged. Also shown in FIG. 5 are the characteristics of a spatial frequency filter using a birefringent crystal filter. Therefore, in order to improve resolution using this method, it is necessary to increase the number of elements in the horizontal direction, which requires increasing the degree of integration or increasing the imaging area. However, such a solution has many problems in terms of manufacturing technology, and therefore, a color system is desired that allows an image pickup plate with as small an area and a small number of elements as possible.

For this reason, Japanese Patent Application No. 50-128752 is proposed as an example of a method that can effectively improve resolution with a smaller area and fewer elements than conventionally considered methods.
The outline of its contents will be explained below.

FIG. 6 shows a mosaic arrangement of color filters, in which 67 is red, 68 is blue, and 69 is a filter element with luminance characteristics or total light transmission. In this case, combinations such as red and cyan, or blue and a different color may be used. If you arrange it like this, each horizontal line will have columns of R, Y, R, Y... and B, Y, B, Y...
A column of point sequential signals is obtained. This signal is illustrated in FIG. 7. In Figure 7, A
indicates the column R, Y, R, Y..., B indicates B,
The columns Y, B, Y... are shown. FIG. 7 shows an image pickup output signal when a magenta (red+blue) color object is imaged in which the blue component is twice the size of the red component. At this time, each color filter R (red), B
(blue), W (all-color transmission) filter. W
= (R+G+2B), so each output signal is R,
2B, Y=(R+2B), and in Figure 7, Y
The signal is the loudest.

A block diagram of a signal processing system for converting this into a standard color television signal is shown in FIG. In FIG. 8, reference numeral 80 denotes a solid-state image pickup plate equipped with the mosaic color filter shown in FIG. The components are removed to produce a luminance signal. On the other hand, the other signal obtained from the solid-state image pickup plate 80 is passed through a bandpass filter 83 whose center is 1/2 of the horizontal drive frequency, and the high frequency component modulated by a mosaic color filter is separated and synchronously detected. A color difference signal (R-Y) or (B-Y) is obtained by synchronous detection in a circuit 84, one directly through a low-pass filter 85, and the other through a one horizontal period delay line 86 to a switch circuit 87. In addition, the line-sequential color difference signal is converted into two consecutive color difference signals and then supplied to an encoder 88 for conversion into a standard color television signal. A pulse generator 89 applies this signal to the solid-state image pickup plate to drive the solid-state image pickup plate, and also supplies pulses in a synchronous relationship with the solid-state image pickup plate to the sampling circuit 81 and the synchronous detection circuit 84. In this way, the bandwidth of the luminance signal will always be 1/4 of the horizontal drive frequency and 1/2 of the sampling frequency, regardless of color purity. This means that the band characteristics of the luminance signal when photographing a black-and-white subject are narrower than in the conventional example, but the amplitude in the high range can be increased, resulting in a clearer image. The combined luminance signal characteristics at this time are shown by the solid line in FIG.

Now, if the overall sensitivity characteristics of each individual including the color filter element shown in Fig. 6 are made equal when capturing a black and white image of the subject, then in the case of black and white images, the output signal will not be sampled and the high frequency signal will be transmitted through only the low pass filter. By removing the component, the bandwidth of the luminance signal can be expanded to 1/2 of the horizontal drive frequency.
This relationship is shown in FIG.

In FIG. 9, A is the horizontal drive frequency, and B
is the sampling frequency. According to the sampling theorem, the luminance signal band is 1/1 of the sampling frequency.
2, that is, only up to point C, but if only a low-pass filter is used without sampling, the bandwidth will expand up to point B. In other words, it has the bandwidth shown by the dotted line. In reality, as explained earlier, a spatial frequency filter is inserted to remove spurious signals and optical blur is produced, so the bandwidth becomes a little narrower. An example of its characteristics is shown in FIG.

Note that in the case of colored objects, the bandwidth of objects with high color purity is narrow up to point C, although this varies depending on the color. However, general objects generally have low color purity, and therefore, in most cases, a wide bandwidth is used, which essentially results in an increase in resolution. However, in the case of a colored subject, so-called line shading occurs in which the amplitude differs for each horizontal line, as is clear from the waveform in FIG.

In FIG. 7, if the signal outputs from scanning n(H) and n+1(H) are denoted by S _o and S _o+1 , the general expression of the signal from each scanning is given by the following equation.

S _o = (R + Y) + (R - Y) _sio 〓 _t = (2R + G + B) + (R - Y) _sio 〓 _t (2) S _{o +1} = (B + Y) + (B - Y) _sio 〓 _t = ( R+G+2B)+(B-Y) _sio 〓 _t (3) However, Y=R+G+B, ω is the repetition angular frequency of R, B, and Y. As explained in Figure 7, a magenta-colored subject with a hint of blue is imaged. In this case, S _o and S _o+1 are given by the following equations. However, it is assumed that the spectral characteristics of this magenta color are R:G:B=1:0:2.

S _o = (R + Y) + (R - Y) _sio 〓 _t = (R + R + 2B) + (R - (R + 2B)) _sio 〓 _t = (2R + 2B) + (-2B) _sio 〓 _t (4) S _{o +1} = (B+Y)+(B-Y) _sio 〓 _t = (2B+R+2B) + (2B-(R+2B)) _sio 〓 _t = (R+4B)+(-R) _sio 〓 _t (5) So, for each horizontal line, S _o , S _o+1 , the R, G, and B components that make up the low-frequency components are different, so when capturing an image of a colored subject, the amplitude differs for each horizontal line, and on a TV screen, the brightness differs for each horizontal line. A so-called line shading phenomenon occurs.

Therefore, if even this line shading can be removed, the resolution can be greatly improved. The present invention aims to eliminate line shading and widen the band of luminance signals with an extremely simple configuration, and one embodiment thereof will be described below.

FIG. 11 shows a block diagram of a color solid-state imaging device according to an embodiment of the present invention. 1st
In FIG. 1, 80, 83 to 89 are exactly the same as those having the same reference numerals in FIG.

The output signal from the solid-state imaging plate 80 is supplied to a bandpass filter 83 and an adder 111. A bandpass filter 83 separates the signal components spatially modulated by the color filter. The separated modulated signal components are supplied to a synchronous detector 84. A reference signal used for synchronous detection is supplied to the synchronous detector 84 from a pulse generator 89 . This reference signal is
It is synchronous with the horizontal transfer clock pulse of the CCD, and its frequency is equal to the horizontal repetition frequency of the color filter.

The synchronous detector performs synchronous detection of the modulated signal using the reference signal. It is well known that by inverting the phase of the reference signal supplied to the synchronous detector, the polarity of the synchronously detected output signal can be inverted.

The synchronously detected modulated signal is passed through a low pass filter 8.
5, is subjected to circuit processing by a one horizontal period delay line 86 and a switch circuit 87, and is then supplied to a color encoder 88.

The synchronously detected modulated signal is also supplied to an adder 111. The adder 111 adds the detected modulation signal and the image sensor output signal.

At this time, the polarity and amplitude of the detected signal to be added to the image sensor output signal are set so that the amplitude of the low-frequency signal obtained from the low-pass filter 112 does not vary from one horizontal scanning line to another.

However, this polarity and amplitude are (2), (3),
As is clear from the formulas shown in (4) and (5), it is determined uniquely. The reason for this will be explained next.

The waveform shown in FIG. 7 shows an image sensor output signal when an object of magenta (blue+red) color with a bluish tinge is imaged. The color filters at this time are R, B, and Y filters, where R is red transparent;
B has a spectral characteristic of transmitting blue light, and Y has a spectral characteristic of transmitting all light. Since Y=(R+G+B), the signals corresponding to each color filter are R, B, and (R+B).
If the ratio of red and blue of the subject is 1:2, then,
The output signals of S _o and S _o+1 , _{that is} , the output _signals of A and B _in FIG. _t (6) S _o+1 = (2B+Y) + (2B-Y) _sio 〓 _t = (R+4B) + (-R) _sio 〓 _t (7)

Therefore, in order to equalize the signal amplitudes and spectral characteristics of the signals of S _o and S _o+1, add 2B to S _o and R to S _o+1 , then both the two signals become equal. (2R+4B), which results in equal signal amplitude and signal spectral characteristics.

In other words, it can be seen that the modulated signal separated from the image sensor output signal is synchronously detected on the low band signal separated from the image sensor output signal, the signal is inverted, and the inverted signal is added to the low band signal.

Therefore, the same effect can be obtained by subtracting the signal obtained by synchronously detecting the modulation signal from the low-frequency signal separated from the image sensor output signal.

Furthermore, the effect of eliminating the line shading phenomenon will be explained using FIG. 12.

FIG. 12 is an example in which the magnitudes of the R and B signals are reversed with respect to the Y signal compared to the case of FIG. 7.

In the combination of the image sensor and the color filter in FIG. 12, the spectral characteristics of the R, B, and Y color filters are R
is red transmission, B is blue transmission, Y is luminance signal characteristic, i.e.,
This is a combination of a color filter and an image sensor having a characteristic of Y=(0.3R+0.6G+0.1B).

FIGS. 12a and 12b schematically show horizontal scanning output signals when a magenta object with a large red component is imaged using the above combination. When the spectral characteristics of a magenta subject with a strong red color are approximately R:B=4:1, the magnitude of R of a is approximately 4, the magnitude of B is approximately 1, and the magnitude of Y is approximately 1.3.

The image sensor output signals Sa and Sb from adjacent horizontal scans when the above object is imaged are expressed by the following equations.

Sa = (4R + 1.3Y) + (4R - 1.3Y) _sio 〓 _t = (5.2R + 0.1B) + (2.8R - 0.1B) _sio 〓 _t (8) Sb = (B + 1.3Y) + (B - 1.3 Y) _sio 〓 _t = (1.1B + 1.2R) + (0.9B - 1.2R) _sio 〓 _t (9) Figures 12c and d are the modulation components of the signals shown in Figure 12a and b.

What should be noted here is that the polarities of the modulation components (AC components) in c and d are opposite to each other.

The fundamental wave components of the signal waveforms shown in c and d can be expressed by the following equation.

Sc=(4R−1.3Y) _sio 〓 _t ={(4R−(1.2R+0.1B)} _sio 〓 _t =(2.8R−0.1B) _sio 〓 _t (10) Sd=−(B−1.3Y) _sio 〓 _t = -(0.9B-1.2R) _sio 〓 _t (11) By synchronously detecting these signals with the reference signal sinωt, a color difference signal of (2.8R-0.1B) is obtained as a low-frequency signal from Sc. A color difference signal of (0.9B-1.2R) is obtained as a low frequency signal from Sd.

12e and f are waveforms obtained by synchronously detecting the waveforms shown in c and d (output signals of the synchronous detector 84). This is a low frequency component. Therefore, Figure 12e
is a low frequency signal appearing in the positive direction, and FIG. 12f is a low frequency signal appearing in the negative direction.

By inverting the polarity of the signals obtained by synchronizing the modulation components as described above and adding them to the image sensor output signal, and removing unnecessary high-frequency components using a low-pass filter, adjacent horizontal scanning The magnitude of the low-frequency signal, that is, the luminance signal, obtained by , and the ratio of each color signal component constituting the signal are equal as shown in the following equation.

Sg = (5.2R + 0.1B) - (2.8R - 0.1B) = (2.4R + 0.2B) (12) Sh = (1.1B + 1.2R) - (0.9B - 1.2R) = (2.4R + 0.2B) ( 13) The diagrams shown in FIGS. 12g and 12h show how the signals obtained by synchronously detecting the modulation components are each added to the image sensor output signal with the polarity inverted. In other words, the magnitude of the signal indicated by the solid line is equal between adjacent horizontal scans.

Therefore, it is possible to improve the horizontal resolution by about twice as much, and it is possible to achieve this objective with a relatively small number of elements. Although the description has been given on the case where a single color imaging plate is used, two plates may be used, or a combination of primary color and complementary color filters may be used.

As described above, according to the present invention, the number of horizontal elements is small in order to obtain the same resolution, and a good color image can be obtained. Furthermore, a high resolution color image can be obtained with the same number of elements. Furthermore, the filter configuration can be simple, and the circuit configuration can be simplified as only a few additional functions are required, making it possible to obtain a low-cost color solid-state imaging device.

[Brief explanation of the drawing]

Figure 1 is a plan view showing an example of a solid-state imaging plate that combines a photosensitive element and a charge transfer element, Figure 2 is a diagram showing an example of a three-color striped filter, and Figure 3 is a plan view showing an example of a three-color striped filter. A block diagram showing a conventional color solid-state imaging device that obtains color signals by pasting a color filter on a solid-state imaging board as shown in FIG. Figure 5 is a characteristic diagram of the spatial frequency filter, Figure 6 is a characteristic diagram of a mosaic color filter, and Figures 7A and B are output waveform diagrams of the imaging plate when the filter shown in Figure 6 is used. , FIG. 8 is a block diagram of a color solid-state imaging device using an element that combines the filter and solid-state imaging number shown in FIG. FIG. 10 is a characteristic diagram of an optical low-pass filter, FIG. 11 is a block diagram showing a color solid-state imaging device in an embodiment of the present invention, and FIG. 12 is a, b, c, d, e, f,
g and h are diagrams showing voltage waveforms at various parts in FIG. 11. 80... Solid-state image pickup plate, 83... Band pass filter, 84... Synchronous detector, 85... Low pass filter, 86... 1 horizontal period delay line, 87... Switch circuit, 88... Encoder, 111... Adder, 112...Low pass filter.

Claims (1)

[Claims] 1 A first color filter having a first spectral characteristic, a second color filter having a characteristic substantially equal to the luminance characteristic, and a third color filter having a third spectral characteristic, A first photosensitive element row in which combinations of first and second color filters and photosensitive elements are alternately and repeatedly provided, and a second row in which combinations of second and third color filters and photosensitive elements are alternately and repeatedly provided. A photosensitive element group having a photosensitive element row repeatedly arranged in a vertical scanning direction, extracts a modulation component by the color filter and a low frequency component from the output signal of the photosensitive element group, and from the low frequency component, A color solid-state imaging device characterized in that a signal obtained by synchronously detecting the modulation component at a repetition frequency of a color filter is subtracted.