JPS6238911B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION A solid-state imaging device using a charge transfer device such as a CCD or a BBD as an imaging device is configured as shown in FIG.

1R, 1G, and 1B are image sensors that capture red, green, and blue color-separated images, and their respective outputs are waveform-shaped by sampling and holding circuits 2R to 2B, and then subjected to γ correction by γ correction circuits 3R to 3B. and γ
The respective corrected outputs SS _R to SS _B are band-limited (about 0.5 MHz) by low-pass filters 4R to 4B, and then matrixed by a matrix circuit 5 to produce red and blue color difference signals R-Y, B-. Y and a low-band luminance signal Y _L are formed.

On the other hand, the γ correction outputs SS _R to SS _B are synthesized by a synthesizer 7 after level adjustment, and a high frequency luminance signal Y _H is extracted by a band pass filter 8, which is combined with the above-mentioned color difference signals RY, B. -Y and low-range luminance signal Y
_L is supplied to an encoder 9 to form a desired color video signal.

The above-mentioned γ correction circuits 3R-3B are normally provided before the low-pass filters 4R-4B. This is because it is necessary to use the γ correction outputs SS _R to SS _B to form the high-band luminance signal Y _H , and the passband of the primary color signals S _R to S _B must be set below the high-band luminance signal band. This is because it cannot be restricted.

In this embodiment, the spatial relative phase between the color-separated images projected on the image sensors 1R and 1B and the color-separated images projected on the other image sensor 1G is π (the pixel arrangement pitch in the horizontal direction is If τ _H , the relationship between the image sensor and the projected image is selected so that the spatial deviation is τ _H /2), and the level adjuster 11R
~11B, red and blue γ correction output SS _R and
The respective levels are adjusted so that the sum of SS _B is equal to the green γ correction output SS _G.

In this way, when the phase of the sampling signal related to the green primary color signal is used as a reference, the phase of the sampling signal related to each of the red and blue primary color signals will be exactly the opposite phase, so when capturing black and white images, the γ correction output
By combining SS _R to SS _B , the sampling signals are canceled out, which makes it possible to remove aliasing distortion mixed into the baseband component of the luminance signal _. You can also choose lower. Furthermore, if the sampling frequency is the same, that is, the number of picture elements in the horizontal direction is the same, aliasing distortion will not occur and the signal band will be wider.

Now, in the solid-state imaging device configured in this way, the sampling angular frequency is ω _s , and the color signal,
In other words, when the frequency band of the color difference signal is ω _c , the angular frequency ω _i of the primary color signal is within the band of ω _s /2−ω _c ≦ω _i ≦ω _s /2 + ω _c ……(1) In some cases, based on the detection action of the γ correction circuit and the phase difference of this detection output, which is the γ correction output, this detected component is mixed into the baseband of the primary color signal, resulting in cross color.

The generation of this cross color will be explained with reference to FIG. 2, taking the red primary color signal S _R as an example.

Now, the red signal S _R is expressed as in equation (2).

S _R R ₀ + R ₁ cos (ω _r t + φ _r ) ………(2) However, R ₀ : DC component R ₁ : High frequency component ω _r : Angular frequency of red signal φ _r : Phase of R ₁ Image sensor 1R Since the imaging output S′ _R sampled by the red signal S _R sampling signal S _sR (phase is θ _r ) is obtained, this imaging output
When dealing with the signal components below the lower sideband of S′ _R , this imaging output S′ _R becomes as follows (second
Figure B).

S' _R = R ₀ + R ₁ cos (ω _r t + φ _r ) + R ₁ cos {(ω _s − ω _r ) t + (θ _r − φ _r )} = R ₀ + R ₁ /2 cos (ω _s /2t + θ _r /2 )×cos {(ω _s /2−ω _r )t+(θ _r /2−φ _r )} ………(3) As is clear from equation (3), the imaging output S′ _R is (ω _s /
The carrier wave cos(ω _s /2t+ _{θ r /}
2) is in a balanced modulated form, so if this imaging output S′ _R is supplied to the γ correction circuit 3R, the frequency component of (ω _s /2−ω _r ) is detected by the detection action of this circuit. is demodulated (Fig. 2C).

This demodulated output (γ correction output) _SSR is band-limited by the low-pass filter 4R (Fig. 2C),
In the end, the output R _L is R _L = R ₀ + R′ ₁ cos {(ω _s /2−ω _r )t+(θ _r /2−φ _r )}=R ₀ +C _R ………(4) However, C _R ]R′ ₁ cos {(ω _s /2−ω _r )t+(θ _r /2−φ _r
)} The phases θ _r and θ _b of the sampling signals of the red signal S _R and the blue signal S _B are in phase, and the phase θ _g of the sampling signal of the green signal S _G is shifted by π from them, so the blue signal S The outputs _{B L and} G _L for the green signal S _G are as follows.

B ₀ +B′ ₁ cos {(ω _s /2−ω _b )t+(θ _r /2−φ _b )}=B ₀ +C _B ………(5) G _L =G ₀ +G′ ₁ cos {(ω _s /2−ω _g )t+(θ _r /2+π/2−φ _g )}=G ₀ +C _G ………(6) However, B ₀ , G ₀ : DC component B′ ₁ , G′ ₁ : High Area components ω _b , ω _g : Angular frequencies φ _b , φ _g of primary color signals _SB , _SG : Phases CB of primary color signals _SB , _SG 〕 _B ′ ₁ cos {(ω _s /2−ω _b ) t+(θ _r /2−φ _b
)} C _G = G′ ₁ cos {(ω _s /2−ω _g )t+(θ _r /2+π
/2−φ _g )} Therefore, considering a black and white subject, the following condition holds, a phase difference of π/2 occurs between the detection outputs C _R , C _B and C _G , and the color difference signal, for example, the red color difference signal R-Y, is R-Y=0.70R _L −0.59G _L − 0.11B _L = 0.59 (C _R − _{C G} ) = 0.59×√2cos {(ω _s /2−ω _r )t + (θ _r /2−φ _r +π/4)} ……(8) ≠0 This Thus, even if a black and white subject is imaged, the color difference signals RY and BY do not become zero, and cross color occurs. If there is no phase difference in the sampling signals, equation (8) becomes zero, but as described above, in order to eliminate the aliasing distortion that occurs in the luminance signal, it is not possible to make all the sampling phases equal.

In this way, cross color is based on the detection effect of the γ correction circuit and the sampling phase difference,
If the primary color signal exists outside the band ω _s /2±ω _c shown by equation (1), the frequency component after the detection output will exist outside the passband of the low-pass filter 4R, so In such cases, cross color will not occur.

Therefore, the present invention limits the frequency band of the primary color signal input to the γ correction circuit, and as shown in FIG. Among the bands, the signal band component of equation (1) is limited.

In this way, the high frequency component shown in Figure 2A
Even if the imaging output S' _R including R ₁ is input, it is removed by the band suppression circuit 20R, so the γ correction output becomes zero and cross color does not occur.

Note that the high-frequency luminance signal Y _H is transmitted through the band suppression circuit 20R.
Since it is formed based on the imaging outputs _S'R to _S'B of the previous stage of ~20B, there is no influence from band limitation. 21 is a low-pass filter, and 3Y is a γ correction circuit.

Here, for example, the sampling frequency f _s is
4.0MHz, and if the frequency band of the color difference signal is 0.5MHz as usual, the band suppression circuit 20R~20
In B, only the band from 1.5 MHz to 2.5 MHz is limited, so even if such band limitation is performed, there is no effect on color difference signal processing. Therefore, the band suppression circuits 20R to 20B are 0.5 to 1.5MHz.
A low-pass filter with a cutoff frequency of .

As explained above, according to the present invention, since γ correction is performed after band limiting,
The sampling phases are different, and the γ correction circuits 3R to 3
Even with the detection effect of B, cross color does not occur. Therefore, there is no deterioration in image quality even when capturing black and white images.

In addition, in the above-mentioned embodiment, 3 chips (3 boards) are used.
The present invention was applied to a solid-state imaging device in which the phase difference of the sampling signal was selected to be π in the format, but R, G,
The present invention can also be applied to all types of solid-state imaging devices in which the sampling phases of the three B primary colors are different from each other. Of course, the number of image pickup elements may be two chips or one chip.

[Brief explanation of the drawing]

Fig. 1 is a system diagram of the device used to explain the present invention, Fig. 2 is an explanatory diagram used to explain its operation, and Fig. 3 is an explanatory diagram used to explain its operation.
The figure is a system diagram showing an example of the present invention. 1R to 1B are image pickup elements, 3R to 3Y are γ correction circuits, 5 is a matrix circuit, 9 is an encoder, 2
0R to 20B are band suppression circuits.

Claims (1)

[Claims] 1. In a solid-state imaging device having a solid-state imaging device in which the sample phase of another primary color signal is different from that of a given primary color signal, each primary color signal obtained from the solid-state imaging device is passed through a band suppression circuit, and then γ While performing the correction, forming a color difference signal from the γ-corrected primary color signal, and forming at least a part of a luminance signal from the primary color signal before being supplied to the band suppression circuit, and by the band suppression circuit,
1. A cross color removal circuit for a solid-state imaging device, characterized in that a cross color removal circuit for a solid-state imaging device is configured to suppress at least a band having a bandwidth of the color difference signal above and below the center around 1/2 of the sampling frequency of each primary color signal.