JPS63276381A - Picture input device - Google Patents

Abstract

PURPOSE:To improve the S/N of a video signal and to obtain a picture with high resolution by clamping a feed through period of an output signal from a solid-state image pickup element and extracting only the effective signal portion. CONSTITUTION:A video signal from a solid-state image pickup element 11 is inputted to a clamp circuit 12 and part of the feed through period is clamped by a clamp pulse CP and subjected to DC recovery at a prescribed voltage. The DC recovery signal from the clamp circuit 12 is given to an extraction circuit 13 and only a requested color signal is extracted. The color signal extracted is outputted via amplifiers 14G, 14R, 14B, low pass filters 15G, 15R, 15B, a processing circuit 20 and an encoder 21.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an image input device using a semiconductor solid-state image sensor, and in particular to improving the signal-to-noise ratio of a video signal and obtaining a high-resolution image. This invention relates to an image input device that can perform

[Conventional technology]

Image input devices using semiconductor solid-state image sensors are becoming smaller and
It is rapidly becoming popular because it can achieve lower power consumption, longer life, and lower price. As is well known, the semiconductor solid-state image sensor of such an image input device is constructed by arranging a large number of pixels, 200,000 or more, on a 10 (mm) square semiconductor substrate, and the light incident on these pixels is Accordingly, a photocharge is generated and the photocharge is temporarily stored in each pixel.
This is then read out using a solid-state scanning method.The solid-state scanning method for reading out photoelectric charges from such a semiconductor solid-state image sensor uses circuitry for a large number of pixels arranged independently on a semiconductor substrate. The photoelectric charge accumulated in the pixels is read out one after another by applying a scanning pulse, and there are an X-Y addressing method and a charge transfer method. Targeted at image input devices.

FIG. 5 is a schematic diagram showing the output section of the semiconductor solid-state image sensor of the above-described image input device.

In FIG. 5, transfer electrodes 2 Hl, 2 Hz, 2 H3, 2H4 and an output gate 3 are formed on a semiconductor substrate 1 via an insulating film, and on the right side of the output gate 3 there is a A floating junction 4 is formed. On the right side of the floating junction 4 in the figure, a reset electrode 5 is placed on the group [1 through an insulating film].
A reset drain 6 is provided in the substrate 1 adjacent to the electrode 5. A field effect transistor (FET) is formed in which the floating junction 4 serves as a source, the reset electrode 5 serves as a gate, and the reset drain 6 serves as a drain. Therefore, the reset electrode 5 will be hereinafter referred to as a reset gate 5. In addition, Floating Junction 4 also has
FETl-transistor TPI+TFt, "rr3.
The input terminal of a sense amplifier circuit 7 having a two-stage source follower configuration consisting of T□ is connected. Terminal ■ of the amplifier circuit 7
The (+) pole of the battery 8 is connected to □, and the terminal 5 of the amplifier circuit 7 is connected to □. and the (-) pole of the battery 8 are connected to ground. In addition, transfer electrodes 2 Hl, 2 H-, 2H
A predetermined pulse or the like is applied to the z, 2 H-1 output gate 3, reset 5, and reset drain 6. The photoelectrically converted signal charges are transferred to the transfer electrode 2H.
, 2H□2H, , 2H, , are transferred to the right in the figure through the substrate 1, detected by the floating junction 4, and input to the sense amplifier circuit 7. 'The sense amplifier circuit 7 amplifies the signal and outputs it to the output terminal V. It is configured so that it can be output from the UA.

The operation of the output section configured in this way is shown in FIGS. 5 and 6.
This will be explained with reference to the flowchart shown in the figure.

The transfer electrodes 2H on the substrate 1 are given predetermined pulses φH1, φH2, φH3 at 2Hz, 2Hs to move the signal charges to the right in the figure.Next, the transfer electrodes 2H on the base Fi1 are , the pulse φH4 shown in the sixth (A) is applied to collect the signal charge under the transfer electrode 2H (period jl""F), and the % signal charge is detected at the floating junction 4 (period t3 to t8). , see FIG. 6(E)), and by applying the reset pulse R3 to the reset gate 5 during the period t4 to t as shown in FIG. 6(B), the reset pulse R3 as shown in FIG. ) will be obtained. On the other hand, the signal detected at the floating junction 4 is amplified by the sense amplifier circuit 7 and sent to the output terminal V. It's taken out from there. Here, a reset pulse is applied to the reset electrode 5 on the substrate 1 at the timing shown in FIG. 6 to reset the floating junction 4 to the reference potential. Incidentally, the output signal includes a period (t) in which a reset pulse is applied to the reset gate 5 and the reset gate 5 is on.
a~ts), floating junk 4. This includes reset noise consisting of thermal noise from the FET consisting of the recent gate 5° and reset drain 6, and noise due to a mixture of amplitude fluctuations of the reset pulse applied to the reset gate 5. Therefore, if this output signal is averaged by a low-pass filter, the averaged output signal will contain low frequency components due to reset noise, and the reproduced image will be significantly degraded.

For this reason, conventionally, as a means to remove this reset noise and improve the signal-to-noise ratio (S/N), the reference potential during the stable period of the output signal is clamped to a certain constant potential, and then the signal during the effective signal period is A correlated double sampling method for sampling potentials has been proposed.

[Problem that the invention seeks to solve]

By the way, separation of color signals of a color semiconductor solid-state imaging device has conventionally been carried out using the above-mentioned correlated double sampling method. For example, in the case of a color semiconductor solid-state imager that employs vertical stripe filters of R (red), G (green), and B (blue), the field-through period (sixth
After clamping the signals t3 to t4 (t, to tb) in FIGS. 3 and 4, the color signals were separated by sampling effective signal portions corresponding to each color signal. Therefore, the color signal R
For 1G and B, in this case, the signal of one pixel is averaged over three pixels. Next, as methods for obtaining a luminance signal, there are two methods: one that samples only C(8), and one that samples the effective period of all signals.

In the method of sampling only the G signal, since the sampling point is 173 of all pixels, the resolution is extremely reduced, and as the number of pixels decreases, there is a problem with the magnitude of the luminance signal.

The method of sampling the valid period of the golden signal is simply sampling, so G as its luminance component is
, B, and R are controlled by the spectral transmittance of the color filter.
In addition to the fact that HL is necessary and it is difficult to design a filter, in the case of a chromatic object, the signal amount of each color is different, so there is a problem that the difference in signal amount appears in the luminance signal and becomes a false signal.

Furthermore, if the signal amount of each sampled color signal is normalized and then added to create a luminance signal, there is a problem that sufficient resolution cannot be obtained because the pixel information of each color is averaged.

In addition, due to sampling, there is also noise generated when frequency components higher than Nyquist fall back into the signal band.

It is an object of the present invention to solve the above problems and to
An object of the present invention is to provide an image input device that can improve N and obtain high-resolution images.

[Means for solving problems]

An image input device according to the present invention that achieves the above object includes a solid-state image sensor that converts a signal incident on a semiconductor substrate into a signal charge and outputs an electric signal based on the signal charge, and an output from the solid-state image sensor. The present invention is characterized in that it includes a DC regeneration circuit that regenerates DC by clamping the feed-through period of the signal, and a sampling circuit that extracts only the effective signal portion of the signal from the DC regeneration circuit.

Another invention provides a solid-state imaging device that converts an optical signal incident on a semiconductor substrate into a signal charge and outputs an electrical signal based on the signal charge, and a feed-through period of an output signal from the solid-state imaging device. A DC regeneration circuit that clamps and regenerates DC, a sampling circuit that extracts only the effective signal portion of the signal from the DC reproduction circuit, and amplification control of each color signal obtained from the sampling circuit individually to obtain a predetermined amplitude. The present invention is characterized in that it includes a variable gain amplifier circuit that outputs an output signal of the variable gain amplifier circuit, and a mixing circuit that mixes each color signal from the variable gain amplifier circuit and outputs the resultant signal as a luminance signal.

[For production]

In the present invention, a DC regeneration circuit performs DC regeneration during the feed-through period after the output signal from the solid-state image sensor is reset, thereby removing low frequency noise caused by the reset. Further, only the valid signal period from the DC regeneration circuit is extracted by the extraction circuit.

This reduces aliasing noise associated with sampling.

Furthermore, the signals extracted by the extraction circuit are individually amplified by variable gain amplifier circuits to produce output signals of a predetermined amplitude, and these are mixed by a mixing circuit to obtain a luminance signal.
You can get a high resolution screen.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

1 to 4 are diagrams shown to explain the present invention in detail, FIG. 1 is a block diagram showing the embodiment, FIG. 2 is a diagram showing a solid-state image sensor of the embodiment, and FIG. FIG. 3 is a flow chart shown to explain the operation of the same embodiment, and FIG. 4 is a circuit diagram showing a specific example of the sampling circuit of the same embodiment.

The embodiment shown in FIG. 1 includes a solid-state ↑ layer image element 11 that can convert an optical signal on a semiconductor substrate into a signal charge and output it as a video signal, and a feed-through period after reset of the output signal from this solid-state image sensor. A clamp circuit 12 is used as a DC regeneration circuit that clamps a portion of the DC signal with a clamp pulse CP and regenerates the DC voltage to a predetermined voltage, and extracts only the required color signal from the DC reproduction signal from the clamp circuit 12 as the DC regeneration circuit. Therefore, the extraction pulse G P l+
GPz, a gate circuit 13C gate-controlled by GP3; a sampling circuit 13 consisting of 13R, 13B; and each color signal SG, S from the sampling circuit 13;
A. Amplification circuits 14G, 14R, 14B that amplify Ss
and low-pass filters 15G, 15R, 1 that band-limit the output signals from each amplifier 14G, 14R, 14B.
5B, and each color signal S c, S *, S m obtained by the sampling circuit 13 is input to the G signal section 11. Variable gain amplifier circuits 16G, 16R, 16B that control and amplify the amplitude of the R signal section, B signal Sm using the reference value as a reference value, and a mixing circuit that mixes each signal obtained from each amplifier circuit 160, 16R, 16B. 17, an amplifier circuit 18 for amplifying the signal from the mixing circuit 17, a low-pass filter 19 for band-limiting the signal from the amplifier circuit 18, and this low-pass filter 1.
9 and the low pass filters 15G, 15R, 15B
A processing circuit 20 that processes signals from the processing circuit 20, an encoder 21 that converts the signals from the processing circuit 20 into NTSC signals, and a clamp pulse CP and a sampling pulse CP based on the horizontal drive pulse φH that drives the solid-state image sensor 11.
, , G P , CP , and a control circuit 22 that outputs .

Furthermore, the formula expressing the stimulus values of the three primary colors of the luminance signal is Ey for the luminance signal, E R for the R signal value, and EG for the G signal value.

When the B signal value is E, E y-0, 30E, + 0.59E, + 0.
11E, -------(1) is given.

Figure 2 (1) shows color filters IIIG and IIIR.
, 111B are arranged vertically, and each filter 1
The line rows of 11G, IIIR, and IIIB are repeated in order to form a stripe shape, and each filter IIIG
, IIIR, 111B, a photoelectric conversion section 111 is configured.

A first signal charge transfer section 112 for transferring photoelectrically converted signal charges is vertically provided on each side of these photoelectric conversion sections 111, and a second signal charge transfer section is provided below these signal charge transfer sections 112. 113 is provided. An output amplification circuit 114 is provided that detects and outputs charges transferred from the second signal charge transfer section 113. From this output amplification circuit 114, G and R are output.

A signal B is output.

In FIG. 2(n), the second signal charge transfer section 113 is composed of two, and one signal charge transfer section 113 receives only the G signal, and the other signal charge transfer section 113 receives only the G signal. R signal/B
It is designed so that signals can be transferred, and each transfer unit 11
3G, 113RB has an output amplifier circuit 114G, u4
RB is provided. G from the output amplifier circuit 114G
The signal is output from the output circuit 114RB, and the R signal/B signal is output from the output circuit 114RB.
Each is output.

In addition, in FIG. 2 (I), it is a G=R"-B filter, and in FIG. 2 (n), it is a G-R-G-B filter. Therefore, the output signal is as shown in FIG. G-R-B
In FIG. 2 (I[), the signal is transferred to the second signal charge transfer unit 113.
The G signal is output from the G signal, and the R signal/B signal power (respectively) is output from the second signal charge transfer unit 113RB.

The operation of the embodiment configured as described above is illustrated in FIGS. 1 to 3.
Tj! This will be explained with reference to J. In the following explanation, the second
A description will be given assuming that the solid-state image sensor shown in FIG. (1) is used.

As shown in FIG. 3(a), the output signals from the solid-state image sensor 11 have different amplitudes for the G signal, R signal, and B signal. The feedthrough period of this output signal (Fig. 3)
-j a ), the control circuit 22 supplies the clamp pulse CP (FIG. 3(b) t2 to t3) with the same frequency as the horizontal drive pulse φH to the clamp circuit 12, and the feedthrough period (FIG. 3 tl- A part of t a' ) is clamped and DC regenerated to a predetermined potential. Next, a sampling pulse (315(c)t) is applied to the first gate circuit 13G of the sampling circuit 13.

~ts) and extracting the effective signal portion (t, ~tS in FIG. 3), a G signal Sc as shown in FIG. 3(f) is obtained.

In addition, the control circuit 22 supplies the clamp pulse CP (from t+z to t13 in FIG. 3) with the same frequency as the horizontal drive pulse φH to the clamp circuit 12, thereby starting the feedthrough period (tll to t, 4 in FIG. 3). A part of the DC voltage is clamped to regenerate a DC of a predetermined voltage, and then a sampling pulse (FIG. 3(d) t+a
~t+s) to obtain the effective signal portion (Fig. 3, 11 to t24)
), an R signal portion, l, as shown in FIG. 3(g) will be obtained.

Similarly, a clamp pulse CP (tzz-tzs in FIG. 3(b)) is applied from the control circuit 22 to the clamp circuit 12 to clamp a part of the feedthrough period (11+ to t24 in FIG. 3) to obtain a predetermined voltage. Then, a sampling pulse shown in FIG. 3(e) is given to the third gate circuit 13B of the sampling circuit 13 to reproduce the effective signal portion (tz
4-t2. ), a B signal part like the third (h) can be obtained.

If you repeat the above, G-4R→B→G→R→B→-...
The color signal S ,, S *, S that repeats...
m will be obtained.

The three color signals Ss, S+t, and Ss separated by the extraction circuit 13 in this way are sent to the amplifier circuit 14G.
, 14R.

After each amplification by 14B, low pass filter 15
The signal is band-limited by G, 15R, and 15B and provided to the processing circuit 20. On the other hand, the three color signals S, S, Sm separated by the extraction circuit 13 are always in the G signal part 0. Amplifying circuits 16G and 16R amplify the R signal part 6 using the reference value as a reference value, and the B signal part 6 with the relationship expressed by equation (1).

After being amplified by 16B, it is applied to the mixing circuit 18.

In the mixing circuit 18, each of the amplifier circuits 16C;, 16R
, 16B with a specific relationship are mixed to produce a mixed signal (luminance signal) as shown in FIG. 3(i).
is obtained. Regarding such a luminance signal, the amplifying circuits 16R and 16B perform an amplifying operation as described above, and the G-R-B signal thus amplified is sent to the mixing circuit 17.
Since the pixel portions where the color signal is missing are accurately interpolated with other color signals, the signals of all pixels can be used as the luminance signal.

Therefore, the band of the luminance signal is, in principle, up to the Nyquist limit determined by the number of pixels of the solid-state image sensor, and the maximum resolution can be obtained.

This luminance signal is then amplified by an amplifier circuit 18, band-limited by a low-pass filter 19, and then sent to a processing circuit 2.
given to 0.

After being processed, the color signal and luminance signal supplied to the processing circuit 20 are converted into N and TSC signals by the encoder 21.

As described above, according to this embodiment, only the signal portion is extracted from the separated color signal, resulting in a signal that does not contain reset noise or aliasing noise.

Further, according to this embodiment, when focusing on one separated color signal, the other two separated color signals contain luminance information at the signal position where this signal is missing, so these signals are By using the signal as a reference and controlling the signal amount according to the imaging conditions and then adding it, a luminance signal that is the average of all pixels is obtained, which is equivalent in principle to a black and white television camera. pixel information will be obtained.

FIG. 4 is a circuit diagram showing a specific circuit example of the main part of this embodiment.

In FIG. 4, a signal from the solid-state image sensor 11 is supplied to the clamp circuit 12 via the amplifier circuit 23. In addition to the solid-state image sensor 11, the clamp circuit 12 receives a clamp pulse CP and a signal for specifying a clamp level.

The output from the clamp circuit 12 is sent to the gate circuits 13G, 13R, and 13B of the sampling circuit 13 via the amplifier circuit 24.
is being supplied to. At each gate 13G, 13R, 13B of the extraction circuit 13, extraction pulses G P , CP , , G P are connected to a capacitor 25G.

It is designed to be supplied via 25R and 25B.

The gate circuit 13G is a transistor T rl O to T P
It is composed of a double balanced differential current switch circuit composed of I3, resistors RIO to RH, and a diode D1. Each gate circuit 13R, 13B is also constructed of the same double balanced differential type current switch circuit.

Further, the signal from the clamp circuit 12 is transmitted via the amplifier circuit 24 to the connection point between the emitter of the transistor TPIS and the collector of the transistor Tr17 of the double-balanced differential current switch circuit constituting each of the gate circuits 13G, 13R, and 13B. are supplied to each. Extraction pulse G P +, G P2+ CP,
is also supplied to the base of the transistor T11°+Tr+3 of each double-balanced differential current switch circuit. Each output of the extraction circuit 13 is connected to a transistor T r+ of each double balanced differential current switch circuit.
It is designed to be taken out from the collector of I+Tr+3.

In this way, each gate circuit 13G of the extraction circuit 13.

The reason why double balanced differential current switch circuits are used for 13R and 13B is as follows.

That is, the extraction circuit 13 needs to have a function of reliably extracting only the effective signal portion in order to sufficiently improve the band by signal addition and to eliminate aliasing noise due to sample and hold. In addition, the output signal of the solid-state image sensor 11 has a clock frequency of approximately 7 (M
Since the frequency is higher than H2), the normal MO
In S gates and switch circuits, even the presence of a small amount of capacitance, such as the capacitance of the switch circuit or the capacitance due to wiring, causes a hold effect, causing the inconvenience that the aperture of the signal to be added becomes wider. Therefore, each gate circuit 13G is provided with a double balanced differential current switch circuit as shown in FIG. 4, which does not cause such inconvenience and can perform a reliable extraction operation.

This was adopted for 13R and 13B.

Incidentally, the present invention can be modified and applied as follows.

(a) In the above embodiment, the extraction pulse G P l+ G
P.

Although GP has been described using a rectangular wave, the present invention is not limited to this, and a sine wave may also be used. Considering the noise of the sampling pulse itself, a sine wave is more effective.

(b) In the above embodiments, the solid-state image sensing device has been described as a COD, but is not limited to this, and may be a CPD, for example.

(c) In this embodiment, the force explained in the example of a television camera (not limited to this, the force can be applied to various image input devices).

(d) In this embodiment, the case where there is one second signal charge transfer section (horizontal register) 113 has been described, but the case may be such that there are two second signal charge transfer sections (horizontal register) 113. In this case, pixel information can be embedded in the reset and feed-through periods, which is even more effective.

(e) In this embodiment, the primary color filter is G-R-B, G-=R-G-B, but of course G-Y.

(yellow) -C, (cyan), C=Y, -G-2Cy
It may also be a complementary color filter such as.

Moreover, it is not limited to a vertical stripe filter.

〔Effect of the invention〕

As described above, according to the present invention, only the effective signal portion is extracted from the output signal from the solid-state image pickup device, so that the video signal signal is eliminated from the conventional cent noise and aliasing noise due to sampling. This has the effect of significantly improving the noise ratio. Further, according to the present invention, since the extracted color signals are amplified with a specific relationship and these are added, it is also possible to obtain a high-resolution image.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing a solid-state image sensor used in the embodiment, FIG. 3 is a time chart shown to explain the operation of the embodiment, and FIG. FIG. 5 is a circuit diagram showing a specific circuit example of the extraction circuit used in the same embodiment, FIG. 5 is a schematic diagram showing a conventional example, and FIG. 6 is a time chart shown to explain the operation of the conventional example. 11-・Semiconductor solid-state image sensor, 12-clamp circuit (DC regeneration circuit), 13-extraction circuit, 13G, 13
R, 13B-111 gate circuit, 14G, 14R
, 14B, 18-・Amplification circuit, 15C;, 15R
, 15B, 19--low pass filter, 16G,
16R, 16B--variable gain amplifier circuit, 17--mixing circuit, 20--processing circuit, 21--encoder, 22-
・-Control circuit. Agent Patent Attorney Tomo Murakami - Figure 2 (I) (1[) Figure 4

Claims (3)

[Claims] (1) A solid-state image sensor that converts an optical signal incident on a semiconductor substrate into an electric charge and outputs an electrical signal based on this signal charge, and performs DC regeneration by clamping the feed-through period of the output signal from the solid-state image sensor. An image input device comprising: a DC regeneration circuit; and a sampling circuit that extracts only a valid signal portion of a signal from the DC reproduction circuit. (2) The image input device according to claim 1, wherein the sampling circuit is constructed of a double-balanced differential current switch circuit. (3) A solid-state image sensor that converts an optical signal incident on a semiconductor substrate into a signal charge and outputs an electrical signal based on the signal charge, and a solid-state image sensor that clamps the feed-through period of the output signal from the solid-state image sensor. A DC regeneration circuit that regenerates DC;
a sampling circuit that extracts only an effective signal portion of the signal from the DC reproduction circuit; a variable gain amplifier circuit that individually amplifies and controls each color signal obtained from the sampling circuit to output an output signal of a predetermined amplitude; An image input device comprising: a mixing circuit that mixes each color signal from the variable gain amplifier circuit and outputs the mixture as a luminance signal.