JPH05914B2 - - Google Patents

Description

Detailed Description of the Invention (Technical Field) The present invention relates to an imaging apparatus using an imaging device, and in particular to a signal processing circuit for reducing aliasing distortion and a nonlinear processing circuit for improving the dynamic range of the signal processing circuit. The present invention relates to an imaging device equipped with a KNEE circuit having input/output characteristics.

(Prior Art) Conventionally, this type of color imaging apparatus uses an imaging device (such as a CCD) that includes a plurality of color separation filters having different spectral characteristics, for example, in a stripe pattern.

For example, as shown in Figure 1, a red light (R) transmission filter,
Let us consider a case where a green light (G) transmission filter and a blue light (B) transmission filter (hereinafter simply referred to as an R filter, a G filter, and a B filter) are sequentially arranged on the surface of an imaging device. The light that enters the imaging device through such a filter and optical system is spatially sampled by the color stripe filter and imaging device described above, but in this case, the number of pixels of the imaging device or the pitch of the color stripe filter is sampled spatially. Spatial frequency components of the incident light that correspond to 1/2 or more of the determined spatial sampling frequency cause aliasing distortion. This will be explained using FIG.
In each of the diagrams A, B, and C, the horizontal axis represents frequency and the vertical axis represents signal level.

The incident light sampled on the imaging device is read out from the imaging device as an imaging signal by photoelectric conversion, etc., but if we focus on only R (or only G, or only B) of this imaging signal, its repetition frequency is 1/3 of the read frequency. If this repetition frequency is c, the baseband component and band wave component of the incident light due to sampling will be as shown in figure A, and the shaded part in the figure is called the aliasing distortion component. When this signal is passed through a low-pass filter with characteristics as shown in the diagram B, this aliasing component will remain mixed with the baseband component,
This component causes a significant deterioration in image quality on a display. A method for reducing such aliasing distortion is described in Japanese Patent Application Laid-open No. 120281/1981. In other words, when an achromatic object image is photographed as shown in C in the same figure, if the color separation filter is designed so that the level of the point sequential signal output from the imaging device is 1:1:1, the sideband components will be reduced. By canceling each other out, aliasing distortion can be reduced.

According to this method, aliasing distortion can be reduced at least for achromatic screens.

Of course, such an effect cannot be obtained on a screen with high color saturation, but human visibility is low in the high frequency range when it comes to colors, so it can be ignored.

However, in the case described above, if the color temperature of the photographing light source differs depending on the purpose or location of photographing, the point-sequential signal level becomes unbalanced, eventually resulting in aliasing distortion. Figure 3 shows, for example, the color temperature.
Showing the spectral energy at 3200〓 and 6000〓,
FIG. 4 is a diagram illustrating a drawback when a color separation filter is designed so that the level of a point-sequential output signal is constant at 3200, for example. If you set a filter so that R, G, and B are 1:1:1 at a color temperature of 3200〓, as shown in Figure 4, at 6000〓, the long wavelength side, that is, the R side will be weak, and the B side will be strong. So,
The sideband vector is biased toward the cyan (Cy) side as shown in the fourth diagram, and aliasing distortion occurs.

Furthermore, imaging devices are generally highly sensitive to infrared light, which may deviate from the visual sensitivity. To prevent this, an infrared cut filter is installed in the optical path of light entering the imaging device. Unevenness in the thickness of the outer cut filter that occurs during the manufacturing process may cause variations in the spectral sensitivity characteristics of the filter, resulting in fluctuations in the R signal level.

The following method is known as a method for eliminating such defects.

That is, this method uses a mechanical color temperature correction filter. In this method, color temperature correction filters are generally prepared, for example, one for daylight, one for fluorescent light, and one for tungsten, and the above correction filters are switched depending on the shooting location. The disadvantages of this method are that several types of correction filters are required, and that the generation of aliasing distortion cannot be completely prevented because the level adjustment of the point-sequential signal is rough.

Furthermore, conventional imaging devices have the drawback of large noise and poor image S/N. For example, Tokkosho
In the television camera using a three-electrode image pickup tube described in Publication No. 55-51395, R, G,
A high-frequency luminance signal is obtained by adding the B signal. That is, the output signals R, G, and B of the three electrodes become as shown in FIG. 5 by scanning with an electron beam, and a luminance signal Y is obtained by combining the respective color signals. In this case, the noise generated in the image pickup tube is generated uniformly in both the signal component region and the invalid component region, so the noise in the combined luminance signal becomes approximately √3 times, and the S/N decreases by that amount. There were some drawbacks.

In addition, in such imaging devices, there is a tendency to lower the power supply voltage of the signal processing circuit in order to downsize and reduce power consumption, and currently, a 5V power supply is common. In order to generate a high-quality video signal using such a low power supply, it is necessary to appropriately process the signal level and equivalently expand the dynamic range of the signal processing circuit. Generally, as a method of expanding the dynamic range in video signal processing, a circuit called a KNEE circuit that nonlinearly suppresses high-intensity signals is known. However, this
If the KNEE circuit was applied directly to signals read out from an imaging device, aliasing distortion would increase in luminance signals, and white balance errors would occur in color signals, resulting in a significant drop in image quality.

FIGS. 6A and 6B are explanatory diagrams of the drawbacks of the above-mentioned prior art. C is an explanatory diagram for the present invention, the contents of which will be described later.

First, illustration A will be explained. As described above, the output level of each color signal of an imaging device often differs depending on the color temperature of the light source at the time of photographing. Here, as an example of explanation, it is assumed that an achromatic object is to be photographed, and that the R and G signal levels at that time are at the same level, and the B signal level is at a slightly lower level. That is, as the amount of light increases, each signal level exhibits the characteristics shown by the solid lines of R, G, and B shown in the figure. At that time, the saturation signal level of the signal is expressed by the symbol Vsat,
Normally, the saturation levels of R, G, and B are the same, so
R and G are saturated at the light amount shown at point a, and B is saturated at the light amount shown at point b. In this state, in order to expand the dynamic range of the circuit, if we apply KNEE characteristics to the R, G, and B signals at a predetermined signal level V _KNEE , the R, G, and B signals will be R', G', This becomes the signal B′. Next, the white balance of these R′, G′, and B′ signals,
That is, when level matching is performed, as shown in Figure B, the signal levels of R'G' and αB' (α is a coefficient for level matching) do not match above the V _KNEE level.
(Here, V _WC is the white clip level, and signals higher than this are blocked by the circuit.) As a result, at the signal level in the shaded area shown in the diagram, the luminance signal level does not match, causing aliasing distortion. Become. Another drawback is that white balance cannot be maintained for color signals.

(Objective) It is an object of the present invention to provide an imaging device equipped with a signal processing circuit that can eliminate the above-mentioned drawbacks and easily adjust the signal level in a point-sequential manner.

(Embodiment) An embodiment of the present invention will be described by way of example of an imaging apparatus using a CCD type imaging element, as shown in FIG.

The image sensor shown in Figure 7 is a frame transfer type.
It is a CCD. First, information charges photoelectrically converted in the imaging section 1 corresponding to each color filter of the strump filter are transferred at high speed to the memory section 2 during the vertical retrace period of TV synchronization by driving pulses φPI and φPS.
Furthermore, for each vertical transfer of one horizontal line of information charges accumulated in the memory section 2, the information corresponding to each stripe filter is transferred to the horizontal shift register SR1,
It is distributed and transferred to SR2 and SR3.

That is, as shown in FIG. 8, in this embodiment, the memory section 2
Information for one horizontal line of is distributed to shift registers SR1 to SR3 for each color information, respectively, and from horizontal shift registers SR1, SR2, and SR3, R,
G and B signals are output. Therefore, register SR1,
SR2 and SR3 constitute separation means for separating color signals.

FIG. 9 is a block diagram of a signal processing circuit for signals read out from the CCD. clock IC30,
Imaging device 1 driven by driver 20
For example, the first filter shown in the figure is attached to the surface of the 0 (CCD in this embodiment), and R, G, and B signals corresponding to the color separation filters are separately obtained as output signals. This signal is subjected to DC reproduction in the clamp circuit 40, and guided to the gain control circuit 50 as the next stage level control means.
By a control signal of a level adjustment circuit (not shown), R,
The G and B signals are made to the same level. It is even better to use a feed pack clamp circuit that feeds back the DC potential of the input signal of the switch circuit 60 to the clamper as the clamp circuit. With this gain control, aliasing distortion during luminance signal synthesis can be suppressed to a minimum. Next, the output signal of this gain control circuit 50 is sent to a switch circuit 60, which is a sequential means for forming a luminance signal at the next stage according to the present invention, and undergoes normal signal processing such as gamma correction or white clipping, and is converted into an NTSC signal. The output signal is guided to a process encoder circuit 70 in which a circuit is integrated. Next, the operation of the switch circuit 60 will be explained based on FIG. The illustrations S1, S2, and S3 are shown in Figure 7.
This is the output signal of the CCD. In this example, it is assumed that the drive pulse for the horizontal shift register is a three-phase drive pulse equivalent to the signal waveform shown in FIG.

When these signals S1, S2, and S3 are extracted by the switch pulses of the control signals SW-R, SW-G, and SW-B of the switch circuit, and the extracted signals are added, a luminance signal shown as Y in the figure is obtained. That is, the same signal Y as the spatial sampling of the color separation filter is obtained, and the resolution is very good. In this way, when a luminance signal is generated by extracting and adding only the portions necessary as a luminance signal by switching, noise addition is eliminated and there is no deterioration of S/N.

Reference numeral 55 in FIG. 9 is a white balance circuit, and any circuit may be used as long as it is intended to substantially match the R, G, and B signal levels of the luminance signals. Generally, white balance of color signals requires 5 to 6 bit information, but in this example, process information is required.
It is built into the encoder circuit. With this configuration, when the color filter is replaced with a complementary color filter, the circuit can be considered separately for the luminance system and the color system, resulting in a simple circuit.

The embodiment shown in FIG. 11 is a circuit block diagram when the horizontal shift register of the CCD shown in FIG. 5 is driven with in-phase pulses.

In this case, when combining the luminance signals, R, G,
In order to make the B signal the same as the spatial sampling of the color separation filter, the G and B signals need to be delayed with respect to the R signal. The delay circuit may be a normal delay line or a sample and hold circuit. If the driving methods differ between the horizontal shift registers in this way, circuit means for returning to the original spatial sampling state is required. This circuit means may be located at any position before the means for synthesizing luminance signals.

Next, FIG. 12 is a diagram showing an embodiment of a KNEE circuit as a nonlinear suppression means for expanding the dynamic range of the circuit. If the KNEE circuit is inserted after matching the levels of the R, B, and G signals as in this embodiment, the signal level lower than the white clip level V _WC , that is, between the signal levels to be used, as shown in FIG. In this case, since the signal levels of R', G', and αB' match, no problem occurs in the luminance signal or color signal. The embodiment shown in FIG. 12 is a switch circuit 60.
This is an example in which the KNEE circuit 65 is inserted in the latter stage of the
Figure 13 shows the stage after the gain control circuit 50.
7 is a diagram showing an example in which a KNEE circuit 65 is inserted. FIG.
In this case, only one KNEE circuit is required, which simplifies the configuration.

(Effects) As explained above, the following effects can be obtained by combining the luminance signals from the signal components after level matching of the output signals of the imaging device.

First, since there is no need to use a color temperature correction filter, the configuration becomes simpler.

Second, by matching the levels of each color signal that makes up the luminance signal, aliasing distortion can be significantly reduced, and as a result, a high-bandwidth video signal can be obtained.

Thirdly, the S/N ratio is improved because the luminance signals are synthesized by switching only the effective components of each color signal and converting them into point-sequential signals.

Furthermore, since the KNEE characteristic is applied after level matching of each color signal, a video signal with a wide dynamic range and no false signals can be obtained.

Although this embodiment has been described using a solid-state image pickup device, it may be of an image pickup tube type, or may be configured using other multiple output methods instead of the three output method.

[Brief explanation of the drawing]

Figure 1 is a schematic diagram of a color separation filter, Figure 2A,
B and C are illustrations of aliasing distortion, Fig. 3 is a spectral energy diagram due to differences in imaging light sources, Fig. 4 is an illustration of aliasing distortion due to unbalance of point-sequential signal levels, and Fig. 5 is a luminance signal of the conventional method. An explanatory diagram of the synthesis method, Figures 6A and B are explanatory diagrams of the problems caused by adding KNEE characteristics in the conventional method, and Figure 6C is an explanatory diagram of the problems caused by adding the KNEE characteristic in the conventional method.
Figure 7 is a schematic diagram of the CCD used in the embodiment of the present invention, Figure 8 is a diagram explaining the transfer of each color signal to the horizontal shift register, Figure 9 is the imaging device of the present invention. A first embodiment diagram of
FIG. 10 is an explanatory diagram of the luminance signal synthesis method in FIG. 9, FIG. 11 is a diagram of the second embodiment of the present invention, and FIG.
2 and 13 are diagrams of embodiments of the present invention in which a KNEE circuit is inserted. 10 is an imaging device, 40 is a clamp circuit, 5
0 is a gain control circuit, 60 is a switch circuit, 45 and 46 are delay circuits, and 65 is a KNEE circuit.

Claims (1)

[Scope of Claims] 1. An imaging means combined with a plurality of color separation filters having a plurality of different color spectral characteristics, and a signal in the row direction of the imaging means that is distributed to each color signal corresponding to each of the color separation filters. a plurality of signal readout paths for respectively outputting the signals; a level control means for respectively controlling the level of each color signal outputted via the plurality of signal readout paths; non-linear processing means for performing predetermined non-linear processing on each color signal; and point-sequential processing by extracting each color signal whose level is controlled by the level control means and non-linearly processed by the non-linear means using a switch means. 1. An imaging device comprising: point sequentialization means for forming a broadband luminance signal.