JPH06315154A - Color imaging device - Google Patents

Abstract

(57) [Abstract] [Purpose] To improve the resolution without increasing the pixel density of the solid-state imaging device or increasing the number of pixels. [Structure] The image that has passed through the lens 2 is decomposed by a dichroic prism 4, and red, green, and blue subject images are obtained by the CCD 11
Images are formed on R, 11G, and 11B. The CCD 11G is left by the piezoelectric element 3 in the horizontal direction and the vertical direction by a length that is an integral fraction of the pixel pitch, and the interpolation processing unit 5 is provided.
It is characterized in that the still image is made high-definition by synthesizing by G, the frame synthesizing unit 7, and the image synthesizing unit 8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image input technique, and more particularly to a high-definition color image pickup apparatus using a color separation prism and a plurality of solid-state image pickup elements.

[0002]

2. Description of the Related Art Conventionally, as a method for picking up a color image with high precision, a method using a color separation prism called a dichroic prism has been used. This method
The light after passing through the lens is red by the dichroic prism,
In this method, a plurality of color components such as green and blue are separated, subject images for each color component are formed at a plurality of positions, and the plurality of subject images are captured by a plurality of solid-state image pickup devices.

FIG. 16 shows the structure of a conventional color image pickup device using a dichroic prism. In FIG. 16, 4 is a dichroic prism, 4R, 4G and 4B are end faces of the dichroic prism 4 on the exit side, 1R and 1R.
G and 1B are solid-state image pickup devices (collectively referred to as 1),
2 is a lens. The subject images of the respective color components are formed on the exit side end faces 4R, 4G, 4B of the dichroic prism 4, respectively. Then, these subject images are picked up by the solid-state image pickup devices 1R, 1G, and 1B attached to the emission side end faces 4R, 4G, and 4B with an adhesive or the like, and an image signal for each color component is generated.

[0004]

The resolution of the color image pickup device using the conventional dichroic prism 4 is as follows.
It depends on the number of pixels of the solid-state image sensor 1 to be used. As a method of increasing the number of pixels of the solid-state imaging device 1, a method of making full use of a microfabrication technique and increasing the pixel density is the mainstream, but there is also a method of increasing the size of the solid-state imaging device 1 to increase the number of pixels. However, the method of increasing the pixel density has a problem that the process becomes difficult, and further, there is a problem that the sensitivity decreases due to the reduction of the pixel area. In recent years, H
Although the solid-state image sensor 1 of 2 million pixels for DTV has been developed, this number of pixels is close to the limit, especially in consideration of the decrease in sensitivity, and it is not possible to further increase the number of pixels to improve the resolution at present. It has become difficult.

On the other hand, the solid-state image sensor 1 without changing the pixel density
The method of increasing the number of pixels to increase the number of pixels lowers the manufacturing yield and increases the cost. For this reason, its applications are limited to scientific purposes such as astronomy and military purposes. In addition to the above problem, in the conventional method for increasing the resolution by increasing the number of pixels in one solid-state image sensor 1, the driving speed of the solid-state image sensor 1 and the A / The problem of increasing the D conversion speed and the like is becoming more serious.

It is an object of the present invention to provide a color image pickup device having an improved resolution without increasing the image density of the solid-state image pickup element or increasing the number of pixels.

[0007]

In a color image pickup apparatus according to the present invention, at least one solid-state image pickup element is synchronized with a field or frame period, and the length of the pixel pitch is divided by an integer, which is a horizontal direction or a vertical direction. Or a shift means for shifting in both horizontal and vertical directions (hereinafter referred to as an image shift method). Here, the solid-state image pickup device to which the image shift method is applied is not fixed to the exit side end face of the dichroic prism.

Further, according to the present invention, a moving area detecting means for detecting a moving portion from the picked-up image is provided.

Further, the solid-state image pickup device to which the above image shift method is applied is used for picking up a subject image of a color component which is closest to a luminance component among the color-separated color components.

[0010]

In the present invention, a case will be described in which red, green and blue subject images separated by the dichroic prism are picked up by three solid-state image pickup devices. In this case, the color component closest to the luminance component is the green component, and the solid-state imaging device to which the image shift method is applied captures a subject image of the green component.

According to the present invention, the green component image is made finer by the image shift method. FIG. 15 shows a principle diagram of high definition by the image shift method. In the solid-state image sensor,
Photoelectric conversion units (photosensitive units) that read two-dimensional optical image information and convert them into electric signals are discretely arranged as shown in FIG. 15, and do not directly contribute to imaging between adjacent photoelectric conversion units. There is a non-photosensitive area. The image shift method is a method for increasing the definition of an input image by utilizing the existence of this non-photosensitive portion and doubling the sampling points of the image.

That is, the relative positional relationship between the image input in a certain frame (FIG. 15A) and the subject image and the pixels of the solid-state image sensor in the next frame is 1/2 of the pixel pitch.
By combining the shifted input images (Fig. 15 (b)),
In each frame image, the optical image information of the area corresponding to the non-photosensitive portion of the solid-state image sensor is sampled in the other frame image, and as a result, as shown in FIG.
As shown in (c), this is equivalent to effectively reading an image with double the pixel density, and high definition can be achieved. Figure 15
For the sake of simplicity, the method of interpolating two images in the horizontal direction has been described. However, high definition can be achieved by interpolation in both the horizontal direction and the vertical direction, and furthermore, 1/3 of the pixel pitch is possible.
It is possible to increase the density three times in one direction by shifting the positions one by one.

The image shift method is a method for realizing high definition by synthesizing a plurality of frames (or fields) as described above, and therefore cannot be applied because a multiple image is generated in the moving area. Therefore, in the present invention, the moving area detecting means is used, and the high definition method by the image shift method is applied only to the stationary area. As a result, only the still area is made high definition, but considering that the visual characteristics of the human eye are inferior to the still area in the visual characteristics of the moving area, only the still area is made high definition. As a result, the entire screen becomes high definition visually.

Further, in the present invention, only the static area of the green component is made high definition, but it is generally known that a high definition image can be visually reproduced by making only the luminance component high definition in a color image. ing. Therefore, a high-definition color image can be output by making only the color component image close to the luminance component high definition as in the present invention.

The present invention focuses on the fact that the required resolution is different between the moving area and the still area, and that the resolution of the color image largely depends on the resolution of the green component. An image is captured with characteristics suitable for components, and a high-definition color image pickup device is provided by a solid-state image pickup element having a small number of pixels.

[0016]

【Example】

[Embodiment 1] FIG. 1 shows a schematic configuration of a first embodiment of the present invention. In FIG. 1, 11R, 11G, and 11B are CCDs.
Area sensor (hereinafter, simply referred to as CCD), 2 is a lens, 3 is a piezoelectric element, 4 is a dichroic prism, 4
R, 4G and 4B are exit side end faces of the dichroic prism 4, 5R, 5G and 5B are interpolation processing units, 6R and 6G,
6B is a moving area detection unit, 7 is a frame composition unit, 8 is an image composition unit, and 9R, 9G and 9B are output terminals. Note that FIG.
In the above, the amplifier and the A / D converter are omitted.

In the first embodiment, a CCD is used as a solid-state image sensor.
Is used. Of the three CCDs, the CCDs 11R and 11B for picking up the red component image and the blue component image are adhered and fixed to the emission side end faces 4R and 4B of the dichroic prism 4, as in a normal three-plate CCD camera. CCD 11G
The piezoelectric elements 3 are attached to the upper, lower, left and right sides and are arranged close to the emission side end surface 4G of the dichroic prism 4. With this piezoelectric element 3, the CCD 11G is vertically and horizontally shifted by a width of 1/2 of the pixel pitch in synchronization with the frame period.

The signals picked up by the CCDs 11R, 11G and 11B have a pixel density of 4 times the number of pixels picked up by the CCD (2 times each in the horizontal and vertical directions) by interpolation processing in the interpolation processing sections 5R, 5G and 5B, respectively. To do. Each interpolation processing unit 5
The interpolating method in R, 5G, and 5B interpolates in the field in the horizontal direction, while using the signals from the moving region detection units 6R and 6B in the vertical direction, interpolates in the field in the moving region and in the still region. Performs static / dynamic adaptive interpolation processing that is inter-field interpolation.

Further, the image signal from the CCD 11G is combined by the frame combining section 7. FIG. 2 shows a synthesizing method in the frame synthesizing unit 7. FIG. 2A is an array diagram of the photoelectric conversion units of the CCD, which are shifted in synchronization with the frame cycle by repeating the left, bottom, right, and top by the image shift method. Therefore, when the images captured in each frame are combined by the image combining unit 8, the positional relationship of pixels in the combined image is as shown in FIG. here,
If P1 is a pixel of the image captured in the nth frame, P
2 is a pixel in the (n + 1) th frame, P3 is a pixel in the (n + 2) th frame, and P4 is a pixel in the (n + 3) th frame.

The image synthesizing section 8 receives the image signal from the interpolation processing section 5G and the image signal from the frame synthesizing section 7, and further receives the motion detection signal from the moving area detecting section 6G. When the motion detection signal is large, the image signal from the interpolation processing unit 5G is output, and when the motion detection signal is small, the image signal from the frame synthesizing unit 7 is output. The signal combined by the image combining unit 8 is output from the output terminal 9G as an output signal of the green component image. On the other hand, the image signals of red and blue components are interpolated by the interpolators 5R and 5B, and then output from the output terminals 9R and 9B, respectively.

A known method used in NTSC is used as a method of generating a motion detection signal in the motion area detection units 6R, 6G and 6B. In addition to the method described above, a method described in a document (see Japanese Patent Application No. 5-22204) can be used as a method of generating a motion detection signal in the moving area detection unit 6G.

The signal from the frame synthesizing unit 7 input to the image synthesizing unit 8 is a high definition green component image,
Since this is output when the motion detection signal is smaller than a certain level, the still region of the green component image is made finer. [Embodiment 2] FIG. 3 shows a schematic configuration of a second embodiment of the present invention. The optical system configuration of the lens 2, dichroic prism 4, CCD 11R, 11G, 11B, piezoelectric element 3 and the like in this embodiment is the same as that in the first embodiment. In this embodiment, the output signals of the moving area detection units 6R and 6B are used as the motion detection signals used by the image composition unit 8. That is, the motion detection signal obtained from the red component image and the blue component image is used. Then, when one of these two motion detection signals is large, the image signal from the interpolation processing unit 5G is output, and when the motion detection signal is small, the image signal from the frame synthesis unit 7 is output. The signal combined by the image combining unit 8 is output as an output signal of the green component image from the output end 9G.

In the second embodiment, the same operation as in the first embodiment is performed except for the above points. [Embodiment 3] FIG. 4 shows a schematic configuration of a third embodiment of the present invention. In FIG. 4, 111G and 112G are CCD and 4
1 is a dichroic prism, 41R, 41B, 411
G and 412G are the exit side end faces of the dichroic prism 41. It should be noted that details of the configuration of the dichroic prism 41 are omitted in FIG.

In this embodiment, a dichroic prism 41 for splitting incident light into red, blue and two green lights is used. That is, the subject images of red, blue and two green components are transferred to the four emission side end faces 41 of the dichroic prism 41.
An image is formed on R, 41B, 411G, 412G, and this is imaged by four CCDs 11R, 11B, 111G, 112G. One of the two CCDs 112G for picking up the green component image has the piezoelectric elements 3 attached to the top, bottom, left and right, and is arranged close to the emission side end surface 412G of the dichroic prism 41. In addition, another CCD 111 for capturing a green component image
CCD 1 for capturing G, red component images, and blue component images
1R and 11B are adhered and fixed to the exit side end faces 411G, 41R and 41B of the dichroic prism 41, respectively.

The CCD 112G is vertically arranged by the piezoelectric element 3,
The width is shifted in the horizontal direction by half the pixel pitch in synchronization with the frame period. Then, the image synthesizing unit 7 synthesizes the image signals in the same manner as in the first embodiment.

The signals picked up by the CCDs 11R, 111G, and 11B are made to have the same pixel density as the image output from the frame synthesizing unit 7 by the interpolation processing in the interpolation processing units 5R, 5G, and 5B, respectively. The interpolation method in the interpolation processing units 5R and 5B is in-field interpolation in the horizontal direction,
On the other hand, in the vertical direction, static / moving adaptive type interpolation processing is performed by using the signals from the moving area detecting units 6R and 6B and performing inter-field interpolation in the moving area and inter-field interpolation in the still area. Further, in the processing by the interpolation processing unit 5G, the intra-field interpolation processing is performed on the entire screen regardless of the moving area or the stationary area.

The image synthesizing section 8 receives the image signal from the interpolation processing section 5G and the image signal from the frame synthesizing section 7, and further receives the motion detection signal from the moving area detecting section 6G. When the motion detection signal is large, the image signal from the interpolation processing unit 5G is output, and when the motion detection signal is small, the image signal from the frame synthesizing unit 7 is output. The signal synthesized by the image synthesizer 8 is output as a green component image output signal from the output terminal 9G. On the other hand, the image signals of red and blue components are interpolated by the interpolators 5R and 5B, and then output from the output terminals 9R and 9B, respectively.

A known method used in NTSC is used as a method of generating motion detection signals in the moving area detecting units 6R and 6B. The signal from the frame synthesizing unit 7 input to the image synthesizing unit 8 is a high definition green component image,
Since this is output when the motion detection signal is small,
The static area of the green component image is made finer. [Embodiment 4] FIG. 5 shows a schematic configuration of a fourth embodiment of the present invention. Lens 2, dichroic prism 41, CCDs 11R, 11B, 111G, 112 in this embodiment
The optical system configuration of G, the piezoelectric element 3, etc. is the same as that of the third embodiment. In the present embodiment, in the interpolation processing by the interpolation processing unit 5G, in addition to the image signal of the CCD 111G, the CCD 112G
The image signal picked up in is also used. The interpolation processing method in the interpolation processing unit 5G will be described with reference to FIGS.

In FIGS. 6 to 9, the pixels shown by the solid squares are the pixels of the green component image picked up by the CCD 111G, and the pixels shown by the circle are the pixels picked up by the CCD 112G. Also, the shaded and shaded squares and circles are the pixels sampled in each field. Further, in FIGS. 6 to 9, Pq
(m, n) (q = 1 to 4, m = i, i + 1, n = j, j + 1) indicates a pixel position. Here, i and i + 1 are CCDs 11R, 11B and 111, respectively.
G and 112G are horizontal pixel addresses, and j and j + 1 are vertical pixel addresses (line numbers). Further, in FIGS. 6 to 9, k, k + 1 ... Are line numbers in the image after the interpolation processing.

Since the CCD 112G takes an image by the image shift method, the pixel position is different for each frame. Now, it is assumed that the pixel arrangement in the nth frame is the same as the pixel position of the CCD 111G (position indicated by P1 (m, n)). However, it is assumed that the CCD 111G and the CCD 112G are arranged so that the pixel array is shifted by one line. The pixel arrangement in the nth frame is shown in FIG. FIG. 6A shows an odd field pixel arrangement, and FIG. 6B shows an even field pixel arrangement. Since both CCDs 111G and 112G are arranged one line apart, the CCD 111G is
In contrast, the pixel of the j-th line is sampled, whereas the pixel of the j + 1-th line is sampled in the CCD 112G. In the odd field of the nth frame, the interpolation processing unit 5G outputs k, k + 2, ... Line images. Therefore, in the output image from the interpolation processing unit 5G, the signals from both CCDs 111G and 112G are output as the signals of the odd line and the even line, respectively. However, since there is no sampling data in the pixel indicated by P2 (m, n), interpolation is performed by taking the average value of the pixels indicated by P1 (m, n) on the left and right and setting it as the value of that pixel.

FIG. 6B shows the pixel arrangement in the even field of the nth frame. CCD11 in this field
In 1G, the pixels in the j + 1th line are sampled, and in the CCD 112G, the pixels in the jth line are sampled. However, in this field, the interpolation processing unit 5
The lines output from G are k + 1, k + 2, ...
Pixels on this line (pixels indicated by P3 (m, n) and P4 (m, n)) are not sampled. Therefore,
From the pixels sampled by the interpolation method according to the equations (1) and (2), the values of the pixels indicated by P3 (m, n) and P4 (m, n) are obtained and output.

[0032]

[Expression 1] P3 (i, j) = {P1 (i, j) + P1 (i + 1, j) + P1 (i, j + 1) + P1 (i + 1, j + 1)} / 4 …… (1) P4 (i, j) = {P1 (i, j) + P1 (i, j + 1)} / 2 …… (2) In the odd field of the (n + 1) th frame in FIG. The pixel arrangement is shown. In this field, the pixels of the jth line are sampled by the CCD 111G, and the CCD 11G
In 2G, the pixel of the j + 1th line is sampled. Here, the CCD 111G samples the same pixel (pixel indicated by P1 (m, n) on the j-th line) as in the odd field of the previous frame, whereas the CCD
In 112G, the pixel position is CCD by the image shift method.
The pixel is shifted to the right by ½ pixel pitch (one pixel pitch on the output side) on the image pickup plane, and the sampling pixel becomes the pixel indicated by P2 (m, n) on the j + 1th line. In this field, an image of k, k + 2, ... Lines is output from the interpolation processing unit 5G. Therefore, in the pixel indicated by P2 (m, n) in the k line in the output image, the left and right P1
The average value of the pixels represented by (m, n) is taken as the value of that pixel, while the pixel represented by P1 (m, n) on the k + 2 line is the pixel value represented by P2 (m, n) on the left and right. The average value is taken as the value of that pixel.

FIG. 7B shows the pixel arrangement in the even field of the (n + 1) th frame. CCD in this field
111G samples the pixels on the j + 1th line, and CCD 112G samples the pixels on the jth line. However, the lines output from the interpolation processing unit 5G in this field are k + 1, k + 3, ..., And the pixels (pixels indicated by P3 and P4) on this line are not sampled. Therefore, the values of the pixels indicated by P3 and P4 are obtained from the sampling pixels according to the equations (3) and (4) and output.

[0034]

[Equation 2] P3 (i, j) = {P1 (i, j) + P1 (i + 1, j) +2 x P2 (i, j + 1)} / 4 (3) P4 (i, j) = {2 x P1 (i, j) + P2 (i-1, j + 1) + P2 (i, j + 1)} / 4 (4) In the (n + 2) th frame of FIG. The pixel arrangement in an odd field is shown. In this field, the pixels of the jth line are sampled by the CCD 111G, and the CCD 11G
In 2G, the pixel of the j + 1th line is sampled. The CCD 112G samples the same pixel (pixel indicated by P1 (m, n) on the j-th line) as in the odd field of the previous frame, whereas the CCD 112G
Then, the pixel position is shifted downward by 1/2 pixel pitch (1 pixel pitch on the output side) on the image pickup surface of the CCD by the image shift method, and the sampling pixel is P3 (m, n) of the j + 1th line. It becomes the pixel shown. In this field, an image of k, k + 2, ... Lines is output from the interpolation processing unit 5G. Therefore, P in the k line in the output image
For the pixel indicated by 2 (m, n), the average value of the pixels indicated by P1 on the left and right is taken as the value of that pixel, while it is indicated by P1 (m, n) and P2 (m, n) on the k + 2 line. For pixels, the values of these pixels are obtained from the sampling pixels according to equations (5) and (6) and output.

[0035]

[Equation 3] P1 (i, j + 1) = {P3 (i-1, j + 1) + P3 (i, j + 1) + P1 (i, j)} / 3 (5) P2 ( i, j + 1) = {4 x P3 (i, j + 1) + P1 (i, j) + P1 (i + 1, j)} / 6 (6) The n + 2th pixel in FIG. 8B. The pixel arrangement in the even field of the frame is shown. In this field, the pixels of the j + 1th line are sampled by the CCD 111G,
In 112G, the pixel of the jth line is sampled. In this field, from the interpolation processing unit 5G to k + 1, k
Images of +3, ... Lines are output, but the pixels indicated by P3 and P4 in these lines are not sampled. Therefore, for the pixel indicated by P4 on the k + 1 line, the average value of the pixels indicated by P3 on the left and right is taken as the value of that pixel, while for the pixel indicated by P3 and P4 on the k + 3 line, equations (7) and (8) are used. Then, the values of these pixels are obtained from the sampling pixels and output.

[0036]

[Equation 4] P3 (i, j + 1) = {P1 (i, j + 1) + P1 (i + 1, j + 1) + P3 (i, j + 2)} / 3 (7) P4 (i, j + 1) = {4 x P1 (i, j + 1) + P3 (i-1, j + 2) + P3 (i, j + 2)} / 6 (8) FIG. 9 (A) shows a pixel arrangement in the odd field of the (n + 3) th frame. In this field, the pixels of the jth line are sampled by the CCD 111G, and the CCD 11G
In 2G, the pixel of the j + 1th line is sampled. The CCD 111G samples the same pixel (pixel indicated by P1 (m, n) on the j-th line) as in the odd field of the previous frame, whereas the CCD 112G
Then, the pixel position is shifted to the left by ½ pixel pitch (one pixel pitch on the output side) on the image pickup surface of the CCD by the image shift method, and the sampling pixel becomes the pixel indicated by P4 on the j + 1th line. In this field, an image of k, k + 2, ... Lines is output from the interpolation processing unit 5G. Therefore, P2 (m,
In the pixel indicated by n), the average value of the pixels indicated by P1 (m, n) on the left and right is taken as the value of that pixel, while k + 2
For the pixels indicated by P1 (m, n) and P2 (m, n) in the line, the values of these pixels are obtained from the sampling pixels according to equations (9) and (10) and output.

[0037]

[Equation 5] P1 (i, j + 1) = {P1 (i, j) +2 x P4 (i, j + 1)} / 3 (9) P2 (i, j + 1) = {P1 (i, j) + P1 (i + 1, j) + 2xP4 (i, j + 1) + 2xP4 (i + 1, j + 1)} / 6 (10) FIG. 9B shows the n + 3. The pixel arrangement in the even field of the frame is shown. In this field, the pixels of the j + 1th line are sampled by the CCD 111G,
In 112G, the pixel of the jth line is sampled. In this field, from the interpolation processing unit 5G to k + 1, k
Images of +3, ... Lines are output, but the pixels indicated by P3 and P4 in these lines are not sampled. Therefore, for the pixel indicated by P4 on the k + 1 line, the average value of the pixels indicated by P3 on the left and right is taken as the value of that pixel, while for the pixel indicated by P3 and P4 on the k + 3 line, equations (11) and (12) are used. Then, the values of these pixels are obtained from the sampling pixels and output.

[0038]

[Equation 6] P3 (i, j + 1) = (P4 (i, j + 2) + P4 (i + 1, j + 2) + 2xP1 (i, j + 1) + P1 (i + 1, j +1)} / 6 (11) P4 (i, j + 1) = {2xP1 (i, j + 1) + P4 (i, j + 2)} / 3 (12) More than 4 frames ( Processing of 8 fields) is sequentially repeated. In the fourth embodiment, the interpolation processing unit 5 described above is used.
Same as Example 1 except for the treatment with G. In this embodiment, since the interpolation processing uses the signal from the CCD 112G,
The moving area of the green component image can also be made fine. [Embodiment 5] FIG. 10 shows a schematic configuration of a fifth embodiment of the present invention.
Shown in. Lens 2, dichroic prism 41, CCDs 11R, 11B, 111G, 112 in this embodiment
The optical system configuration of G, the piezoelectric element 3, etc. is the same as that of the third embodiment. In the fifth embodiment as well, in the interpolation processing by the interpolation processing unit 5G, in addition to the image signal of the CCD 111G, the CCD 112
The image signal captured by G is also used. Further, in the present embodiment, the frame synthesizing unit 7 performs frame synthesis by using the green component image signal from the CCD 111G in addition to the image signal of the CCD 112G.

FIG. 11 shows an image shift method using the CCD 112G in the fifth embodiment. In FIG. 11, the pixels indicated by the solid squares are the pixels of the green component image captured by the CCD 111G, and the pixels indicated by the circles are the CCD.
It is a pixel imaged at 112G. In the image shift method used in each of the above embodiments, 1/2 of the pixel pitch is shifted to the right,
While one pixel samples four positions in four frames by shifting in the order of bottom, left, and top, in the fifth embodiment, the CCD111 is shifted in the order of bottom, left, and top right.
Three positions are sampled in three frames except the position sampled by G. And frame composition is CC
3 frames of pixels and CC imaged by D112G
This is performed by synthesizing pixels for one frame captured in D111G.

The operation of the fifth embodiment is the same as that of the fourth embodiment except for the processing in the frame synthesizing section 7. [Embodiment 6] FIG. 12 shows a schematic configuration of a sixth embodiment of the present invention.
Shown in. In FIG. 12, 7R and 7B are inter-field interpolation processing units, 8R, 8G and 8B are image combining units, and
R, AB and 6RGB are adders, and HF is a high pass filter. The optical system configuration of the lens 2, dichroic prism 4, CCD 11R, 11G, 11B, piezoelectric element 3 and the like in the sixth embodiment is the same as that of the first embodiment. In each of the above embodiments, the high definition of the still area of the red and blue component images is realized by the interpolation between the two fields of the current field and the previous field, but in the sixth embodiment, the signal after this interpolation processing is green. By adding the high frequency component of the component,
Realizes higher-definition red and blue component images. That is, of the green component image signals synthesized by the frame synthesizing unit 7G, the signals that have passed through the high-pass filter HF are converted into blue component and red component image signals that are interpolated by the inter-field interpolation processing units 7R and 7B, respectively. Adder AR, A
Add with B. This process is a process for realizing high definition of the red and blue component images by utilizing the correlation between the high frequency components of the red and blue component images and the green component image.

In order to perform the above-mentioned high-definition processing, the condition is that the target pixel is not in the moving region from the present field to the past 8 fields. Therefore, the moving area detection unit 6
For R and 6B, it is not possible to use the known method used in NTSC or the like as in the above-mentioned embodiment. Therefore, in this embodiment, all the brightness changes between the past eight fields are detected. FIG. 13 shows the moving area detection unit 6R in the sixth embodiment.
The calculation method of the motion detection signal in and 6B is shown. Now, assuming that the current field is the m-th field, P1 (i,
It is assumed that the motion coefficient of the pixel indicated by j) is calculated. The CCDs 11R, 11 for the red and blue component images
Since B is fixed, only the pixels indicated by the solid squares in FIG. 14 are sampled. The signal value of the pixel P1 (i, j) in the m-th field is P1 (i, j)
If represented by _m , the moving area detection unit 6R in the present embodiment.
The motion detection signal k _R in the above equation is given by the equation (13).

[0042]

[Formula 7] k _R = | P1 (i, j) _m −P1 (i, j) _m-2 | + | P1 (i, j) _m-2 −P1 (i, j) _m-4 | + | P1 (i, j) _m-4 −P1 (i, j) _m-6 ｜ ＋ ｜ P1 (i, j) _m-6 −P1 (i, j) _m-8 ｜ ＋ ｜ P1 (i, j) _m-1 −P1 (i, j) _m-3 ｜ ＋ ｜ P1 (i, j) _m-3 −P1 (i, j) _m-5 ｜ ＋ ｜ P1 (i, j) _m-5 −P1 ( i, j) _m-7 | (13) However, since all CCDs are driven by the interlace method, if the j line including the pixel P1 (i, j) is scanned in the mth field, m-1, m-3, m-5, m-7
Since the j line is not scanned in each field of, the signal value of the pixel P1 (i, j) cannot be obtained. Therefore, in these fields, the pixel P1 (i, j-1) on the upper and lower lines scanned in that field
And the average value of the signal values of P1 (i, j + 1) is used. In addition, the motion detection signal k _B in the motion area detection unit 6B
Is _{calculated in the} same way as k _R.

In the sixth embodiment, in order to make the moving regions the same in the red, green and blue component images, one moving detection signal k is used to determine the moving region and the still region of each color component image. The motion detection signal k is calculated by the equation (14).

[0044]

K = k _R + k _G + k _B (14) Using the motion detection signal k, the image combining units 8R, 8G, and 8B combine the still area image and the moving area image, respectively.

Also in this embodiment, since the CCDs for picking up the red and blue component images are fixed, these images do not contain high frequency components, but the correlation with the high frequency components of the green component image is utilized. By adding this, high definition of red and blue component images is realized.

It is needless to say that various modifications can be made without departing from the spirit of the present invention. For example, although the CCD is used as the solid-state image pickup device in the above-described embodiment, it can be implemented regardless of the type of the solid-state image pickup device.

Further, in the above-mentioned embodiment, the green component image of the static area is generated by synthesizing an image obtained by shifting the CCD by ½ of the pixel pitch in one direction, but it is ⅓ of the pixel pitch.
Alternatively, it is also possible to shift by 1/4 and combine images with higher resolution.

In the above embodiment, the piezoelectric element 3 is used for the CCD.
The image of the object is shifted on the image pickup device by using the image shift method by the shift means for vibrating 11R, 11B, 11G, 111G, 112G and the like. For example, the transparent plate is vibrated between the dichroic prism and the image pickup device. The present invention can be implemented by using any shifting means such as a method.

Further, in the above embodiment, the 3CCD system and the 4CCD system using the dichroic prism have been described, but the present invention can also be implemented by a method in which two CCDs are used to capture images of red and blue components. Is possible.

[0050]

As described above, according to the present invention, in a color image pickup apparatus having a color separation prism and a plurality of solid-state image pickup elements, at least one solid-state image pickup element is synchronized with a field or frame period to reduce the pixel pitch. Since it has a shift means for shifting in the horizontal direction or the vertical direction, or both the horizontal direction and the vertical direction by the length of a fraction of an integer, it is possible to capture an image with characteristics suitable for each area and color. That is, with respect to a still area that determines the definition of an image, it is possible to take a sufficiently high-definition image by taking time even if an image pickup device having a small number of pixels is used. Further, the moving area can be picked up at the same speed as the conventional moving image pick-up, and a natural moving picture can be picked up. From the above,
The color image pickup device of the present invention can pick up a high-definition moving image even if a solid-state image pickup device having a small number of pixels is used. Therefore, a high-definition moving image pickup device such as an HDTV can be realized at low cost. Furthermore, by implementing the present invention using a high-resolution image pickup device for HDTV or the like, it becomes possible to take an ultra-high-definition moving image that has never been seen before.

Since the present invention is provided with the image synthesizing means for synthesizing a plurality of continuous frame or field images and the moving area detecting means for detecting the moving area from the picked-up image, the moving area can be detected and taken out. Therefore, when transmitting to the receiving side, it is sufficient to send only the signal in the moving region and the still region which were the moving region one frame before, and the transmitted signal can be easily compressed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a first embodiment of the present invention.

FIG. 2 is a diagram showing a frame synthesizing method in the first embodiment.

FIG. 3 is a block diagram showing a schematic configuration of a second embodiment of the present invention.

FIG. 4 is a block diagram showing a schematic configuration of a third embodiment of the present invention.

FIG. 5 is a block diagram showing a schematic configuration of a fourth exemplary embodiment of the present invention.

FIG. 6 is a diagram illustrating a pixel positional relationship between a captured image and an output image according to a fourth embodiment of the present invention for each frame and field.

FIG. 7 is a diagram illustrating a pixel positional relationship between a captured image and an output image according to a fourth embodiment of the present invention for each frame and field.

FIG. 8 is a diagram showing a pixel positional relationship between a captured image and an output image according to a fourth embodiment of the present invention for each frame and field.

FIG. 9 is a diagram showing a pixel positional relationship between a captured image and an output image according to a fourth embodiment of the present invention for each frame and field.

FIG. 10 is a block diagram showing a schematic configuration of a fifth exemplary embodiment of the present invention.

FIG. 11 is an explanatory diagram of an image shift method used in the fifth embodiment.

FIG. 12 is a block diagram showing a schematic configuration of a sixth embodiment of the present invention.

FIG. 13 is a diagram showing a method of calculating a motion detection signal in a motion area detection unit in the sixth embodiment.

FIG. 14 is a diagram showing a pixel positional relationship between a captured image and an output image according to the sixth embodiment.

FIG. 15 is a diagram illustrating the principle of resolution enhancement by the image shift method according to the present invention.

FIG. 16 is a schematic configuration diagram of a color imaging device using a conventional dichroic prism.

[Explanation of symbols]

 2 lens 3 piezoelectric element 4 dichroic prism 5R interpolation processing section 5G interpolation processing section 5B interpolation processing section 6R moving area detecting section 6G moving area detecting section 6B moving area detecting section 7 frame combining section 8 image combining section 9R output terminal 9G output terminal 9B Output terminal 11R CCD area sensor 11G CCD area sensor 11B CCD area sensor 41 Dichroic prism

Claims (2)

[Claims] 1. A color image pickup apparatus comprising a color separation prism and a plurality of solid-state image pickup devices, wherein at least one solid-state image pickup device is synchronized with a field or frame period, and is horizontally divided by an integer fraction of a pixel pitch. A color image pickup device comprising a shift means for shifting in a vertical or vertical direction, or in both horizontal and vertical directions. 2. A color image pickup apparatus according to claim 1, further comprising: an image synthesizing unit for synthesizing a plurality of continuous frame or field images, and a moving region detecting unit for detecting a moving region from the picked-up image. .