JPH07123302B2 - Color video camera - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color video recording device for ultra-high-speed photography, which is equipped with a light intensifying device, and more particularly, to ultra-high-speed moving objects such as rockets, explosion, destruction, turbulence, discharge phenomenon, and chemical reaction. It is preferably used for scientific measurement of phenomena and movement of microorganisms under a microscope.

[0002]

2. Description of the Related Art To photograph the above phenomenon, it is necessary to photograph with an ultra-high speed camera. In ultra-high-speed shooting, the exposure time becomes very short, so the amount of light per screen may drop, and shooting may not be possible. A typical example is photographing a dynamic phenomenon under a microscope. Under a microscope, the amount of light is originally small, and even a small movement becomes very fast as the magnification becomes higher. Therefore, in order to respond to this movement, the number of shooting screens per unit time, that is,
When the frame rate is increased, the exposure time is shortened and the light quantity is further limited.

A microchannel plate type image intensifier for enhancing weak light has been recently used as a means for attaining high sensitivity under the condition of a small amount of light as described above.
(Hereinafter referred to as MCP type II) is provided.

The above MCP type II has the structure shown in FIG. 11, and the cathode that emits electrons when light is incident on the rear surface of the fiber glass (incident fiber glass 2) that is an incident window attached to the front part of the vacuum tube 1 The photocathode 3 is provided, and the microchannel plate 4 having a large number of pores with a minute diameter of about 10 μm is installed behind the photocathode 3 with a required space. The micro channel plate 4
A fiber glass 5 is installed at a certain distance behind the fiber glass 5, and a screen made of a substance that emits light when electrons collide with the front surface of the fiber glass 5 is provided as a fluorescent screen 6. The light receiving surface 9 of the image pickup device 8 is connected to the rear surface of the fiber glass 5 via the fiber coupling 7 in the MCP type II.

In the above MCP type II, the light incident from the front surface of the incident fiber glass 2 and striking the photocathode 3 as shown by the arrow in FIG. 11 is converted into electrons according to the amount of light, and this is converted. The amount of light enters many holes of the microchannel plate 4. In the microchannel plate 4, one photoelectron collides with the wall of a small hole to generate an avalanche phenomenon in which a large number of secondary electrons are emitted, and this phenomenon is repeated to increase the number of electrons. A large number of photoelectrons emitted from the microchannel plate 4 collide with the phosphor screen 6 while being accelerated by the electric field, and emit enhanced light.

According to the MCP type II having the above action,
Due to the effect of increasing the number of electrons by the microchannel plate and accelerating the electrons by the electric field, the amount of incident light can be increased to about tens of thousands of times. Therefore, an MC that has a light enhancement function of 10,000 times or more when shooting under weak light
The use of P-type II is extremely effective in ultra-high-speed photography.

[0007]

As described above, the MCP type II is a very powerful optical device, but in principle the image is black and white. This is because the photocathode on the front surface of the microchannel plate emits electrons not only for visible light, but also for light of any frequency such as infrared rays, ultraviolet rays, and X-rays. This is because it will be information only.

Therefore, although it is not easy to achieve colorization using the MCP type II, in image processing, color information has a much larger amount of information than black-and-white information, which facilitates processing. For example, when tracking the trajectory of a rocket, in monochrome photography, only bright points corresponding to the trajectory of one object are selected from many bright points on the screen without being affected by other bright points. It's not easy to do. In that case, if color information is obtained by color imaging, the bright point to be traced can be easily identified.

For the above-mentioned reasons, it is desired to use the MCP type II in the ultra-high-speed image pickup apparatus to achieve colorization, and the following three methods are possible. 3-plate type Shutter or method of sequentially inserting filters of three colors on the light source side Method of mounting color filter array on the front of MCP type II

First, as is well known, the above-mentioned three-plate type video camera has a structure shown in FIG.
A dichroic prism 12 is provided behind the dichroic prism, and a layer for transmitting or reflecting only three primary colors (usually blue, red, and green) is provided in the prism 12 and the dichroic surface 13A, 13 is provided.
B is provided. Three image pickup devices 14A, 14B, 14C are arranged so as to face the three exit surfaces of the prism 12,
A color image is obtained by synchronizing the photographing by the three image pickup devices.

The MCP is used in the color image pickup device of the three-plate type.
When the type II is attached, as shown in FIG. 12, the MCP is provided between the image pickup devices 14A, 14B and 14C and the prism, respectively.
It is possible to adopt a method in which only the light of a predetermined color that has passed through the color filter and is subjected to color separation in advance via the type IIs 16A, 16B, and 16C is made to have an increased light amount by the MCP type II and is received by each image sensor. A red image, a green image, and a blue image respectively appear on the three image pickup devices, and a color image can be obtained by combining these.

However, when the above three-plate type is used,
There is a drawback that a positional shift is likely to occur when synthesizing three images. Further, since three image pickup devices and three MCP type II are used, respectively, there are disadvantages that the device becomes large and the cost becomes high.

Next, the shutter or the light source side 3
Consider how to add color filters in sequence. In principle, if the three continuous screens that are shot at short intervals of 1 / 60th of a second or less transfer the red, green, and blue light components, the movement (shift) between the three is sufficient. When it is very small, a color image can be obtained by combining the three images. As a method of changing the color of these three continuous images, on the light source side,
There are a method of outputting three colors of red, green, and blue one after another, and a method of sequentially applying filters of three colors when entering a camera.

Specifically, for example, three strobes are prepared, and these are used as light sources of three colors and are synchronized with the shuttering of the camera. Alternatively, a rotary type three-color filter may be attached in front of one white light source or in front of the camera and rotated in synchronization with the shuttering of the camera.

In the above method, there is a problem in the synchronization technique, and the outer shape becomes large. In many cases, it is impossible to change the color on the light source side or the shutter side depending on the type of photographing, and for example, the color on the light source side cannot be changed in photographing under a microscope or photographing in the daytime.

Next, the above color filter array
A method using (hereinafter, referred to as CFA) will be examined. In this case, as shown by the alternate long and short dash line in FIG. 11, the front surface of the incident fiber glass 2 is used as an imaging position by a lens, and the CFA 10 is attached to that portion, or the incident light is converted into parallel light and then the photoelectric surface is passed through the CFA. It is possible to adopt a method of making the light incident on the light source 3. This method is in principle the same as the case where the CFA is mounted on the front surface of the image pickup element adopted in a normal color video camera.

When the CFA is mounted on the front surface of the above-mentioned normal image pickup device, the CFA is closely attached to the front surface of the image pickup device,
The pitch of A is set to be the same as the pixel pitch of the image sensor. Therefore, for example, the amount of light received by the pixel below the red filter is proportional to the amount of red light above the red filter. Therefore, consider four adjacent pixels,
(Red, green, blue, green (R, G, B, G)), or (blue,
If you add filters such as yellow, green, and white (no filter)), you can know the brightness of each color component in the area covered by these four pixels.
Color images can be played back.

On the other hand, the method of attaching the CFA to the front surface of the MCP type II incident fiber glass 2 has the following problems. First, it is not easy to match the positional relationship between the CFA and the image sensor. When a CFA is mounted on the front surface of a normal image sensor, a color filter may be overlaid on each pixel and baked, so that the color filter should be installed without any error in the extension of the manufacturing method of a normal integrated circuit. I can.

However, when the CFA is attached to the front surface of the MCP type II incident fiberglass 2, as shown in FIG. 11, between the CFA 10 and the image sensor 8, the fiberglass 2 of the incident window, the photocathode 3, and the MCP 4 are provided. , The fiber glass 5 and the fiber coupling 7 are interposed, and a space exists between them. Therefore, since the distance between the CFA 10 and the image sensor 8 is several cm and a plurality of optical device elements are interposed therebetween, it is very difficult to match the pitch of the CFA and the image sensor. Even if it can be adjusted at the time of shipment from the factory, it is almost impossible to prevent the displacement caused by deformation due to thermal stress or mechanical stress.

Secondly, the light is dispersed by the time it enters the image pickup device from the CFA, and color mixing occurs in the image of the image pickup device. For example, as shown in FIG. 13, the red light that has passed through one red window of the CFA 10 diffuses when it enters the incident fiber glass 2 and also when it collides with the photocathode 3. The photocathode 3 becomes an electron corresponding to the light intensity, the photoelectron diffuses before entering the MCP 4, and the electron emitted from the MCP 4 also diffuses before hitting the fluorescent screen 6, and further, the fiber coupling from the fluorescent screen 6. When it is incident on 7, it also diffuses in the process of coupling the fiber coupling 7 and the image sensor 8. Therefore, for example, the CF of the light transmitted through the CFA 10
It is estimated that the total diffusion width D from A10 to the light receiving surface of the image sensor formed by the pixel 8a is about 40 μm.

On the other hand, as a matter of course, it is desirable that the pitch of the CFA 10 on the front surface of the photocathode 3 and the pitch of the pixels 8a of the last image pickup device 8 are about the same, and the total diffusion width is 4
Since it is about 0 μm, the pitch of the CFA 10 and the pixel pitch of the image sensor 8 must be at least 40 μm or more. However, the pixel pitch of an ordinary image sensor is 10
Since it is on the order of ˜20 μm, if an attempt is made to obtain a color image using MCP type II, color overlap at each pixel inevitably occurs.

On the other hand, it is also possible to use an ordinary small pixel size of the image pickup device, consider several to several tens of pixels as one set of pixels, and adjust the width to the mesh size of CFA.

Further, the pixel size of the image pickup device is the same as the above, and a normal small one is used, and at the stage of processing the photographed image, the pixel region is divided into each color later, and the pixel group of the red region is formed. It is also possible to consider a method in which the pixel groups in the blue area and the pixel groups in the green area are processed by eliminating color overlap. However, in the case of this method, it is necessary to set a color-coded section for each image, which has a drawback that the image processing becomes extremely complicated.

The method of mounting the color filter array on the front surface of the MCP type II as described above has not been realized due to problems such as alignment, color mixing, and resolution.

Although there are various problems in colorization using the MCP type II, as described above, color information has a much larger amount of information than black and white information in image processing, and various kinds of processing are easy. And for accuracy, MCP type I
There is a strong demand for the provision of a video camera capable of ultra-high-speed shooting that uses I-color in combination.

The present invention has been made to achieve the above-mentioned demand, and has been made for the purpose of achieving high-sensitivity colorization in a video photographing apparatus for ultra-high-speed photographing using MCP type II.

[0027]

A light intensifying device is arranged in front of an image pickup device, incident light is converted into photoelectrons by a photoelectric surface of the light intensifying device, and then converted into light having an increased light amount by a fluorescent surface. An imaging device that emits light, receives this light at each pixel of the image sensor, and stores and processes the output signal of each pixel in a processing unit,
A color filter array in which three primary colors are arranged in a grid pattern is arranged on the front surface of the fiber glass or the photoelectric surface of the entrance window of the light intensifying device, and the processing unit sets a plurality of adjacent pixels of the image pickup device as a set. The relationship between the output of each pixel in response to incident light and the intensity of the color component of the incident light at the center point of the pixel set is stored as a regression equation obtained from the condition of the least square error. A color video recording apparatus configured to estimate the intensity of the color component at the center point of each pixel group from the output signal of each pixel based on the regression equation.

The regression coefficient of the above regression equation is preferably stratified according to the intensity of incident light.

Specifically, the light intensifying device comprises a micro channel plate type image intensifier, and the pitch of the color filter array is set to the fiber glass front surface or the photoelectric surface of the entrance window of the micro channel plate type image intensifier. 1 of the total diffusion width until the incident light incident on one pitch of the color filter array placed on the front surface of the
/ 2 times or more and 2 times or less.

More specifically, the color filter array preferably includes red, green, blue or yellow, green and blue filters.

[0031]

Since the color video recording apparatus according to the present invention has the above-mentioned structure, when the incident light reaches the light receiving surface of the image pickup device from the color filter array through the light intensifying device, each pixel becomes an incident light. The response output is sent to the processing unit as an image signal of the image sensor. In the processing unit, the image signal is stored, and based on the regression equation between the output recorded in advance corresponding to the incident light of each pixel and each color component of the incident light at the center of each pixel group, It is possible to reproduce the color image by estimating the intensity of the color component of the incident light in each image set.

When the regression coefficient is stratified by the intensity of the incident light, the intensity of each color component of the incident light can be estimated from the output of the pixel with high accuracy even if the intensity of the incident light changes.

[0033]

Next, the present invention will be described in detail with reference to the embodiments shown in the drawings. As shown in FIG. 1, the overall structure of the color video photographing apparatus according to the present invention is the same as the conventional one, and a video camera body 20 is provided with an image pickup unit 22 for converting the light imaged by the lens 21 into an electric signal. The image pickup device 30 provided in the image pickup unit 22 (shown in FIG. 2)
The image signal is output to the signal processing unit 24 provided in the processing unit 23 which is separate from the video camera main body 20, and is sent from the signal processing unit 24 to the memory unit 25. An image forming computer 26, an image forming device 27, a display 28, and an external recording device 29 are connected to the memory section 25.

As shown in FIG. 2, the structure of the image pickup unit 22 is substantially the same as the structure shown in FIG. 11. Therefore, the same members are designated by the same reference numerals and the description thereof will be omitted. That is, in the image pickup unit 22, a microchannel plate type image intensifier (hereinafter referred to as MCP type II) 31 is installed in front of the light receiving surface 30a of the image pickup element 30, and the incident fiber glass 2 of the MCP type II 31 is installed. A color filter array 33 (hereinafter abbreviated as CFA) on the front surface.
Therefore, the amount of light is increased by the MCP type II 31, and the color is achieved by the device of the CFA 33.

As in the case of FIG. 11 described above, the MCP type II 31 converts incident light into electrons at the photocathode 32 on the rear surface of the incident fiberglass 2 attached to the front part of the vacuum tube 1, and converts the converted photoelectrons into microelectrons. Doubled through the channel plate 4, the doubled electrons are converted into light by the fluorescent surface 6 on the front surface of the fiber glass 5, and emitted to the light receiving surface 30a of the image pickup device 30 through the fiber coupling 7.

As shown in FIG. 3, the CFA 33 of this embodiment is an RGB type CFA in which color filters 35 of three primary colors of 40 μm × 40 μm of red, green and blue are sequentially arranged in a grid pattern mask pattern. . In this embodiment, CFA3
The pitch P1 of 3 is set to the same degree as the total diffusion width D of the incident light from the photocathode 32 to the light receiving surface 30a after passing through the CFA, that is, 40 μm.

This is because when the pitch P1 of the CFA 33 is set to be sufficiently larger than the total diffusion width D and the pixel pitch P2 of the image pickup device 30 described later is set to be sufficiently smaller than the pitch P1 of the CFA 33, as shown in FIG. As shown in FIG. 5, a certain group of pixels G (indicated by hatching in the figure) responds to a single color component, and the remaining pixels G respond with a mixture of color components. By focusing only on the group of pixels that respond only to this single color component, the intensity of that color component can be directly known without using the regression equation. Therefore, after joining the MCP type II with CFA to the image pickup device and then sequentially obtaining red light, green light, and blue light, it is possible to investigate which device responds to which light. However, if a pixel group that responds only to each single color component is investigated, a color image can be reproduced from this. However, in this case, the pitch of the CFA has to be at least twice as large as the total diffusion width D or more, so that there is a drawback that the spatial resolution of the screen becomes low.

On the other hand, in order to maintain the spatial resolution, C
It suffices to make the pitch P1 of FA smaller, but if the pitch P1 is made significantly smaller than the total diffusion width D, the pitch P1 of FIG.
As shown in, the colors are completely mixed on the light receiving surface of the image sensor, and the colors cannot be separated. Therefore, in the present embodiment, the pitch P1 of the CFA 33 is set to be approximately the same as the total diffusion width D as described above. The pitch P1 of the CFA 33 is not limited to the above-mentioned 40 μm, and a similar function can be expected if the pitch P1 is slightly smaller than the total diffusion width D, but at least half the total diffusion width D or more. On the other hand, as described above, the desired spatial resolution cannot be obtained if the total diffusion width D is set to twice or more. Therefore, the total diffusion width D is set to 1/2 times or more and twice or less. Is preferred.

As shown in FIG. 7, the image pickup device 30 of this embodiment has pixels G arranged in a grid pattern, the number of pixels is 256 rows × 256 examples, the pixel pitch P2 is 40 μm, and the pixel pitch is Color filter array 33 for PG
Is set to be equal to the pitch P1.

As described above, in the present embodiment, the pitch P1 of the color filter array 33 is set within the range of 1/2 times or more and 2 times or less of the total diffusion width D of the incident light in the MCP type II31. In addition, since the pixel pitch P2 of the image sensor 30 is set to be approximately the same as the pitch P1 of the color filter array 33, the color filter array 3
The light once decomposed in 3 is the image sensor 3 as shown in FIG.
There is some mixing on the light receiving surface 30a of 0.

However, the light once decomposed by the color filter array 33 is not completely mixed on the light receiving surface 30a, so that the pixel or green which slightly reacts to the red color component of the incident light, or the green. Alternatively, a pixel that slightly strongly reacts to the blue color component occurs. However, for example, even a pixel that is slightly strongly responsive to the red color component is responsive to the green and blue color components to some extent.

In the present embodiment, for all the pixels G of the image pickup device 30, four pixels adjacent in the vertical and horizontal directions are set as one set, and for each pixel set consisting of these four pixels,
The output in which each pixel responds to the incident light and the incident light red,
A regression equation is created between the intensities of the three color components of green and blue. Using this equation, from the difference in the intensities of incidence on the above four pixels, the incidence at the center point of the four pixels is made. The intensities of the three color components of light, red, green, and blue, are estimated.

The regression coefficient of the above regression equation is obtained by calibration. That is, each of red, green, and blue monochromatic light is made to enter the imaging unit 20 a plurality of times, and at that time,
The current intensity generated in response to each pixel in the set is examined, and the regression coefficient is obtained from the condition of the least square error.

Hereinafter, regarding a set of four pixels G1, G2, G3 and G4 which are adjacent to each other in the vertical and horizontal directions shown in FIG.
The above regression equation and regression coefficient will be described in detail. Each pixel G1, G2, G3, G4 has a current intensity generated in response to incident light having a certain color component as X ₁ , X ₂ , X ₃ , X ₄ ,
The intensity of the red, green, and blue color components of the incident light at the center point P of the four pixels G1, G2, G3, and G4 with respect to the incident light is represented by Y ₁ , Y ₂ , and Y ₃ below, respectively. Build up.

[0045]

[Equation 1]

Here, β mn (m = 1 to 3, n = 1 to 4) is a regression coefficient, and α ₁ , α ₂ and α ₃ are constants. The value of the regression coefficient βmn in this equation (1) is obtained by calibration. As the calibration method, as described above, the red, green, and blue monochromatic lights whose intensities are known are respectively incident K times, and at that time, each pixel G1, G2, G3, G4 responds to the incident light. The current intensities X ₁ , X ₂ , X ₃ , X ₄ are obtained, and the regression coefficient βmn is calculated from these values using the condition of the least square error. Here, in the case of the monochromatic light, the color components other than the color component are 0. For example, in the case of red monochromatic light having an intensity of 1, Y ₁ = 1 and Y ₂ , Y ₃ = 0. From the current intensities X ₁ , X ₂ , X ₃ , X _{4 of} each pixel and the intensities Y ₁ , Y ₂ , Y ₃ of the red, green, and blue color components at the center point P, the average values at the time of calibration, respectively. Pull

[0047]

[Equation 2]

Let us say. Note that Y _ma is the average value of the color components, X
_na indicates the average value of the current intensity of each pixel. From this equation (2), the above equation (1) becomes

[0049]

[Equation 3]

It becomes Expanding equation (3) above, m =
1, that is, for the red color component,

[0051]

[Equation 4]

It becomes Here, ε ₁ k is an error. Therefore, from equation (4)

[0053]

[Equation 5]

In addition,

[0055]

[Equation 6]

Here, from the condition of the least square error,

[0057]

[Equation 7]

From equation (7),

[0059]

[Equation 8]

Putting,

[0061]

[Equation 9]

Therefore,

[0063]

[Equation 10]

It becomes The regression coefficients β ₁₁ , β ₁₂ , β ₁₃ , and β ₁₄ are determined by substituting the intensity of the red monochromatic light used for calibration and the response of each pixel into the above equations (8) and (10). For the green color component and the blue color component, the regression coefficients β ₂₁ , β ₂₂ , and β are set in the same manner as the red color component described above.
₂₃ , β ₂₄ and the regression coefficients β ₃₁ , β ₃₂ , β ₃₃ , β ₃₄ can be determined.

In this embodiment, the light receiving surface 30 of the image pickup device 30 is used.
Regression coefficients as described above are prepared for all combinations of four pixels G forming a that are adjacent to each other in the vertical and horizontal directions, and regression coefficients are obtained in advance by calibration, and they are processed by a processing device. 23 in the image processing device 27.

Furthermore, the input / output relationship between the intensity of the incident light and the current intensity of the pixel response is usually non-linear, and the value of the regression coefficient βmn is different when the average intensity of the incident light is strong and weak. Therefore, in this embodiment, the average intensity of the incident light is stratified into three levels of high, medium, and low, and the value of βmn is obtained for each level.

As described above, since the relationship between the intensity of the color component of the incident light and the output of each pixel is expressed by the regression equation, if the CFA 33 and the image pickup device 30 are misaligned, the calibration is performed again. Should be done.

Since the color video photographing apparatus of this embodiment has the above-mentioned structure, the CFA 33 is used at the time of photographing.
The current intensity of the response of each pixel G to the incident light incident on the image pickup surface 30a of the image pickup device 30 through the photocathode 32, the microchannel plate 4, and the phosphor screen 6 is stored in the memory unit 25 via the signal processing unit 24. Accumulated. During image reproduction, the processing unit 23 causes the four pixels G1, G2,
The average intensity of the pixel responses X ₁ , X ₂ , X ₃ , X ₄ is calculated for each of G3 and G4, and the regression coefficient βmn corresponding to the three levels of strength, medium strength, and weakness is applied to the above equation (3), and incident light is input. It is possible to reproduce the color image by estimating the color components of.

As described above, in the present embodiment, since the regression coefficient βmn is obtained in advance by the calibration as described above and the regression coefficient βmn is recorded in the processing device 23, the incidence of an arbitrary color component is made. Each color component at the center point P can be estimated from the equation (3) for light.

FIGS. 9A to 9D show the second embodiment of the present invention. In the second embodiment, with respect to all the pixels of the image sensor 30, three pixels G1 ′, G2 ′, G that are adjacent in the vertical and horizontal directions are used.
3 ′ as one set, the intensity of the red, green, and blue color components of incident light Y ₁ , Y ₂ , Y ₃ and the pixels corresponding thereto are calculated for these three sets of pixels in the same manner as in the first embodiment. Response X ₁ ,
A regression equation between X ₂ and X ₃ is obtained, and the red and green incident lights at the center point P ′ of the three pixels are calculated from the difference in the incident intensity on each of the three pixels using this regression equation. The intensities of the three color components of blue are estimated.

As shown in FIG. 9, one central point P '
The method of selecting three pixels that are adjacent to each other in the vertical and horizontal directions is 4
There is communication. For these three pixels G1 ', G2', G3 ',

[0072]

[Equation 11]

And

[0074]

[Equation 12]

It becomes Therefore, if the coefficient βmn 'is determined, the equation
From (11), the coefficient βmn 'is added to the current intensity of each pixel G ₁ ', G ₂ ', G ₃ ' responding to the incident light having an arbitrary color component.
If it is written, the intensity of each color component at the center point P ′ can be calculated.

The calibration for obtaining the coefficient βmn 'is performed as follows. That is, red monochromatic light having a known intensity is made incident on the image pickup element and each pixel G ₁ ', G ₂ ', G
₃ 'current intensity of _{X 1', X 2 ',} X 3' examined. At this time Y ₁
Since Y ₂ and Y ₃ = 0 are known, the coefficients β ₁₁ , β ₁₂ and β ₁₃ can be obtained by dividing X ₁ ′, X ₂ ′ and X ₃ ′ by Y ₁ ′. Similarly, when green and blue monochromatic lights of known intensities are incident and the intensity current of each pixel at that time is divided by the intensity of the monochromatic color, the coefficients β ₂₁ , β ₂₂ , β ₂₃ and the coefficient β are respectively obtained. ₃₁ , β ₃₂ , β ₃₃ can be obtained.

As described above, there are four ways of selecting a set of three pixels with respect to one central point P ', as shown in FIGS. 6A to 6D. In this embodiment, these four ways are selected. Coefficient matrix βij
Is obtained, a set of three pixels that maximizes the value of det {βij} is adopted, and this is stored in the processing unit 23. Furthermore, the second
Also in the embodiment, similarly to the above-described first embodiment, the intensity of incident light is stratified into three stages of strong, medium and weak, and different βij values are used for each stage.

The other structure and operation of the second embodiment are the same as those of the above-mentioned first embodiment, and the description thereof will be omitted.

The present invention is not limited to the above embodiment, but various modifications can be made. For example, in the above-described first and second embodiments, the pitch of CFA is set to be about the same as the diffusion width, and the pixel pitch of the image sensor is set to be the same as the pitch of CFA. However, the pixel pitch of the color filter array is set. It may be set smaller than the pitch. When the pixel pitch is set in this way, the color resolution can be further improved.

In both the first and second embodiments, the stratification of three levels of strong, medium, and weak is performed corresponding to the intensity of the incident light, but the number of stages is further increased according to the shooting conditions to perform fine stratification. The stratification may be performed, or conversely, the stratification may not be performed.

When the color components are estimated with a set of three pixels as in the second embodiment, an image pickup device in which the pixels are arranged in a staggered arrangement as shown in FIG. 10 may be used. in this case,
There are one way to select three pixels for one center point P ′.

The color filter array is not limited to the RGB type as described above, and for example, a YGB type color filter array having yellow, green and blue filters may be used.

[0083]

As is apparent from the above description, in the color video recording apparatus according to the present invention, the color filter array is mounted on the photocathode of the MCP type image intensifier which is a light intensifying device suitable for ultra high speed. , The relationship between the received light intensities of several adjacent pixels of the pixels of the image pickup device and the intensities of the color components of the three primary colors at their center points is stored in advance in the processing means, and the above-mentioned several of them are output from the output signal of each pixel. Since the intensity of the color component of the incident light in the set of pixels is estimated, it is possible to perform ultra-high-speed color photography without increasing the size of the device, and it is not necessary to align the color filter with the image sensor.

Further, the color video apparatus according to the present invention is
With the above-mentioned configuration, there is no need to illuminate the subject, and it is possible to achieve ultra-high-speed color imaging even under a microscope, and also for ultra-high-speed color imaging of large devices and moving objects. It is suitable, and has various advantages such as a large amount of information in measurement photography and the accuracy of measurement processing can be improved.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing an entire color video shooting apparatus according to a first embodiment.

FIG. 2 is a side view of essential parts showing a state in which a color filter is attached to the front surface of an MCP type II photocathode.

FIG. 3 is a partial schematic view showing a mask pattern of a color filter array.

FIG. 4 is a schematic diagram showing a mixed state of light beams of respective colors when the pitch of the color filter array is set to be sufficiently larger than the total diffusion width.

FIG. 5 is a schematic view showing a mixed state of light beams of respective colors when the pitch of the color filter array is set sufficiently larger than the total diffusion width.

FIG. 6 is a schematic diagram showing a mixed state of light beams of respective colors when the pitch of the color filter array is set to be significantly smaller than the total diffusion width.

FIG. 7 is a partial schematic diagram showing a pixel array of an image sensor.

FIG. 8 is a schematic diagram showing a mixed state of light rays of respective colors when the pitch of the color filter array is set within the range of this embodiment.

FIG. 9 is a partial schematic diagram showing how to select three pixels in the second embodiment.

FIG. 10 is a partial schematic diagram showing an array of pixels in a modification of the present invention.

FIG. 11 is a view similar to FIG. 2 for explaining a conventional problem.

FIG. 12 is a diagram for explaining a problem when the conventional No. 3 equation is used.

FIG. 13 is a view for explaining a problem when a color filter is attached to the front surface of the photocathode.

[Explanation of symbols]

 2,5 Fiber glass 4 Micro channel plate 23 Processing unit 30 Imaging device 31 MCP type II 32 Photoelectric surface 35 Color filter G pixel

Claims (3)

[Claims] 1. A light intensifying device is arranged in front of an image pickup device,
After converting the incident light into photoelectrons on the photoelectric surface of the light intensifying device, the light is converted into light with an increased amount of light on the fluorescent surface and emitted, and this light is received by each pixel of the image sensor, and the output signal of each pixel is output. An image pickup apparatus for storing and processing in a processing unit, wherein a color filter array in which three primary colors are arranged in a grid pattern is arranged on the front surface of the fiber glass or the photoelectric surface of the entrance window of the light intensifying device. , A group of a plurality of adjacent pixels of the image pickup device, the relationship between the output of each pixel in response to incident light and the intensity of the color component of the incident light at the center point of the pixel set Color video recording is stored as a regression equation obtained from the condition of least square error, and the intensity of the color component at the center point of each pixel group is estimated from the output signal of each pixel based on the regression equation. apparatus. 2. The color video recording apparatus according to claim 1, wherein the regression coefficient of the regression equation is stratified by the intensity of incident light. 3. The light intensifying device comprises a microchannel plate type image intensifier, and the pitch of the color filter array is arranged on the front surface of the fiber glass or the photocathode of the entrance window of the microchannel plate type image intensifier. 1/2 of the total diffusion width until the incident light incident on one pitch of the arranged color filter array reaches the light receiving surface of the image sensor
3. The color video recording apparatus according to claim 1, wherein the color video recording apparatus is set to be more than twice and less than twice.