JP6842691B2 - Color image sensor, color image imager, and color image imager - Google Patents

Description

The present invention relates to an imaging method for efficiently and high-quality imaging of a color image using a single image sensor and an interpolation method of the image captured by the imaging, and a particularly characteristic color filter array and corresponding interpolation. The present invention relates to imaging and signal processing for obtaining a high-quality color image signal suitable for information compression by processing.

<Color image system>
Not limited to moving images, color image signals are formed by the three primary colors of R (Red, red), G (Green, green), and B (Blue, blue), and imaging and display are performed in this form. On the other hand, as a form for transmission and recording, a component color signal formed by luminance and two color differences is used. These are obtained by a predetermined conversion from the three primary colors, and the color difference is often subsampled. Specifically, non-subsampled 4: 4: 4, horizontally sub-sampled 4: 2: 2, both horizontal and vertical sub-sampled 4: There is 2: 0. Each moving image color format is defined in detail in the standard and is widely used in television broadcasting and the like.

<Color imaging>
In color image imaging, there are a three-plate method in which R, G, and B image sensors are provided, and a single-plate method in which R, G, and B are distributed to each cell (pixel) by a single image sensor. The three-plate system is becoming a special existence in recent years, and it does not exist in still cameras, and almost all of the 4K image cameras with 3860 x 2160 pixels are single-panel systems even for moving images.

In the case of a single plate, the R, G, and B color filters are arranged in the light receiving cells (pixels), but the three types of filters should be as equal and uniform as possible. However, since it is not easy to divide it into three equal parts in two dimensions, what is called a Bayer array as shown in FIG. 1 (a) is common, with half for G and four minutes for R and B, respectively. Allocate 1. The reason why the ratio of G is larger than the others is that the ratio of G to the luminance component is the highest and the resolution of the luminance signal is not deteriorated.

Different from this, there is one using complementary colors. Cyan (complementary color of R) C = G + B, magenta (complementary color of G) M = R + B, and yellow (complementary color of B) Y = R + G are used instead of R, G, and B. As the complementary color component, the R component is (Y + MC) / 2 and the white component is (Y + C + M) / 2, so that it is multiplied by 1/2, so that the dark current noise generated in the imaging is reduced. In CCD sensors, it was mainly used for video cameras, but the spectral characteristics of complementary color filters are not always ideal, so color reproducibility is poor. In FIG. 1B, one color is left as G, and the color reproducibility is improved. However, since there are four colors, it is not necessarily advantageous in terms of resolution.

Further, in addition to R, G, and B, those using white (W) have also been proposed. The arrangement is shown in FIG. 1 (c). R and B are the same as the Bayer sequence, and half of G is replaced with W. The W filter is a filter that transmits all visible light with only infrared rays and ultraviolet rays removed, and since light energy can be used effectively, noise is reduced like a complementary color filter. It has been attracting attention in recent years because of its good color reproducibility.

<Color plane interpolation processing (demosaiking)>
In a single-panel color imaging device, each color of the image (RAW image) obtained from the image sensor is in a state where cells (pixels) are thinned out, so a color plane in which all the pixels are aligned by interpolating the missing pixels. (For example, Patent Document 1). This missing color complementary process is called demosaiking. The basic processing is performed by a two-dimensional interpolation filter corresponding to the thinning structure of each color plane. In the case of the Bayer array, the interpolation filter is different because the thinning structure is different between G and R (B). In this case, the obtained resolution depends on the thinning structure, and is two-dimensionally half for G and one quarter for R and B.

Some improved techniques adaptively switch interpolation according to the local shape of the image. If the adaptive process works well, the resolution will improve, but it may result in erroneous interpolation. There is also a method using the correlation between R, G, and B colors, which is effective for an image having a high correlation between colors, that is, a low saturation, but there is a risk of erroneous interpolation in a high saturation portion.

Japanese Unexamined Patent Publication No. 2003-179941

The color image imaging method uses color filters of the three primary colors R (red), G (green), and B (blue). In the image sensor of the single-plate color image imaging device, these are arranged for each cell (pixel). In this case, since the number of pixels of each color is half of the whole even with the largest number of G, the spatial resolution is significantly inferior to that of the single color (monochrome) or three-plate sensor. Each color plane is obtained by demosaiking, but the luminance signal converted from it has poor frequency characteristics, and there is unevenness due to demosiking. When the luminance component is deteriorated in this way, a prediction residual is likely to occur in the motion compensation image-to-image prediction at the time of information compression, and the amount of generated information becomes unnecessarily large.

The present invention has been made in view of the above problems in the prior art, and an object of the present invention is to provide a configuration for obtaining a high-quality color image signal suitable for information compression.

As a result of diligent studies on a configuration for obtaining a high-quality color image signal suitable for information compression, the present inventor came up with the following configuration and arrived at the present invention.

That is, according to the present invention, the color filter is a single-plate color image sensor in which a color filter that transmits light in the entire visible light is arranged on each light receiving cell of a single image sensor. , A first light color filter that transmits more of the first primary color component, a second light color filter that transmits more of the second primary color component, and a third light color filter that transmits more of the third primary color component. A color image sensor including the above is provided.

Further, according to the present invention, the color filter is a single-plate color image sensor in which a color filter that transmits light in the entire visible light is arranged on each light receiving cell of a single image sensor. , A first light color filter that transmits the first primary color component less, a second light color filter that transmits the second primary color component less, and a third light color filter that transmits the third primary color component less. A color image sensor including the above is provided.

Further, according to the present invention, it is a single plate type color image imaging apparatus including the color image imaging element, and corresponds to the first light color image corresponding to the first light color filter and the second light color filter. A means for acquiring a second light-colored image and a third light-colored image corresponding to the third light-colored filter, a means for converting the three types of the light-colored image into a white component and a color difference component by a matrix calculation, and the white component. And a color image imaging apparatus including means for converting the color difference component into three types of primary color images by matrix calculation are provided.

Further, according to the present invention, it is a single plate type color image imaging device provided with the color image imaging element, and a low-pass filter is used from a RAW image corresponding to all cells of the color image imaging element to a low frequency band. A means for separating existing white components, a means for obtaining a color difference component from the residuals of the white component and the RAW image, and a means for converting the white component and the color difference component into three types of primary color images by matrix calculation. A color image imaging apparatus including, is provided.

As described above, the present invention provides a configuration for obtaining a high quality color image signal suitable for information compression.

FIG. 1 is a diagram showing a color filter arrangement of a conventional single-plate color imaging device. FIG. 2 is a diagram showing a light color filter arrangement of the first embodiment. FIG. 3 is a diagram showing a functional block of the single plate color imaging apparatus of the first embodiment. FIG. 4 is a diagram showing a partial filter structure according to the first embodiment. FIG. 5 is a diagram showing a light color filter arrangement of the second embodiment. FIG. 6 is a diagram showing a functional block of the single-panel color imaging apparatus of the third embodiment. FIG. 7 is a diagram showing a state of the RAW image space frequency spectrum. FIG. 8 is a diagram showing a spectrum detection band and separation filter characteristics.

Hereinafter, the present invention will be described with reference to the embodiments shown in the drawings, but the present invention is not limited to the embodiments shown in the drawings. In each of the figures referred to below, the same reference numerals are used for common elements, and the description thereof will be omitted as appropriate.

<First Embodiment>
The single-panel color image capturing apparatus according to the first embodiment of the present invention will be described. FIG. 2 shows a color filter included in the single-panel color image capturing apparatus of the present embodiment and its arrangement. The feature of this embodiment is a color filter, and other functions (optical system, cell structure other than the color filter, data transfer method, etc.) in the image sensor are common to those of a normal single-panel color image image pickup device.

In the single-plate color image imaging apparatus of the present embodiment, instead of the R, G, and B color filters, a light color filter that transmits light in the entire visible light range to some extent is used. Specific colors are P (Pink), L (Light Green), and S (Sky Blue), which are light colors obtained by adding a white component to R, G, and B. The light color filter P transmits the first primary color component (R component) more than the remaining primary color components (G component, B component), and the light color filter L transmits the second primary color component (G component). The remaining primary color components (R component, B component) are transmitted more, and the light color filter S transmits the third primary color component (B component) more than the remaining primary color components (R component, G component). Let me. The conversion matrix from R, G, B to P, L, S is the degree of light transmission of each component of R, G, B in the light color filter, but in the light color filter, two other than the primary colors that transmit the most. It is expressed by the following transformation matrix equation, where w is the transmittance of one primary color component.

Here, assuming that half of all transparent colors of other colors, that is, w = 0.5, the following equation is obtained.

Compared with the conventional R, G, B as it is (w = 0), the R, G, B components used are doubled. If the characteristics of this embodiment are to be strengthened, the W component is increased, for example, w = 0.75. On the contrary, when the W component is reduced and, for example, w = 0.25, the conventional R, G, and B are approached.

<Color filter arrangement>
FIG. 2 shows the color filter arrangement of the present embodiment. In the PLSL arrangement shown in FIG. 2A, the Bayer arrangements R, G, and B are converted into P, L, and S by adding white components as they are, L is 1/2, and P and S are 1 respectively. / 4 exists. This arrangement is appropriate for processing where the value of w is small and relatively close to the conventional type.

In the PLPS arrangement shown in FIG. 2B, P and L are interchanged, and P is 1/2 and L is 1/4. In R, G, and B, the proportion of G contained in the brightness was increased to maintain the resolution of the brightness, but in P, L, and S, white is present in all pixels, so it is necessary to increase G. This is because it fades.

In the PLSW arrangement shown in FIG. 2 (c), one L is white W with respect to PLSL, and the conventional RGB W shown in FIG. 1 (c) has W as it is and R, G, and B are P. It can be said that it is made into, L, S. It can be seen that the white W is L and w = 1.0. As mentioned above, the need to increase L has diminished, so by setting it to W, the average value of w will be increased.

<Inverse conversion and characteristics>
As a color image, R, G, B signals are finally required, but the conversion from P, L, S to R, G, B is the same as the conversion from the luminance color difference signal to the RGB signal. It can be realized by matrix calculation. This is the inverse transformation of the transformation from R, G, B to P, L, S, and the one corresponding to equation (1) is expressed by the following equation.

When w = 0.5, that is, the inverse transformation of Eq. (2) is expressed by the following equation.

The R, G, B image signals thus obtained are basically the same as the conventional R, G, B image signals, but are characterized in the influence of noise generated in imaging on the image quality. Specifically, since the W (white) component close to the luminance component is increased, the noise contained in the luminance component is reduced. To show a specific example when w = 0.5, the white component W is obtained by halving the sum of P, L, and S by the following equation.

This means that if the same noise as in the conventional R, G, B imaging occurs in the cells P, L, S, the noise in W is halved. Assuming that the conversion gain from P, L, S to W is G _W , the relationship with w is as follows. The larger w is, the higher the noise reduction effect is.

On the other hand, the color difference component when w = 0.5 is as follows, taking R-G as an example, and the noise is doubled together with the signal component.

If this is the conversion gain G _C , the relationship with w is given by the following equation, and it can be seen that a value where w is close to 1.0 is not realistic.

Since the luminosity factor for the color difference component error is considerably lower than the luminosity factor for the luminance component error, the image quality at w <0.5 matches the visual characteristics as it is. Moreover, since the color sensitivity to a high spatial frequency is significantly low with respect to the color difference component, w> 0.5 is also possible on the premise that noise reduction processing is applied to the color difference component.

<Development processing>
Next, the development process from obtaining the R, G, B image data from the RAW image data will be described. FIG. 3 shows a functional block of the single plate color image capturing apparatus of this embodiment. The photoelectric conversion value (RAW image data) of each cell obtained by the color image image sensor in which the above-mentioned light color filter of each light receiving cell of the image sensor 1 is arranged is given to the demosaiking 2. The demosaiking 2 interpolates the missing pixels of each color plane such as P, L, and S in the same manner as the general processing for R, G, and B, and outputs each color plane in which all the pixels are aligned. In this embodiment, the interpolation process is performed by a fixed linear filter. The interpolated signal may be immediately converted into R, G, and B and output, but the effect of the present embodiment can be further enhanced by performing processing for each component.

Each interpolated color plane signal such as P, L, S is given to the light color matrix 3. The light color matrix 3 converts the white component W, the color difference component R-W, and the color difference component B-W into three components. This process is to perform the conversion to R, G, B and the W formation and subtraction processes at the same time, but it can be realized only by changing the coefficients of the conversion to R, G, B. W, R-W, B-W may be Y, R-Y, B-Y, where white W is the brightness Y.

The converted W signal is given to the W corrector 4, the R-W signal is given to the R-W corrector 5, and the B-W signal is given to the B-W corrector 6. Each corrector performs correction suitable for each signal, but the R-W corrector 5 and the B-W corrector 6 also use the W'signal corrected by the W corrector 4. The W corrector 4 performs sharpening processing in a form of compensating for the deterioration of the frequency in the demosaiking. In the R-W corrector 5 and the B-W corrector 6, a sharpening process similar to W is performed at a conversion point where W, which is an edge, is large, and a flattening process is performed at other parts. The corrected W signal, RW signal, and BW signal are given to the primary color matrix 7. The primary color matrix 7 converts to R, G, B, which can be realized by R = (R-W) + W, G = W-R-B, B = (B-W) + W. The obtained R, G, B signals are output from the R signal output 8, the G signal output 9, and the B signal output 10.

<Light color filter>
In order to realize this embodiment, the above-mentioned light color filter is required. A normal color filter is made by impregnating a transparent film material with a dye, and the color of the filter is determined by the spectral characteristics of transmission. An easy way to make the current R, G, B color filters lighter (P, L, S) is to reduce the amount of dye and the amount of blocked light. .. However, normally, there is no problem if the wavelength (color) to be suppressed is not completely uniform, as long as it is sufficiently suppressed, but if about half of the wavelength (color) is transmitted, the difference in transmittance becomes a problem. For example, when R is set to P, the attenuation tends to be larger in B having a far wavelength than in G having a near wavelength. It is necessary to take measures against such deviation from the ideal filter characteristics. Here, three types of methods are shown.

The first method is to change the matrix. If the variation in transmittance is not remarkable, it is allowed, and the conversion matrix itself is the characteristic. In this case, there will be a slightly different W as shown in the following equation.

Since the inverse transformation is obtained for this matrix, R, G, and B can be appropriately obtained by the inverse transformation. However, since each W changes depending on the sensor, a unique W is required for the inverse conversion.

The second method is a method of obtaining a constant W as much as possible by adjusting the dye. When the transmittance of G is higher than that of B as in the previous example, the transmittance cannot be increased. Therefore, in order to reduce the transmittance of G, a slight amount of dye of a color (magenta) that suppresses only G is added.

The third method is a method in which a white filter (that is, no filter) is provided in a certain spatial area. This situation is shown in FIG. The simplest is, as shown in FIG. 4A, left-right division, in which R, G, B filters are provided only on the left and right left halves in one cell (pixel), and the other half is a white filter. In this case, the change in light in one cell is suppressed by the optical filter, but it is not completely uniform, so that a slight imbalance occurs. It should be noted that it may be divided vertically or diagonally instead of left and right, but the same problem remains.

As a method of balancing at least vertically and horizontally, as shown in FIG. 4 (b), a filter smaller than the cell is placed only in the central portion, and as shown in FIG. 4 (c), a hole is formed in the central portion. There is also a way to open it without a filter. In either case, since the light is focused in the center of each cell, the area of the central portion is smaller than that of the peripheral portion excluding the center. In addition, without a filter, it is desirable to provide a white filter because it also receives light outside the visible light, but it is not necessary to provide only the part without the filter, and a white filter may be provided so as to cover the entire cell. ..

<Second Embodiment>
The single-panel color image capturing apparatus of the second embodiment of the present invention will be described. FIG. 5 shows the color filter of the present embodiment and its arrangement. The difference between this embodiment and the first embodiment described above is the color filter, and the others are common. In the present embodiment, instead of the primary color light color filter used in the first embodiment, it is a filter that transmits light in the entire visible light range, and has a transmittance of only one primary color component out of three types of primary color components. Use a complementary light color filter with the value reduced. The specific colors are bright cyan with R subtracted from white, A (Aqua), bright magenta with V (Violet) with G reduced, and bright yellow with E (Lemon) with B subtracted. To do. The complementary light color filter A transmits the first primary color component (R component) less than the remaining primary color components (G component and B component), and the complementary light color filter V transmits the second primary color component (G component, B component). The G component) is transmitted less than the remaining primary color components (R component, B component), and the complementary light color filter E allows the third primary color component (B component) to pass through the remaining primary color component (R component, G component). Permeate less than component). In Lemon, since Light Green was set to L in the first embodiment, it is set to E as the second character to distinguish it. The conversion matrix from R, G, B to A, V, E is similar to the primary color light color filter used in the first embodiment, and has a degree of light transmission of each component of R, G, B, but is a complementary color system. In the light color filter, the rate at which the transmission is reduced (prevented) with respect to the transmission of each primary color component is represented by the following conversion matrix as c (color).

When c = 0.5, the following equation is obtained.

If the rate of subtraction is increased and c = 2/3, the following equation is obtained.

If c = 1, it becomes a pure complementary color filter.

FIG. 5 shows the color filter arrangement of this embodiment. In the AVEV arrangement shown in FIG. 5A, R, G, and B of the Bayer arrangement are directly A, V, and E, V is 1/2, and A and E are 1/4 each. In the AWEW arrangement shown in FIG. 5 (b), V is changed to white W, and the conversion formula is changed to the following formula.

This is white and white with two types of color differences added, and has a shape similar to the luminance color difference signals Y, Cb, and Cr.

In the conversion from A, V, E to R, G, B, the general formula of the inverse conversion to the formula (10) is represented by the conversion matrix of the following formula.

The inverse transformation corresponding to equation (11) with c = 0.5 is as follows.

The inverse transformation corresponding to equation (12) with c = 2/3 is as follows.

The conversion gain from A, V, E to W G _W and c are shown by the following equation, and the smaller c is, the higher the noise reduction effect is.

On the other hand, _{the relationship between the conversion gains G C} and c of the color difference component is as follows, and it can be seen that a value where c is close to 0 is not realistic.

Comparing G _W and G _C with the first embodiment, w = 0.5 and G _W = 1/2, G _C = 2 in the first embodiment, whereas c = in the second embodiment. At 0.5, G _W = 1 / 2.5 and G _C = 2, and G _{W becomes smaller for the same G C,} which is advantageous in terms of noise reduction effect. This tends to be the same as the characteristics of general primary color filters and complementary color filters. In this respect, complementary color filters are superior, but in terms of spectral characteristics, primary color filters are superior, so superiority or inferiority cannot be simply determined.

<Third embodiment>
The single-panel color image capturing apparatus of the third embodiment of the present invention will be described. FIG. 6 shows a functional block of the single-panel color image capturing apparatus of the present embodiment. In the present embodiment, image signal processing for further improving the image quality is added, and the configuration of the color image image sensor including the image sensor 1 and the light color filter is the first embodiment and the second embodiment. It is the same as the form.

In the present embodiment, RAW images are directly filtered to realize a demosaiking process that emphasizes the frequency characteristics of the W component. In the present embodiment, since all the pixels of the image sensor have a W component, the frequency component of the RAW image has a W component in the low frequency band and a color difference component in the high frequency band. As for how to spread the spectrum, as shown in FIG. 8A, the W component is wide and the color difference component is narrow. A wide band W component can be obtained by separating the W component and the color difference component with a spatial filter suitable for this. However, it is not a pure white component that can be obtained by removing the high frequency component, and in the case of the PLSL arrangement shown in FIG. 2 (a), the number of cells of L is twice as many as the other, so that component W _l is given by the following equation. Will be.

Since G is (1-w) / 4 more than others, W _l is a little closer to brightness.

The two-dimensional frequency spectrum of the RAW image is shown in FIG. In FIG. 7, μ is the horizontal frequency, ν is the vertical frequency, and the end is the maximum frequency determined by the number of cells of the image sensor. In the spectrum, the DC (μ = 0, ν = 0) component is W _l . _{There is a difference between P and S with respect to W l at} high frequencies on the μ and ν axes, that is, R and B color difference components Cr = RW _l and Cb = BW _{l. There is} an L difference with respect to l, that is, a G color difference component Cg = GW _l . Places with all three components are simply indicated as C. As for the spread of the spectrum, since W _l is narrow in C, _{the cutoff (half) frequency when W l} is separated by the two-dimensional filter is as shown by the dotted line in FIG. 8 (b). Even if the filter color or arrangement is changed, the outline of the spectrum does not change.

In the present embodiment, since the W component and the C component are multiplexed, crosstalk associated with separation becomes a problem. At the edge or the like, the spectrum of the W component is widened, and a high frequency component is mixed in C. When the maximum frequency is reached, false color may occur, so it is desirable to remove the maximum frequency component with an optical filter. On the other hand, in the portion where the C signal changes sharply, a frequency component having a high color difference is mixed with the W component, resulting in shape distortion. However, the W signal is also likely to be an edge, and the W signal is mixed into a high frequency, so that it is difficult to understand visually.

The process executed by the single-panel color image imaging apparatus of the present embodiment will be described based on the functional block shown in FIG. The RAW image data obtained by the image sensor 1 is given to the WC separator 11. The WC separator 11 grasps the spectral status of the color difference component for each part of the image, and if the C (color difference) component is weaker than the W (white) component, it is set to W wide band, and if C is strong, it is set to C wide band. .. The cutoff frequency of the separation filter (spatial filter) is 85% of the total band in the W wide band and 50% in the C wide band in W. This is a variable LPF (low-pass filter) with about 8 taps, and is a vertical and horizontal separation type processing. That is, the WC separator 11 adaptively controls the cutoff frequency of the LPF according to the relative amount of the C component with respect to the W component. Then, the C signal is obtained from the residual of the W signal that has passed through the LPF and the original RAW image.

The detection of the C spectrum is naturally in the C band, but since the DC component is not passed through, the band is relatively narrow. The detected value is compared with the spectrum of W. If W and C are separated in advance with medium filter characteristics, detection of each of W and C becomes easy. The characteristics of these filters are shown in one dimension in FIG. 8 (b). The horizontal axis of FIG. 8B is the normalized horizontal frequency, and 1.0 is the maximum frequency determined by the number of cell pixels.

The W signal separated by the WC separator 11 can be used as it is. On the other hand, the C signal is converted from a high frequency component into a baseband by the color difference demodulator 12. This is demodulated by multiplying the signal corresponding to the modulation carrier for each color component. Alternatively, the interpolation may be performed using only the pixels of each component, as in the case of normal demosaiking. The obtained R-W signal and B-W signal are given to the waveform correction noise reducer 13.

The waveform correction noise reducer 13 performs noise reduction together with the same correction process as the RW corrector 5 and the BW corrector 6 shown in FIG. This process is a general non-linear low-pass filter, in which only the small-amplitude component is passed through the low-pass filter. At that time, the value of the small amplitude (non-linear characteristic), which is the strength of the filter, is controlled by the amplification degree (gain) of the photoelectrically converted image signal in the image sensor 1. Originally, the noise generated by the dark current does not change in each cell of the image sensor, so the higher the gain, the more the noise is amplified. Therefore, the value of the small amplitude that causes the filter to act in proportion to the gain is also increased. That is, the waveform correction noise reducer 13 adaptively controls the value of the small amplitude to be cut according to the magnitude of the gain of the photoelectrically converted image signal in the image sensor 1.

When such processing is applied to R, G, B after γ correction and the color difference signal formed from the R, G, B, the frequency characteristic of the luminance signal is deteriorated in the high saturation portion, and the image quality is deteriorated. This is because a part of the luminance signal component is transmitted as a color difference signal by γ correction, which is a non-linear conversion, and becomes more remarkable as the saturation is higher and the pass band is narrower. However, since the luminosity factor for the original color difference component is much lower than the luminance component, if the band is limited before the γ correction is performed, the deterioration of the luminance component as described above does not occur. Further, since the color difference signal basically has little change, the effect of noise reduction by the non-linear filter is high.

The RW signal and the BW signal processed in this way are given to the primary color matrix 7. The primary color matrix 7 is the same as that of the first embodiment, and converts W, RW, and BW into R, G, and B. The converted R signal is given to the γ corrector 14, the G signal is given to the γ corrector 15, and the B signal is given to the γ corrector 16. The gamma correctors 14, 15 and 16 perform gamma correction (y = x ^0.45 ) processing. The gamma-corrected R, G, B signals are output from the R signal output 8, the G signal output 9, and the B signal output 10. Gamma correction is a process required by the standard for moving images, but in this embodiment, it is a feature that noise reduction processing is performed before gamma correction.

Although the present invention has been described above with embodiments, the present invention is not limited to the above-described embodiments, and the actions and effects of the present invention can be exhibited within the scope of other embodiments that can be conceived by those skilled in the art. As long as it works, it is included in the scope of the present invention.

1 ... Image sensor, 2 ... Demodulation, 3 ... Light color matrix, 4 ... W corrector, 5 ... RW corrector, 6 ... BW corrector, 7 ... Primary color matrix, 8 ... R signal output, 9 ... G signal output, 10 ... B signal output, 11 ... WC separator, 12 ... color difference demodulator, 13 ... waveform correction noise reducer, 14, 15, 16 ... γ corrector

Claims (9)

It is a single-plate color image sensor in which a color filter that transmits light over the entire visible light is arranged on each light receiving cell of a single image sensor.
The color filter
A first light color filter that allows the first primary color component of the three primary color components contained in the light to pass through more than the remaining primary color components.
A second light color filter that allows the second primary color component of the three primary color components contained in the light to pass through more than the remaining primary color components.
A third light color filter that transmits a third primary color component out of the three primary color components contained in the light more than the remaining primary color components.
A color image sensor including. It is a single-plate color image sensor in which a color filter that transmits light over the entire visible light is arranged on each light receiving cell of a single image sensor.
The color filter
A first light color filter that transmits the first primary color component of the three primary color components contained in the light less than the remaining primary color components, and
A second light color filter that allows the second primary color component of the three primary color components contained in the light to pass through less than the remaining primary color components.
A third light color filter that transmits a third primary color component out of the three primary color components contained in the light less than the remaining primary color components.
Including
The cell containing the first light color filter, the cell containing the second light color filter, and the cell containing the third light color filter are color image pickup devices including a white filter in a part of each of the cells. A single-panel color image pickup apparatus including the color image pickup device according to claim 1 or 2.
A means for acquiring a first light color image corresponding to the first light color filter, a second light color image corresponding to the second light color filter, and a third light color image corresponding to the third light color filter.
A means for converting the three types of light-colored images into a white component and a color difference component between white by matrix calculation, and
A means for converting the white component and the color difference component into three types of primary color images by matrix calculation, and
Color image imaging device including. Means for performing the sharpening treatment of the white component and
A means for performing a sharpening treatment and a flattening treatment of the color difference component according to the edge of the white component.
The color image capturing apparatus according to claim 3. A single-panel color image pickup apparatus including the color image pickup device according to claim 1 or 2.
A means for separating a white component existing in a low frequency band from a RAW image corresponding to all cells of the color image sensor using a low-pass filter,
A means for obtaining a color difference component from the residual of the white component and the RAW image, and
A means for converting the white component and the color difference component into three types of primary color images by matrix calculation, and
Color image imaging device including. The means for separating
The cutoff frequency of the low-pass filter is adaptively controlled according to the relative amount of the color difference component with respect to the white component.
The color image capturing apparatus according to claim 5. A noise reducing means for cutting a small amplitude component of the color difference component is included, and the noise reducing means adaptively sets a small amplitude value to be cut according to the magnitude of the gain of the photoelectrically converted image signal in the image sensor. Characterized by controlling,
The color image capturing apparatus according to claim 5 or 6. A method of capturing a color image using the color image sensor according to claim 1 or 2.
A step of acquiring a first light color image corresponding to the first light color filter, a second light color image corresponding to the second light color filter, and a third light color image corresponding to the third light color filter.
A step of converting the three types of light-colored images into a white component and a color difference component between white by matrix calculation, and
A step of converting the white component and the color difference component into three types of primary color images by matrix calculation, and
Color image imaging method including. A method of capturing a color image using the color image sensor according to claim 1 or 2.
A step of separating a white component existing in a low frequency band from a RAW image corresponding to all cells of the color image image sensor using a low-pass filter, and
The step of acquiring the color difference component from the residual of the white component and the RAW image, and
A step of converting the white component and the color difference component into three types of primary color images by matrix calculation, and
Color image imaging method including.