KR20070099238A - Image sensor for expanding wide dynamic range and output bayer-pattern image - Google Patents

Abstract

An image sensor for broadening a dynamic range and outputting a bayer image is provided to synthesize two images whose exposure time differs, broaden the dynamic range, and output image data of a different bayer pattern to an existing interface. A pixel unit(410) converts an optical signal of an object inputted through a lens into an electric signal, and outputs bayer pattern data. A high-exposure interpolation unit(422) receives high-exposure bayer pattern data from the pixel unit(410), and converts the received high-exposure bayer pattern data into the first RGB image. A low-exposure interpolation unit(425) receives low-exposure bayer pattern data from the pixel unit(410), and converts the received low-exposure bayer pattern data into the second RGB image. An image synthesizing unit(430) synthesizes the first RGB image with the second RGB image. An RGB-bayer converting unit(440) converts the synthesized image into bayer data corresponding to a predetermined bayer pattern, and outputs the converted bayer data.

Description

Image sensor for expanding wide dynamic range and output Bayer-pattern image}

1 and 2 each show an RGB image with different exposure to light.

3 is an image to which the WDR is applied to the RGB image of FIGS. 1 and 2.

4 is a block diagram of an image sensor according to an exemplary embodiment of the present invention.

5 is a diagram illustrating color information of a Bayer image output through the CFA of the image sensor according to an embodiment of the present invention.

6 is a diagram illustrating color information interpolated corresponding to an RGB image through a color interpolation unit of an image sensor according to an exemplary embodiment of the present invention.

7 is a diagram illustrating a Bayer format preset in an image sensor according to an exemplary embodiment of the present invention.

8 is a diagram illustrating a color gain curve required for performing an image data synthesis operation in a WDR according to an embodiment of the present invention.

9 is a flowchart illustrating a method of outputting Bayer image data after performing WDR in an image sensor according to an exemplary embodiment of the present invention.

10 is an RGB image obtained by combining two images of RGB format having different exposure times.

FIG. 11 is an image obtained by converting the RGB image synthesized as shown in FIG. 10 to Bayer format.

<Explanation of symbols for main parts of the drawings>

410 pixels

422 high exposure interpolator

425 low exposure interpolator

430: image synthesis unit

440: RGB to Baye converter

The present invention relates to an image sensor, and more particularly, to an image sensor capable of outputting a Bayer image after synthesizing two images having different exposure times to widen a dynamic range.

Recently, portable devices (eg, digital cameras, mobile communication terminals, etc.) equipped with image sensors have been developed and sold. The image sensor is configured as an array of small photosensitive diodes called pixels or photosites. The pixels themselves usually do not extract color from the light and convert photons from the broad spectral band into electrons. Thus, in order to record color images with a single sensor, the sensor is filtered so that different pixels receive different color illumination. Sensors of this type are known as color filter arrays or color filter arrays (CFAs), and each different color filter is arranged in a predefined pattern across the sensor.

In determining the quality of an image sensor, one of the important criteria is the dynamic range. Dynamic range generally refers to the maximum range that a signal can be processed without distorting the input signal. In the case of an image sensor, the wider the dynamic range, the better the image can be obtained regardless of the wide range of brightness changes. However, the conventional color image sensor has a disadvantage in that the original color of the image may not be represented well when one or more of red, green, and blue colors are saturated due to a narrow dynamic range. In order to overcome the shortcomings of such a dynamic range, a method of synthesizing images having different exposure times has been attempted. That is, a method of synthesizing different images having different exposure times as shown in FIGS. 1 and 2 as shown in FIG. 3 has been attempted. However, such an image sensor has a problem that the interface with the existing image processing devices that receive the Bayer format image data is not matched by outputting the RGB image data.

Accordingly, an object of the present invention is to provide an image sensor capable of outputting a Bayer type image after widening the dynamic range by synthesizing two images having different exposure times.

Other objects of the present invention will be easily understood through the description of the following examples.

In order to achieve the above object, according to an aspect of the present invention, there is provided an image sensor capable of synthesizing two image signals obtained by varying the exposure time to the same subject to widen the dynamic range and output as a Bayer image.

According to a preferred embodiment of the present invention, the pixel unit for outputting the Bayer pattern data by converting the optical signal of the subject input through the lens into an electrical signal; A high exposure interpolation unit which receives the high exposure Bayer pattern data from the pixel unit and converts the high exposure Bayer pattern data into a first RGB image; A low exposure interpolation unit receiving low exposure Bayer pattern data from the pixel unit and converting the low exposure Bayer pattern data into a second RGB image; An image synthesizer configured to synthesize the first RGB image and the second RGB image; And an RGB to Bayer converter for converting the synthesized image into Bayer data corresponding to a predetermined Bayer format and outputting the Bayer data.

The pixel unit may output two kinds of Bayer pattern data generated by being exposed to a light source for a predetermined time.

The RGB-Bay converter removes a G component and a B component when the pixel to be converted of the synthesized image is an R pixel, and removes an R component and a B component when the pixel to be converted is a G pixel, and the pixel to be converted is a B pixel. The R component and the G component can be removed.

The low exposure interpolation unit and the high exposure interpolation unit may calculate the R component of the G pixel by using the following equation.

, 여기서, Kr1, Kr2는 상기 G 픽셀에 인접한 R 픽셀에 인접한 4개의 G 픽셀들의 평균값을 산출한 후 상기 R 픽셀의 값을 감산하여 산출된 값일 수 있다.

Here, Kr1 and Kr2 may be calculated by calculating an average value of four G pixels adjacent to an R pixel adjacent to the G pixel and subtracting the value of the R pixel.

The low exposure interpolation unit and the high exposure interpolation unit may calculate the B component of the G pixel by using the following equation.

, 여기서, Kr1, Kr2는 상기 G 픽셀에 인접한 B 픽셀에 인접한 4개의 G 픽셀들의 평균값을 산출한 후 상기 B 픽셀의 값을 감산하여 산출된 값일 수 있다.

Here, Kr1 and Kr2 may be calculated by calculating an average value of four G pixels adjacent to the B pixel adjacent to the G pixel and subtracting the value of the B pixel.

The low exposure interpolation unit and the high exposure interpolation unit may calculate the G component of the B pixel by using the following equation.

Here, B2 may be an interpolation target pixel, B1 and B3 may be pixels horizontally adjacent to B2 pixels, B1 'and B3' may be pixels adjacent to B2 pixels vertically, and may be G pixels adjacent to B2 pixels.

The low exposure interpolation unit and the high exposure interpolation unit may calculate the R component of the B pixel by using the following equation.

, 여기서, G는 상기 B 픽셀에 상응하여 산출된 G 성분이며, R1, R2, R3, R4는 상기 B 픽셀에 인접한 4개의 R 픽셀들이고, g1, g2, g3, g4는 상기 B픽셀에 인접한 4개의 R 픽셀에 인접한 4개의 G 픽셀들의 평균값일 수 있다.

Where G is a G component calculated corresponding to the B pixel, and R1, R2, R3, and R4 are four R pixels adjacent to the B pixel, and g1, g2, g3, and g4 are 4 adjacent to the B pixel. It may be an average value of four G pixels adjacent to the R pixels.

The image synthesizing unit may synthesize the interpolation target pixels using the following equation.

,

,

는 칼라 게인 곡선에 의해 산출된 칼라 게인값이며, S는 '0' 초과 '1' 미만의 임의의 실수값일 수 있다.Here, M is the minimum value among the color components of the interpolation target pixel, r, g, b are the color component value of the pixel of the image data input from the high exposure interpolation unit,

,

,

Is the color gain value calculated by the color gain curve, and S may be any real value greater than '0' and less than '1'.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In the following description, the same or corresponding elements will be described with the same reference numerals regardless of the reference numerals in the description with reference to the accompanying drawings.

4 is a block diagram of an image sensor according to an exemplary embodiment of the present invention, and FIG. 5 is a diagram illustrating color information of a Bayer image output through a CFA of an image sensor according to an exemplary embodiment of the present invention. FIG. 6 is a diagram illustrating color information interpolated corresponding to an RGB image through a color interpolation unit of an image sensor according to an exemplary embodiment of the present invention, and FIG. 7 is an image sensor according to an exemplary embodiment of the present invention. FIG. 8 is a diagram illustrating a preset Bayer format, and FIG. 8 is a diagram illustrating a color gain curve required for performing an image data synthesis operation in a WDR according to an exemplary embodiment of the present invention.

Referring to FIG. 4, an image sensor 400 according to an exemplary embodiment of the present invention may include a pixel unit 410, a high exposure interpolation unit 422, a low exposure interpolation unit 425, an image synthesis unit 430, and the like. The RGB to Bayer converter 440 is configured.

The pixel unit 410 (color filter array) performs a function of converting and outputting the signal of the optical object input through the lens into an electrical signal. That is, the pixel unit 410 transmits two image signals acquired by different exposure times to the same subject to the color interpolator 420. The pixel unit 410 uses a Bayer pattern which is advantageous in terms of resolution, and image data having one color information is output for each pixel. For example, image data having only an R component is output from a pixel corresponding to an R pattern, image data having only a G component is output from a pixel corresponding to a G pattern, and an image having only a B component from a pixel corresponding to a B pattern. Data can be output. An example of outputting image data having only one color component in correspondence with a Bayer pattern is illustrated in FIG. 5. Of course, an A / D converter (not shown) that converts analog image data output by the pixel unit 410 into digital image data and outputs the digital image data is connected to the pixel unit 410. That is, the A / D converter converts analog Bayer image data output through the pixel unit 410 into digital Bayer image data and transfers the digital Bayer image data to the high exposure interpolation unit 422 and the low exposure interpolation unit 425.

For example, the pixel unit 410 includes a plurality of pixel units, and each pixel unit includes one photo diode and a MOS transistor for reading charges collected in the photo diode. Since the pixel unit 410 does not have a separate storage unit, in order to continuously read an image, light is exposed to each column and immediately read.

The color interpolation unit 420 includes a high exposure interpolation unit 422 and a low exposure interpolation unit 425. The color interpolation unit 420 receives two different Bayer pattern data obtained by different exposure times from the pixel unit 410. Converts the image into an RGB image and outputs the image to the image combining unit 430.

Here, the color interpolator 420 may interpolate color information required for each pixel by an interpolation method such as bilinear interpolation, neighbor pixel interpolation, effective interpolation, and the like. In this specification, the color interpolator 420 will focus on interpolating color information not present in each pixel using an effective interpolation method.

For example, assume that a Bayer pattern used to output Bayer image data output from the pixel unit 410 is illustrated in FIG. 7. In this case, the pixel unit 410 may generate Bayer image data as shown in FIG. 5 and transmit the Bayer image data to the high exposure interpolator 422 or the low exposure interpolator 425. Hereinafter, since the method of converting the Bayer pattern data to the RGB image in the high exposure interpolation unit 422 and the low exposure interpolation unit 425 is the same, a description will be given of converting the high exposure interpolation unit 422 into an RGB image. .

First, a brief description will be given of a method of calculating the R component and the B component focusing on pixels having only the G component. For convenience of description, a method of interpolating the R component and the B component of the center pixel (ie, (3,3) pixel) 533 in FIG. 7 will first be briefly described. Define (2,3) pixels 523 as "R1 pixels" with (3,3) pixels as the center, (3,2) pixels 532 as "B1 pixels", and (4,3) Define pixel 543 as "R2 pixel" and define (3,4) pixel as "B2 pixel". The average value of four adjacent G components based on the R1 pixel is g1, and the average value of four adjacent G components based on the B1 pixel is g2, and the average value of four adjacent G components based on the R2 pixel is g3 and B2 pixels. Define g4 as the average of four adjacent G components.

In the present specification, pixels used to obtain average values of adjacent G components, such as R1, B2, R3, and B4, will be defined as "reference pixels."

In this way, a new variable is defined by subtracting the component value of the reference pixel from the pixel values of the G components calculated with respect to the reference pixels. For example, the pixel value corresponding to the R1 pixel is subtracted from g1 to be defined as Kr1, the pixel value corresponding to the B2 pixel is subtracted from g2 to be defined as Kb2, and the pixel value corresponding to the R3 pixel is subtracted from g3. Let it be defined as Kr3 and subtract the pixel value corresponding to B4 pixel from g4 as Kb4.

In this state, the R component of the (3,3) pixel may be calculated using, for example, Equation 1 below, and the B component of the (3,3) pixel may be, for example, represented by Equation 2 below. Can be calculated using.

In this manner, the R and B components of pixels having all G components can be calculated.

Hereinafter, a pixel having only a B component will be referred to as a "B pixel", a pixel having only a R component will be referred to as an R pixel, and a pixel having only a G component will be referred to as a G pixel.

Next, a method of calculating the R component and the G component or the B component and the G component of the B pixel or the R pixel will be described.

First, a method of calculating the R and G components of the (3,4) pixel will be described. To calculate the G component of the (3,4) pixel, first calculate the delta value of the B components along the horizontal and vertical axes adjacent to the (3,4) pixel. For example, the delta of the horizontal B component centered on (3,4) pixels is subtracted by subtracting the mean value of the pixels of (3,2) and (3,6) from the (3,4) pixel value. The delta per pixel of the B component can be calculated. In this way, the delta per pixel of the B component is calculated also on the vertical axis. The average value of the delta per pixel of the B component in each of the horizontal and vertical axes will be defined as "X" for convenience. In this state, an average value of four G pixels adjacent to the (3,4) pixel may be calculated and added to X to calculate the G component of the (3,4) pixel.

For example, the high exposure interpolation unit 422 may calculate the G component of the B pixel by using Equation 3.

Here, B2 is a (3,4) pixel value, B1 and B3 are pixel values corresponding to B pixels horizontally adjacent to B2 pixels, and B1 'and B3' are B pixels vertically adjacent to B2 pixels. Is a pixel value corresponding to G1, G2, G3, and G4 are pixel values corresponding to G pixels adjacent to the B2 pixel.

The method of calculating the R component of the (3,4) pixel is as follows. For convenience of explanation, the R pixels adjacent to the (3,4) pixels are defined as R1, R2, R3, and R4. Calculate the average value of four adjacent G components based on R1, R2, R3, and R4, and define g1, g2, g3, and g4. Since this is similar to that described in the method of calculating the R component and the B component of the G pixel, detailed description thereof will be omitted.

In this way, the R component can be calculated by subtracting the average value obtained by subtracting each reference pixel value from the average value of g1, g2, g3, and g4 calculated from the G component values of the (3,4) pixels. For example, the high exposure interpolation unit 422 may calculate the R component of the B pixel using Equation 4 below.

The method for calculating the B component and the G component of the R pixel is the same as the method for calculating the R component and the G component of the B pixel. Therefore, duplicate description thereof will be omitted.

The high exposure interpolation unit 422 may generate image data by interpolating Bayer data input from the pixel unit 410 and color information corresponding to each pixel in the same manner as described above, and transmit the image data to the image synthesis unit 430. have.

In the present specification, the high exposure interpolation unit 422 and the low exposure interpolation unit 425 respectively interpolate color information using different images. Therefore, since the description of the low exposure interpolation unit 425 overlaps with the high exposure interpolation unit 422, the description thereof will be omitted.

The high exposure interpolation unit 422 and the low exposure interpolation unit 425 respectively receive different Bayer pattern images from the pixel unit 410 and correspond to color information of each pixel in a predetermined format by the same method (that is, Interpolates color information corresponding to a predetermined image format) and outputs the image to the image synthesis unit 430.

The image synthesizing unit 430 synthesizes the image data input from the high exposure interpolation unit 422 and the low exposure interpolation unit 425, and transfers the image data to the RGB-bayer conversion unit 440. That is, the image synthesizing unit 430 may synthesize image data exposed to a light source differently by performing a wide dynamic range (WDR).

Hereinafter, in the present specification, image data interpolated by the high exposure interpolation unit 422 is defined as “first image data,” and image data interpolated by the low exposure interpolation unit 425 is defined as “second image data.” Let's do it.

For example, the first image data may be high exposure image data exposed to much light, and the second image data may be low exposure image data exposed to less light than the high exposure image data.

In addition, image data synthesized and output by the image synthesizing unit 430 will be defined as "synthetic image data."

과 같이 정의될 수 있다. 여기서, 컬러 게인을 S라 정의하면, S는

을 만족하는 임의의 값일 수 있으며 고정된 상수 값이 아닐 수도 있다.If the minimum value among the R, G, and B components of each pixel is defined as M, the color gain for each pixel is

It can be defined as Here, if the color gain is defined as S, S is

It may be any value that satisfies and may not be a fixed constant value.

The image synthesizing unit 430 may synthesize the first image data and the second image data using, for example, Equation 5 below.

Here, R ', G', and B 'are color information generated by synthesizing respective color information of the first image data and the second image data.

성분을 산출하기 위한 곡선이다. 여기서, 이해와 설명의 편의를 위해, 본 명세서에서는 제1 이미지 데이터는 고노출 이미지인것으로 가정하며, 제2 이미지 데이터는 저노출 이미지인것으로 가정하여 설명하기로 한다. 도 8을 참조하면, R1과 R2는 그래프의 영역을 나누기 위한 경계값이며, 제1

과 제2

에서 제1

이 클수도 있으며, 제2

이 클수도 있다. 도 8과 같이, 제1 이미지 데이터의 R 성분과 제2 이미지 데이터의 r성분에 상응하여 제1 이미지 데이터의 색상 정보에 의존하는 제1 영역, 제2 이미지 데이터의 색상 정보에 의존하는 제3 영역, 제1 영역과 제3 영역의 중간 영역인 제2 영역으로 각각의 구간을 분리할 수 있다. 도 8에서 제1 영역을 구분하기 위한 R1의 값이 예를 들어, 250(예를 들어, 제1 이미지 데이터 및 제2 이미지 데이터가 8비트 이미지 데이터라고 가정하면)이며, 제3 영역을 구분하기 위한 R2의 값이 예를 들어, 350이라고 가정하자. 이때, 제1 이미지 데이터의 R성분과 제2 이미지 데이터의 r성분을 합산한 R+r의 값이 250이라고 한다면, 제1 이미지 데이터의 R 성분이 포화(saturation)에 근접하지만 아직 미치지 못하였다는 것을 의미한다. 따라서, 이와 같이 제1 이미지 데이터 및 제2 이미지 데이터의 각각의 성분을 합산한 값이 R1 이하이면, 제1 이미지 데이터의 색상 성분이 신뢰도가 더 높다. 따라서, 이 경우, 제1 이미지 데이터의 색상 성분으로 해당 픽셀의 색상 성분을 결정할 수 있다. First, FIG. 8 relates to an R component.

Curve for calculating the components. Here, for convenience of understanding and description, it is assumed herein that the first image data is a high exposure image and the second image data is assumed to be a low exposure image. Referring to FIG. 8, R1 and R2 are boundary values for dividing an area of the graph.

And second

1st in

This may be large, second

This might be big. As shown in FIG. 8, a first area depending on color information of the first image data and a third area depending on color information of the second image data, corresponding to the R component of the first image data and the r component of the second image data. Each section may be divided into a second region, which is an intermediate region between the first region and the third region. In FIG. 8, for example, a value of R1 for dividing the first region is 250 (for example, assuming that the first image data and the second image data are 8-bit image data), and to distinguish the third region. Suppose the value of R2 for is 350, for example. In this case, if the R + r value obtained by adding the R component of the first image data and the r component of the second image data is 250, the R component of the first image data is close to saturation, but not yet reached. Means that. Thus, when the sum of the respective components of the first image data and the second image data is R1 or less, the color component of the first image data is more reliable. Therefore, in this case, the color component of the pixel may be determined as the color component of the first image data.

In addition, when the sum of the respective components of the first image data and the second image data is equal to or larger than R2, it means that the saturation of the color components of the first image data is reached. Therefore, in this case, since the color component of the second image data has higher reliability, the color component of the pixel may be determined as the color component of the second image data. The second region may be represented by a straight line or a curve by a function (experimentally calculated) which is a middle region between the first region and the third region. In FIG. 8, only the method of obtaining the color gain of the R component is described, but the B and G components may also obtain the color gain by generating a color gain curve in the same manner as the R component. In this manner, the image synthesizing unit 430 may generate one image data (that is, composite image data) by adding the first image data and the second image data and transfer the generated image data to the RGB-bayer converter 440.

The RGB-bayer converter 440 converts the composite image data inputted through the image synthesizer 430 into Bayer image data and outputs them to an image processor unit (not shown). For example, the RGB-bayer converter 440 may convert the Bayer image data by discarding two pieces of color information for each pixel corresponding to a predetermined Bayer pattern in the RGB image data.

For example, assume that the RGB image data of FIG. 6 is changed to Bayer image data corresponding to the Bayer pattern of FIG. 7. The RGB-bayer converter 440 may generate 611 pixels as pixels having only G components by leaving only the G components and discarding the R and B components. In this way, the RGB-Bay converter 440 performs an operation of discarding the other two color components by leaving only one color component information in each pixel of the RGB image data using the Bayer format of FIG. 7. The Bayer image data corresponding to the Bayer pattern can be converted.

9 is a flowchart illustrating a method of outputting Bayer image data after performing a WDR in an image sensor according to an exemplary embodiment of the present invention, and FIG. 10 is a RGB format having different exposure times according to an exemplary embodiment of the present invention. An image obtained by synthesizing two images of (RGB format), and FIG. 11 is an exemplary view of an image obtained by converting the synthesized image of FIG. 10 into a Bayer format according to an exemplary embodiment of the present invention. Hereinafter, a process in which two images having different light exposures in the pixel unit 410 are generated and output by a predetermined time difference will be described.

In operation 910, the pixel unit 410 outputs the two Bayer pattern data obtained by varying the exposure time to the subject to the high exposure interpolator 422 and the low exposure interpolator 425.

For example, the pixel unit 410 may convert a signal of an optical object input through a lens into an electrical signal and output the converted signal.

In operation 915, the high exposure interpolation unit 422 or the low exposure interpolation unit 425 interpolates each of the different Bayer pattern data input through the pixel unit 410 by using a preset method to generate the first image data or the second image. The data is converted into data and transferred to the image synthesizer 430.

For example, the color interpolator 420 may interpolate the remaining color information of the target pixel using the color information of the neighboring pixels around the target pixel. Since the color interpolator 420 interpolates respective color information of the target pixel, the color interpolator 420 will be described in detail with reference to FIG. 4, and thus redundant descriptions thereof will be omitted.

In operation 920, the image synthesizing unit 430 synthesizes the first image data and the second image data input from the high exposure interpolation unit 422 and the low exposure interpolation unit 425 using a predetermined algorithm and converts RGB to Bayes. Transfer to section 440.

In operation 925, the RGB-Bay converter 440 removes the color components of each pixel of the composite image data input from the image synthesizer 430, leaving only the color components corresponding to the predetermined Bayer pattern and removing the remaining color components. Convert to an image. That is, the RGB-Bay converter 440 converts the composite image data input in the RGB format from the image synthesizer 430 into Bayer image data corresponding to a predetermined Bayer pattern and outputs the converted Bayer image data to an image processor unit (not shown). Can be.

As shown in Figures 10 to 11, it can be seen that the difference can not be confirmed with the naked eye. For a more sophisticated comparison, using MSE (mean square error) analysis, the color comparison of each pixel is obtained. The MSE of the two images is obtained as '1'. Here, the MSE is a number obtained by dividing the sum of the sum of the two pixels by adding the squares to make the absolute error of each pixel absolute for all pixels. Therefore, if MSE is calculated using the same two images, "0" is obtained. The higher the two images, the larger the MSE value is obtained.

Therefore, since the MSE obtained by comparing the images of FIGS. 10 to 11 is '1', it can be seen that the two images are almost identical images.

As described above, by providing an image sensor and a method for outputting a Bayer image after widening the dynamic range according to the present invention, it is possible to widen the dynamic range by synthesizing two images having different exposure times, and different from the existing interface. This has the effect of outputting Bayer format image data.

Although the above has been described with reference to a preferred embodiment of the present invention, those skilled in the art to which the present invention pertains without departing from the spirit and scope of the present invention as set forth in the claims below It will be appreciated that modifications and variations can be made.

Claims (8)

A pixel unit converting an optical signal of a subject input through the lens into an electrical signal and outputting Bayer pattern data; A high exposure interpolation unit which receives the high exposure Bayer pattern data from the pixel unit and converts the high exposure Bayer pattern data into a first RGB image; A low exposure interpolation unit receiving low exposure Bayer pattern data from the pixel unit and converting the low exposure Bayer pattern data into a second RGB image; An image synthesizer configured to synthesize the first RGB image and the second RGB image; And And an RGB to Bayer converter for converting the synthesized image into Bayer data corresponding to a predetermined Bayer format and outputting the Bayer data. The method of claim 1, And the pixel unit outputs two kinds of Bayer pattern data generated by being exposed to a light source for a predetermined time. The method of claim 1, The RGB-Bay converter removes a G component and a B component when the pixel to be converted of the synthesized image is an R pixel, and removes an R component and a B component when the pixel to be converted is a G pixel, and the pixel to be converted is B. If it is a pixel, the image sensor, characterized in that to remove the R and G components. The method of claim 1, And the low exposure interpolation unit and the high exposure interpolation unit calculate an R component of a G pixel by using the following equation.

Here, Kr1 and Kr2 are values calculated by subtracting the value of the R pixel after calculating an average value of four G pixels adjacent to the R pixel adjacent to the G pixel. The method of claim 1, And the low exposure interpolation unit and the high exposure interpolation unit calculate a B component of the G pixel by using the following equation.

Here, Kr1 and Kr2 are values calculated by subtracting the value of the B pixel after calculating an average value of four G pixels adjacent to the B pixel adjacent to the G pixel. The method of claim 1, And the low exposure interpolation unit and the high exposure interpolation unit calculate a G component of the B pixel by using the following equation.

Here, B2 is an interpolation target pixel, B1 and B3 are pixels horizontally adjacent to the B2 pixel, B1 'and B3' are pixels vertically adjacent to the B2 pixel and G pixels adjacent to the B2 pixel. The method of claim 6, And the low exposure interpolation unit and the high exposure interpolation unit calculate the R component of the B pixel by using the following equation.

Here, G is a G component calculated corresponding to the B pixel, R1, R2, R3, and R4 are four R pixels adjacent to the B pixel, and g1, g2, g3, and g4 are four adjacent to the B pixel. Average of four G pixels adjacent to an R pixel. The method of claim 1, The image synthesizing unit synthesizes the interpolation target pixel by using the following equation.

Here, M is the minimum value among the color components of the interpolation target pixel, r, g, b are the color component value of the pixel of the image data input from the high exposure interpolation unit,

,

,

Is the color gain value calculated by the color gain curve, and S is any real value greater than '0' and less than '1'.

Images