KR20070069890A - Image senser and method for down scaling image - Google Patents

Abstract

An image sensor and an image resizing method are provided to reduce capacity of a line memory which requires high power when an image is resized and acquire images with high quality. An image sensor includes a pixel array(301), a timing controller(309), a row decoder(303), a column decoder, an analog-to-digital converter(307), a horizontal smoothing unit(311), a first image resizing circuit(313), a color interpolator(315), a color space converter(317), a second image resizing circuit(319), an output formatter(321), and a line memory(323). The first image resizing circuit scales down a Bayer image having RGB information. The color space converter converts the Bayer image into an image having YUV information. The second image resizing circuit scales down the image having the YUV information. The line memory provides the same line memory capacity when the first and second image resizing circuits scale down the images.

Description

Image sensor and image reduction method {IMAGE SENSER AND METHOD FOR DOWN SCALING IMAGE}

1 and 2 are block diagrams illustrating an image sensor including an image reduction unit according to the related art.

3 is a block diagram illustrating an image sensor including an image reduction unit according to an exemplary embodiment of the present invention.

FIG. 4 is a diagram illustrating a calculation method of bilinear interpolation of first order reduction in the first image reduction unit of FIG. 3. FIG.

5A and 5B are diagrams illustrating a calculation method of bilinear interpolation of second order reduction in the second image reduction unit of FIG. 3.

FIG. 6 is a photograph (b) illustrating a case in which the Bayer image of the original image is reduced by 1 to 1/2 magnification by the first reduction in the original image (a) and the first image reduction unit of FIG. 3.

7 is a photograph (a) taken by using the image sensor of Figure 1 of the prior art and a photograph (b) taken by using an image sensor of an embodiment of the present invention.

Explanation of symbols on the main parts of the drawings

301: pixel array 303: row decoder

305: column decoder 307: A / D converter

309: timing controller 311: horizontal smooth down

313: First image reduction unit 315: Color interpolation unit

317: color space conversion unit 319: second image reduction unit

321: output formatter 323: line memory

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to semiconductor manufacturing technology, and more particularly, to an image sensor and an image reduction method during a semiconductor device manufacturing process.

In general, image sensors are used in a wide range of fields, from home products such as digital cameras and mobile phones, to endoscopes used in hospitals, and to satellite telescopes around the earth. The CMOS image sensor is a device that converts an optical image into an electrical signal, and employs a switching method in which a MOS transistor is formed by the number of pixels and the output is sequentially detected using the MOS transistor.

Currently, the image sensor of a mobile phone camera is being used as a mega image sensor having more than 1 million pixels. However, since the mobile phone is a portable device, the size of the video signal captured by the image sensor, which is an image processing device, is different from the size of the video signal displayed on the mobile phone. This is due to the limitation of the display capacity of the display unit that can be attached according to the miniaturization of the mobile phone, so that the number of pixels of the image signal read from the image sensor cannot be displayed as it is on the display unit of the mobile phone.

Thus, in order to solve the above problem, the image sensor includes an image reduction circuit to match the magnitude of the image signal photographed through the image sensor to the magnitude of the image signal displayed on the mobile phone.

1 and 2 are block diagrams illustrating an image sensor including an image reduction unit according to the related art.

1 and 2, an image sensor includes a pixel array, a timing controller, a row decoder, a column decoder, and an A / D converter. , Horizontal Smoothing, Image Resizing Circuit, Color Interpolation, Color Space Converter, Output Formatter and Line Memory do.

First, the operation of the image sensor illustrated in FIG. 1 will be briefly described. An analog video signal is output in a matrix form from a pixel array, and a row decoder of a pixel array is selected by a row decoder operating in response to the timing controller. Subsequently, the column decoder selects the column line for the row line and converts the image data into a digital signal in the A / D converter and outputs it. This is called a Bayer image. The Bayer image has R or G or B information for each pixel.

In the horizontal smoothing operation, a smoothing operation is performed by weakening or removing the variation or discontinuity of the Bayer image. Then, the Bayer image is reduced by using a line memory in the image reduction unit, which is performed by bilinear interpolation. Subsequently, the reduced Bayer image is interpolated by the color interpolator so that the Bayer image has R, G, and B information for each pixel.

In addition, the color space converter converts the Bayer image having the R, G, and B information into image data having the Y, U (Cr), and V (Cb) information, and outputs the image to a desired image format using an output formatter.

However, when the Bayer image outputted from the A / D converter is reduced by bilinear interpolation as shown in FIG. 1, when the distance between pixels to share the image information is a far Bayer image (see FIG. 4), when the magnification is 1/2 or more When color interpolation occurs, there is a problem that the pulse color (false color) is caught or the image is too smooth, resulting in poor sharpness.

As an image sensor for solving the above-mentioned problems, the operation of the image sensor of FIG. 2 will be briefly described. The pixel in a row decoder operating in response to a timing controller outputs an analog image signal in a matrix form from a pixel array. Select the row line of the array. Subsequently, the column decoder selects a column line for the row line, and converts the image data into a digital signal in the A / D converter and outputs the digital signal. This is called a Bayer image. The Bayer image has R, G, or B information for each pixel.

In the horizontal smoothing operation, a smoothing operation is performed by weakening or removing the variation or discontinuity of the Bayer image. Thereafter, the Bayer image is interpolated by the color interpolator so that the Bayer image has R, G, and B information for each pixel.

Then, the color space conversion unit converts the Bayer image having the R, G, and B information into image data having the Y, U (Cr), and V (Cb) information, and Y, U (Cr), and V (Cb). The image reduction unit performs the reduction operation using the bilinear interpolation method and the line memory on the image data having the information.

Then, the image data having the reduced Y, U (Cr), and V (Cb) information is output in a desired image format by the output formatter.

The image sensor of FIG. 2 as described above converts image data having Y, U (Cr), and V (Cb) information and reduces it, thereby solving the problem of FIG. 1. This is because the distance between pixels to share the image information is close.

However, when the image sensor of FIG. 1 directly reduces the Bayer image, since the R or B or G information of one pixel has one channel, the capacity of the line memory does not consume much, but the image sensor of FIG. Since Y, U (Cr), V (Cb) information, that is, three channels (but U (Cr), V (Cb) information is complementary to each other, may be regarded as two channels), The problem is that the power consumption of the line sensor, which consumes more power than the image sensor of 1 and uses a lot of chip area, is used.

The present invention has been proposed to solve the above-mentioned problems of the prior art, and provides an image sensor and an image reduction method for reducing the capacity of a line memory which consumes a lot of power during image reduction and obtaining a high quality image. It is done.

According to an aspect of the present invention for achieving the above object, the problem of falling sharpness by the pulse image or excessive smoothing by reducing the Bayer image having the conventional R or B or G information, Y, U (Cr), When the image data with V (Cb) information is reduced, excessive consumption of line memory causes power consumption. Bayer images with R, B, or G information using the same line memory capacity are 1 to 1 First order reduction that reduces to 1/2 magnification and second order reduction that reduces image data having Y, U (Cr), V (Cb) information to 1/2 to 1 / n (n is at least 2 or more) It provides an image sensor and a reduction method of the image sensor to solve the above problems.

Hereinafter, the preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the technical idea of the present invention. .

3 is a block diagram illustrating an image sensor including an image reduction unit according to an exemplary embodiment of the present invention.

Referring to FIG. 3, the image sensor may include a pixel array 301, a timing controller 309, a row decoder 303, a row decoder 305, a column decoder 305, and an A / D converter. 307, A / D Converter, Horizontal Smoothing, Image Resizing Circuit_1, Color Interpolation 315, Color Space Converter 317, Color Space Converter ), A second image reduction unit 319, an output formatter 321, and a line memory 323.

Briefly describing the operation, the pixel array 301 outputs an analog image signal in a matrix form, and selects a row line of the pixel array 301 in the row decoder 303 that operates in response to the timing controller 309. do. Subsequently, the column decoder 305 selects a column line for the row line, and the A / D converter 307 converts the image data into a digital signal and outputs the digital signal. This is called a Bayer image. The Bayer image has R or G or B information for each pixel.

In the horizontal smoothing 311, a smoothing operation is performed by weakening or removing the variation or discontinuity of the Bayer image. Subsequently, the Bayer image is reduced by using a line memory in the first image reduction unit 313, which is performed through bilinear interpolation. In addition, the first image reduction unit 313 reduces the Bayer image by 1 to 1/2 magnification by the first order reduction, which is 1 to 1/2 when the Bayer image is reduced as shown in FIG. 6. This is because, in the case of magnification, the problem of inferior sharpness due to pulse color or excessive smoothing hardly occurs during color interpolation.

Subsequently, the reduced Bayer image is interpolated by the color interpolator 315 so that the Bayer image has R, G, and B information for each pixel.

Then, the color space conversion unit 317 converts the Bayer image having the R, G, and B information into the image data having the Y, U (Cr), and V (Cb) information. The reduction unit 319 performs the second reduction operation. As in the first reduction, the reduction operation is performed through bilinear interpolation, and when the image data having Y, U (Cr) and V (Cb) information is reduced, the distance between pixels to share the image information is close. The problem is that the sharpness of the image does not worsen. The second reduction is to be reduced at a magnification of 1/2 to 1 / n (n is at least two or more).

Subsequently, the output data of the second image reduction unit 319 is output by the output formatter 321 in a desired image format.

In addition, the first and second order reduction operations use the line memory 323. The problem of the waste of the line memory 323 by three (or two) channels is primarily used as shown in FIG. The first reduction operation, the second use of the line memory 323 used in the first reduction by converting the Bayer image into image data having Y, U (Cr), and V (Cb) information. The same amount of line memory 323 is used. That is, the usage of the line memory 323 used for the second reduction that was used on a large scale is set equal to the usage of the line memory 323 used for the first reduction used on a small scale. Thus, power consumption of the line memory 323 is reduced.

For example, assuming a 20x20 image is input, the first image reduction unit 313 (when the maximum reduction ratio is 1/2) is reduced to an image of at least 10x10 using two lines of memory and output. do. The amount of data is increased to 10x10x3 through the color subinterval 315, and the R, G, and B information is changed into Y, U, and V information through the color space converter 317. At this time, the data amount is 10x10x3 as well. The image data changed into Y, U, and V information through the color space converter 317 is reduced to an image of at least 4 × 4 through the second image reduction unit 319 (when the maximum reduction ratio is 1/2). At this time, the amount of memory required when the second image reduction unit 319 is reduced by 1 to 1/2 magnification is 10 × 10 × 2, which is the same as the amount of memory required by the first image reduction unit 313. This is possible by using only one of the U and V information per pixel, since the human eye is insensitive to chrominance, so even if only one of the U and V information is used, image deterioration is not felt.

4 is a diagram illustrating a method of calculating a bilinear interpolation method of first order reduction in the first image reduction unit of FIG. 3, and FIGS. 5A and 5B illustrate a second order reduction in the second image reduction unit of FIG. 3. It is a figure which shows the calculation method of the bilinear interpolation method.

First, referring to FIG. 4, a 1 / n reduction using a Bayer image having R, B, or G information is shown. A bilinear interpolation method using the same is as follows.

(First equation)

or,

(Second equation)

5A and 5B illustrate 1 / n reduction using image data having Y, U (Cr), and V (Cb) information. Bilinear interpolation using the same is referred to in the first and second equations. The pixels g1, g2, g3, and g4 to be replaced with y1, y4, y2, y3 or uv1, uv4, uv2, and uv3, respectively.

7 is a photograph (a) taken using the conventional image sensor 1, and a photograph (b) taken using an image sensor according to an embodiment of the present invention.

As can be clearly seen with reference to FIG. 7, it can be seen that the photograph (b) taken with the image sensor of the present invention is clearer than the photograph (a) taken with the image sensor of the prior art in the part having a human face or fine lines. have. In addition, the photograph (a) taken by the image sensor of the prior art can be seen that a lot of pulse color occurs in the high frequency region (where there are many thin lines), but the photograph (b) taken by the image sensor of the present invention almost occurs I can see that it does not.

As described above, the Bayer image having the conventional R, B, or G information is reduced to reduce the sharpness caused by the pulse image or excessive smoothing, and the image data having the Y, U (Cr), and V (Cb) information. In order to solve the problem of power consumption by reducing the excessive use of the line memory 323, in the present invention, a Bayer image having R, B, or G information while using the same line memory capacity is 1 to 1/2 magnification. Through the first order reduction to reduce the image data and the second order reduction to reduce the image data having Y, U (Cr), V (Cb) information by 1/2 ~ 1 / n (n is at least 2) magnification Solve the problems.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes can be made in the art without departing from the technical spirit of the present invention. It will be clear to those of ordinary knowledge.

As described above, the present invention includes a first image reduction unit and a second image reduction unit to solve the problem of inferior sharpness due to pulse color and excessive smoothing.

In addition, the power consumption is reduced by using the same line memory capacity in the first reduction operation and the second reduction operation.

Claims (9)

An image sensor which processes a pixel data output from a pixel array and generates a Bayer image, A first image reduction unit configured to first reduce the Bayer image having RGB information; A color space converting unit converting the first reduced Bayer image to have YUV information; A second image reduction unit configured to second reduce the image having the YUV information; And Line memory that provides the same line memory capacity when the first and second shrinks Image sensor comprising a. According to claim 1, The image sensor may further include a color interpolator configured to perform color interpolation on the Bayer image reduced in the first image reducer. According to claim 1, And the first reduction and the second reduction are to be reduced by using bilinear interpolation. According to claim 1, The first reduction is an image sensor, characterized in that the reduction of 1 to 1/2 magnification. The method of claim 1, And said second reduction is a reduction of 1/2 to 1 / n (n is at least 2) magnification. An image sensor which processes a pixel data output from a pixel array and generates a Bayer image, First reducing the Bayer image having RGB information using a line memory; Color interpolating the first reduced Bayer image; Converting the color interpolated image to have YUV information; And Reducing the image having the YUV information to the same line memory capacity using the line memory; Image reduction method of the image sensor comprising a. The method of claim 6, The first reduction and the second reduction step is to reduce by using bilinear interpolation method. The method of claim 6, The first step of reducing the image reduction method of the image sensor, characterized in that for reducing to 1 to 1/2 magnification. The method of claim 6, And the second reduction step is to reduce the magnification by 1/2 to 1 / n (n is at least 2 or more).

Images