JP2010268389A - Apparatus and method for generation of zoom image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a zoom image generating apparatus which uses a linear interpolation method and also includes measures to cope with moire reduction and contour blurring. <P>SOLUTION: The zoom image generating apparatus is configured to generate a zoom image, by calculating coordinate values on a source image by zooming coordinate values of respective pixels constituting an image after zoom, to coordinate values corresponding to coordinate values of pixels constituting a source image before zoom, and then by calculating pixel values of the zoom image through linear interpolation using pixel values of the surrounding source image surrounding the coordinate values. The zoom image generating apparatus includes: a coordinate transforming section for transforming coordinate values of pixels to be interpolated on the source image into coordinate values within a predetermined region inside a periphery including pixels closest to the pixels of the source image of the periphery surrounding the coordinate values; an interpolation operating section for calculating pixel values of the pixels after zoom through linear interpolation based on the transformed coordinate values. Furthermore, the predetermined region is adjusted in accordance with the size of a zoom rate to become more effective for measures to cope with moire reduction and contour blurring. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a scaled image generating apparatus and a scaled image generating method for enlarging or reducing an image by performing linear interpolation using pixel values of peripheral pixels.

Conventionally, copying machines, facsimile machines, printers, and the like are known as image processing apparatuses capable of enlarging / reducing an original image.
When performing such enlargement / reduction, an interpolation method is widely used in which pixel values of a plurality of pixels existing in a certain area in an image are used to generate new pixel values within the area. Typical examples include the nearest neighbor method (Nearest Neighbor), the linear interpolation method (Bilinear), and the cubic interpolation method (Bicubic).

  The nearest neighbor method is an interpolation processing method for generating the pixel value of the existing pixel closest to the interpolation pixel, that is, the pixel value of the nearest existing pixel as the pixel value of the interpolation pixel. For this reason, a pixel value that preserves the sharpness of the edge can be generated in the vicinity of the edge, but on the other hand, jaggies appear where the curves and diagonal lines appear jagged and the image quality deteriorates.

The linear interpolation method is an interpolation processing method that generates a pixel value of an interpolation pixel by performing a second-order linear interpolation using a weighting factor based on the position of the interpolation pixel with respect to an existing pixel.
Therefore, although the image can be smoothed, the edge portion in the image is also processed smoothly, so that the edge is dulled, the image is blurred, and the image quality is deteriorated. There is also a problem that moire occurs at a specific magnification.

The cubic interpolation method is one of the cubic nonlinear interpolations, and includes four existing pixels around the interpolation pixel and twelve existing pixels around the four existing pixels. This is an interpolation processing method for performing arithmetic processing using a total of 16 existing pixels.
As a result, the image is smoothed and edge dullness is prevented, so that a smooth and natural image can be generated. On the other hand, it requires extremely complex arithmetic processing compared to the nearest neighbor method and linear interpolation method. Therefore, it is impossible to suppress an increase in the circuit scale for executing the arithmetic processing, and as a result, it is impossible to suppress an increase in the apparatus scale.

As described above, since the conventional interpolation method has drawbacks, the following interpolation processing techniques have been proposed.
In the data processing apparatus and method of Patent Document 1, by changing the data point using a random number when scaling the original image data, and thereafter performing interpolation calculation to create the scaled image data, Moire is being reduced.
Further, in the image enlargement method of Patent Document 2, the contour is blurred by switching between the nearest neighbor method (zero-order hold method) and the linear interpolation method when the interpolation position is a certain distance or more from the original pixel and other cases. Is reduced.

JP 09-128528 A JP-A-11-96348

  However, Patent Document 1 does not provide an improvement measure for blurring the outline, and Patent Document 2 does not reduce moire.

  The present invention has been made in consideration of the above-described circumstances, and provides a variable-magnification image generation apparatus and a variable-magnification image generation method that use a linear interpolation method and have both moiré reduction and a countermeasure for blurring the outline. The purpose is to do.

  In order to solve the above-described problem, the scaled image generation device of the present invention corresponds to the coordinate value of each pixel constituting the image after scaling with the coordinate value of the pixel constituting the original image before scaling. Scales to the coordinate value to calculate the coordinate value on the original image, and uses the pixel value of the surrounding original image surrounding the coordinate value to calculate the pixel value of the scaled image by linear interpolation to generate the scaled image In the image scaling generation device, the coordinate value of the pixel to be interpolated on the original image is within a predetermined region in the periphery including the pixel closest to the pixel of the surrounding original image surrounding the coordinate value. A coordinate conversion unit that converts to a coordinate value and an interpolation calculation unit that calculates a pixel value of the pixel after scaling by linear interpolation based on the converted coordinate value are provided.

  Accordingly, it is possible to provide a zoomed image generating apparatus that uses a linear interpolation method and has both a moire reduction and a countermeasure for blurring the outline.

  Further, in the coordinate conversion unit, bleeding and moire can be further reduced by adjusting the predetermined area in accordance with the size of the magnification.

  Further, the scaled image generation method of the present invention scales the coordinate value of each pixel constituting the image after scaling to the coordinate value corresponding to the coordinate value of the pixel constituting the original image before scaling. In an image scaling generation method for calculating a coordinate value on an original image and calculating a pixel value of the scaled image by linear interpolation using pixel values of a surrounding original image surrounding the coordinate value to generate a scaled image A coordinate value of the pixel to be interpolated on the original image is converted into a coordinate value in a predetermined region in the periphery including the pixel closest to the pixel in the peripheral image surrounding the coordinate value, Based on the converted coordinate value, the pixel value of the pixel after scaling is calculated by linear interpolation.

  Accordingly, it is possible to provide a scaled image generation method that uses a linear interpolation method and has both a moire reduction and a countermeasure for blurring the outline.

  Further, in the above-described zoom image generation method, bleeding and moire can be further reduced by adjusting the predetermined area according to the size of the zoom ratio.

  According to the present invention, moire and blurring of an image can be reduced even if the image is scaled using a linear interpolation method.

It is a figure explaining the general linear interpolation method (one dimension). It is a figure explaining a general linear interpolation method (two dimensions). It is a figure explaining the coordinate transformation of this invention (one dimension). It is a figure explaining the coordinate transformation of this invention (two dimensions). It is a figure explaining the adjustment of the range of a projection area | region. It is a block diagram which shows the structure of the magnification image generation apparatus of this invention. It is a density distribution figure when a linear interpolation method is applied after the coordinate conversion of this invention. It is a flowchart which shows the process sequence of the scaling image generation apparatus of this invention. FIG. 6 is a density distribution diagram using a zoom image generation method, a general linear interpolation method, and a nearest neighbor method according to the present invention.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments according to a scaled image generation apparatus and a scaled image generation method of the present invention will be described with reference to the drawings.

First, a general linear interpolation method will be described using an example of obtaining a density value at the midpoint between two pixels. In FIG. 1, A is the density value of the pixel whose X coordinate value is X0, B is the density value of the pixel whose X coordinate value is X1, and C is the density value of a point separated from the coordinate X0 by the distance X.
The density value C is obtained by the interpolation formula of (Formula 1).

Next, a method for obtaining the density value of the interpolation point expanded two-dimensionally will be described with reference to FIG. In FIG. 2, four points A ₀₀ , A ₁₀ , B ₀₁ , B ₁₁ are adjacent pixels (grid points), S 1 is a distance between points A ₀₀ and B ₀₁ (or A ₁₀ and B ₁₁ ), and S 2 is The distance between point A ₀₀ and point A ₁₀ (or point B ₀₁ and point B ₁₁ ), X is the distance from point A ₀₀ to point P in the X-axis direction of interpolation point P, and Y is in the Y-axis direction of interpolation point P. The distance from the point A ₀₀ to the point P is assumed. Further, the density values of the four points A ₀₀ , A ₁₀ , B ₀₁ , B ₁₁ and the point P are C ₀₀ , C ₁₀ , C ₀₁ , C ₁₁ and C, respectively.
The density value C is obtained by the interpolation formula of (Formula 2).

Next, the coordinate conversion of the present invention will be described.
In the present invention, a coordinate value before scaling corresponding to the coordinate value of the pixel after scaling (also referred to as the coordinate value of the interpolation point) is obtained, and a predetermined area including a pixel closest to the coordinate value of the interpolation point is determined. Project to a point (also called coordinate transformation).
Hereinafter, the predetermined area is referred to as a projection area, and a parameter for limiting the range of the projection area is provided. Hereinafter, this parameter is expressed by N.

  FIG. 3 is a diagram for explaining the coordinate transformation of the present invention, and shows the X coordinate axis for the sake of simplicity. Point A and point B are adjacent pixels, point C is the midpoint between point A and point B, and S is the distance between point A and point B. Point D is a point separated from point A by a distance N, and point E is a point separated from point B by a distance N.

In the coordinate conversion of the present invention, a point on the line segment AC is projected onto a point on the line segment AD, and a point on the line segment BC is projected onto a point on the line segment BE.
That is, if the point P on the line segment AB is separated from the point A by the distance X, the projected point P ′ is the distance X ′ from the point A as in (Expression 3) and (Expression 4). Calculated as a distant point.

FIG. 4 is a diagram for explaining a two-dimensional projection region. The projected area is an area indicated by diagonal lines in FIG. 4A, and is represented by a square having a side length of N for each pixel.
In the coordinate transformation of the present invention, the points in the four divided areas divided by the midpoints of the four adjacent pixels as shown in FIG. 4B are converted into the points in the projection area of the pixels included in the divided areas. It is to project.
That is, the point P ₁ is in the projection area of the point A ₀₀ , the point P ₂ is in the projection area of the point A ₁₀ , the point P ₃ is in the projection area of the point B ₀₁ , and the point P ₄ is in the projection area of the point B _11. Projected to a point.

  Here, the parameter N is appropriately set so as to obtain the best image quality according to the characteristics of the original image and the scaling factor (enlargement ratio / reduction ratio). However, if the scaling ratio (reduction) is small, it is important to prevent bleeding. Therefore, the value of the parameter N is made smaller (FIG. 5A), and if the scaling ratio (enlargement) is increased, interpolation is performed to smoothly enlarge the value. Since the position of the point is important, it is preferable to adjust the projection area in terms of image quality in accordance with the magnification by increasing the value of the parameter N (FIG. 5B).

Next, the zoom image generating apparatus of the present invention will be described.
FIG. 6 is a block diagram showing the configuration of the scaled image generation apparatus of the present invention. In the figure, the scaled image generation apparatus includes a coordinate generation unit 1, a coordinate conversion unit 2, an interpolation calculation unit 3, and an image data storage. The unit 4 and the scaled image data storage unit 5 are also included.

The image data storage unit 4 is a storage device that stores image data fetched from a scanner and image data sent from a PC or a photographing device.
The scaled image data storage unit 5 is a storage device that stores image data obtained by performing a scaling process on the image data stored in the image data storage unit 4.
The image data storage unit 4 and the scaled image data storage unit 5 are of a known data recording method, and in addition to the image data, information such as the number of pixels of the image width and height, and the distance between the pixels is recorded. Shall be.

When the scaling factor is given, the coordinate generation unit 1 calculates the coordinate value on the image before scaling corresponding to the coordinate value of each pixel of the image after scaling, and uses the calculated coordinate value as the coordinate conversion unit Send to 2.
For example, the magnification ratio of the width is zx, the magnification ratio of the height is zy, and the coordinate value (x, y) corresponding to the coordinate value (i, j) of the pixel of the image after scaling is calculated by (Equation 5). Is done.

ここで、ｉ，ｊは整数であり、ｘ、ｙ、ｚｘ、ｚｙは実数である。

Here, i and j are integers, and x, y, zx, and zy are real numbers.

  The coordinate conversion unit 2 interpolates the coordinate values (x ′, y ′) obtained by converting the coordinate values (x, y) sent from the coordinate generation unit 1 by the above-described (Expression 3) or (Expression 4). Send to part 3. At this time, the distances S <b> 1 and S <b> 2 between the pixels of the image before scaling are acquired from the image data stored in the image data storage unit 4.

  Based on the coordinate values (x ′, y ′) after the coordinate conversion sent from the coordinate conversion unit 2 and the image data stored in the image data storage unit 4, the interpolation calculation unit 3 calculates from (Equation 2). The density value of the image after scaling is calculated and written in the scaled image data storage unit 5.

First, the interpolation calculation unit 3 obtains coordinate values (lattice points) of four pixels surrounding coordinate values (x ′, y ′) after coordinate conversion in the image data before scaling.
If the value of the integer part of the X coordinate value x ′ is m and the value of the integer part of the Y coordinate value y ′ is n, the coordinate values of the four pixels are (m, n), (m + 1, n), (M, n + 1), (m + 1, n + 1).

  Next, density values and coordinate values (x ′, y ′) at four pixels (m, n), (m + 1, n), (m, n + 1), (m + 1, n + 1) of the image data before scaling. Then, the density value at the coordinate value (i, j) of the pixel after scaling is obtained by substituting the distance between the four pixels into (Equation 2).

  By calculating the above coordinate generation, coordinate conversion, and interpolation operations for all the pixels of the image after scaling, the image data after scaling is written into the scaled image data storage unit 5.

  FIG. 7 is a density distribution when the linear interpolation method is applied after coordinate conversion according to the present invention. As shown in FIG. 7, since an intermediate density between the density A and the density B that tends to cause blurring is not used, moire caused by repetition of the density that is blurred and a clear density close to the original pixel can be reduced.

Next, the processing procedure of the zoomed image generating apparatus will be described using the flowchart of FIG.
The user stores original image data obtained by a scanner, a PC, a photographing device, or the like in the image data storage unit 4 in advance, and designates the storage location and magnification of the original image from the scaled image generation device or an external device. The image scaling processing function is activated.

  When the image scaling process function of the scaled image generating apparatus is activated, it takes in the storage location and magnification of the designated image data, and activates the coordinate generation unit 1 (step S1).

The coordinate generation unit 1 uses the given scaling factor (width zx, height zy) to correspond to the coordinate value (i, j) of the image pixel before scaling, x, y) is calculated by (Formula 5), and the calculated coordinate value (x, y) is sent to the coordinate conversion unit 2 (step S2).
Thereafter, one pixel is selected from the pixels constituting the image after scaling, and the processes of steps S2 to S7 are performed.

  The coordinate conversion unit 2 performs coordinate conversion of the coordinate value (x, y) sent from the coordinate generation unit 1 according to (Expression 3) or (Expression 4), and the converted coordinate value (x ′, y ′). Is sent to the interpolation calculation unit 3 (step S3). At this time, the distances S <b> 1 and S <b> 2 between the pixels of the image before scaling are acquired from the image data stored in the image data storage unit 4.

Next, the interpolation calculation unit 3 obtains the coordinate values of four pixels surrounding the coordinate values (x ′, y ′) after conversion in the image data stored in the image data storage unit 4 (step S4).
Next, the density value is obtained by substituting the density values of these four pixels, the coordinate values (x ′, y ′), and the distances S1 and S2 between the four pixels into (Equation 2) (step S5). .
The obtained density value is written in the scaled image data storage unit 5 as the density value of the coordinate value (i, j) of the scaled pixel (step S6).

  By performing the processing from step S2 to S6 for all the pixels of the image after scaling, the image data after scaling is written into the scaled image data storage unit 5 (step S7).

Next, using an image (linear pattern) in which the densities 0, 0, 255, 0, 0, 255,... Are repeated, the scaled image generation method of the present invention, the general linear interpolation method, and the nearest neighbor method are used. The interpolation result will be described.
FIG. 9 is a density distribution diagram when the image of the linear pattern in the above example is enlarged to 111%. Here, the value of the parameter N representing the range of the projection area in the zoom image generation method of the present invention is set to 25% of the distance between the pixels (that is, N = S / 4).

Considering a line having a density value of 255 as a contour and paying attention to a density value of 200 or more, the peak portion calculated by the zoom image generation method of the present invention has a value of 200 or more, compared to a general linear interpolation method. It can be seen that blurring of the contour can be avoided.
On the other hand, the nearest neighbor method is the best result when only the contour is viewed, but has a drawback that a smooth image cannot be created when enlarged.

For example, consider a case where a gradation image (density 250, 230, 210,...) Is enlarged by 400%.
In the nearest neighbor method, the density does not become smooth as the density 250, 250, 250, 250, 230, 230, 230, 230, 210, 210, 210, 210,.
On the other hand, in the general linear interpolation method and the scaled image generation method of the present invention, the density is as in density 250, 245, 240, 235, 230, 225, 220, 215, 210, 205, 200, 195,. Since the changes smoothly, it can be seen that the gradation is also smooth.

  Therefore, it can be seen that the scaled image generation method of the present invention is sufficiently smooth compared to the nearest neighbor method.

  As for the reduction of moire, it is noticeable that the maximum and minimum of the density of the peak portion are periodically rippled every time the pixel advances. This is also because the peak value is stable at 200 to 255 density. It can be seen that the moire can be reduced as compared with the conventional linear interpolation method of 150 to 255 density.

If the present embodiment is configured as described above, it is possible to simultaneously improve moire reduction and image blur reduction.
Further, when the enlargement ratio is far from 100%, the importance of the interpolation position is increased by increasing the range of the projection area, so that the image can be enlarged smoothly.
On the other hand, when the reduction ratio is far from 100%, density blur can be reduced by reducing the range of the projection area.

  In the above embodiment, the image is described as monochrome, but it can also be applied to a color image. In the case of a color image, a scaled image can be obtained by performing the above-described interpolation operation for each color component.

  In addition, the present invention is not limited to the above-described embodiment, and various modifications and corrections can be made without departing from the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 ... Coordinate generation part, 2 ... Coordinate conversion part, 3 ... Interpolation calculating part, 4 ... Image data memory | storage part, 5 ... Variable magnification image data memory | storage part.

Claims (4)

The coordinate value of each pixel constituting the image after scaling is scaled to the coordinate value corresponding to the coordinate value of the pixel constituting the original image before scaling, and the coordinate value on the original image is calculated. In an image scaling generation device that generates a scaled image by calculating a pixel value of a scaled image by linear interpolation using pixel values of a surrounding original image surrounding the value,
A coordinate conversion unit that converts the coordinate value of the pixel to be interpolated on the original image into a coordinate value in a predetermined region in the periphery including the pixel closest to the pixel in the peripheral image surrounding the coordinate value And an interpolation calculation unit that calculates a pixel value of the pixel after scaling by linear interpolation based on the converted coordinate value.   The scaled image generation apparatus according to claim 1, wherein the coordinate conversion unit adjusts the predetermined area in accordance with a scale factor. The coordinate value of each pixel constituting the image after scaling is scaled to the coordinate value corresponding to the coordinate value of the pixel constituting the original image before scaling, and the coordinate value on the original image is calculated. In an image scaling generation method for generating a scaled image by calculating a pixel value of the scaled image by linear interpolation using the pixel values of the surrounding original image surrounding the value,
The coordinate value of the pixel to be interpolated on the original image is converted into a coordinate value in a predetermined region in the periphery including the pixel closest to the pixel in the peripheral image surrounding the coordinate value, and the conversion A scaled image generation method characterized by calculating a pixel value of a scaled pixel by linear interpolation based on the coordinate value obtained.   The scaled image generation method according to claim 3, wherein the predetermined area is adjusted in accordance with a scale factor.

Images