JPH10126612A - Pixel high speed interpolating system for enlarged picture and image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute an interpolation processing at high speed in spite of an enlargement ratio while resolution is kept by executing a single interpolation processing by means of increasing a pixel on a remaining hit map after a linear interpolation processing is executed with a power value. SOLUTION: The bit map of m×n, which is developed on a memory, is enlarged to R-times in the direction of X and to S-times in the direction of Y by a bit map enlargement means 11. When the power number U and/or V of '2' becoming R-α U>=0 and S-αV>=0 exists, an interpolation multiple calculation means 12 obtains the inverse string of the power value of '2' from '2' to U and V. A linear interpolation means 13 repeats linear interpolation in the direction of X in adjacent picture elements with the inverse of the power value of '2' from '2' to U. Then, linear interpolation is repeated in the direction of Y with the inverse of the power value of '2' from '2' to V and the new picture element is decided. When the bit map part which is not interpolated exists, the simple interpolation means 14 simply interpolates the part which is not interpolated with the respective pixels.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly, to a high-speed pixel interpolation method of an enlarged image on a bitmap for enlarging and printing an image.

[0002]

2. Description of the Related Art The size of an input image read into an image memory is usually 640 pixels × 480 pixels in a still image mode. If this image is directly printed out as it is, an output image having a size determined by the printing resolution is obtained. In the case of, if you want to output an image in the size of a hand or a business card, or output it to a size such as A4 or A3, or reduce or enlarge the image vertically or horizontally at a certain magnification, or enlarge it vertically It is also necessary to reduce the size in the horizontal direction.

In the case of enlarging an image, simply enlarging the distance between adjacent original pixels makes it impossible to obtain an image having a resolution capable of maintaining image quality. Therefore, the position (coordinates) of the original pixels is adjusted to a position corresponding to the enlargement magnification. , And rearranged on the bitmap. Then, new pixels are positioned between the original pixels, and the enlarged bitmap is interpolated with new pixels to obtain enlarged image data.

[0004] In this case, linear interpolation and simple interpolation are typical examples of pixel interpolation methods. [Linear interpolation method]
The linear interpolation method is a method of performing resolution conversion by linearly interpolating between two adjacent points A and B as shown in FIG. 7 to maintain a constant level of resolution.

In FIG. 7, two adjacent points are denoted by A (0, Ay),
Assuming that B (1, By), a newly inserted point P (Px,
Py) can be expressed as follows.

Py = (By−Ay) × Px + Ay (1) where Py is the gray level of point P for which Py is to be obtained, Ay is the gray level of a known point A (8-bit positive integer), and By is the known point B. (A positive integer of 8 bits), and Px is the distance of point P from point A. Also, 0 ≦ Px ≦ 1.

[0007] This is expanded two-dimensionally and known points are represented by A,
Assuming that B, C, and D are used and R is the insertion point, as shown in FIG. 8, Py = (By−Ay) × Px + Ay (2) Qy = (Dy−Cy) × Px + Cy (3) When a new point R is obtained from the two points P and Q, Ry = (Qy−Py) × Rx + Py (4), so that the equations (2) and (3) are substituted for Py and Qy, and Ay, By, When Cy and Dy are arranged, the point R
Can be expressed by the following equation (5).

Ry = (1-Rx) × (1-Px) × Ay
+ (1-Rx) × Px × By + Rx × (1-Px) × Cy
+ Rx × Px × Dy (5) Equation (5) is a general equation for two-dimensional linear interpolation.

[Simple interpolation method] In the simple interpolation method, as shown in FIG. 9, the positions of the original pixels A, B, C, and D after enlargement are obtained, and the magnifications are R and S (both are positive integers). In this method, the original pixels are copied in the X direction R-1 times, and after the copying in the X direction is completed, the original pixels are copied S-1 times in the Y direction to fill the interval.

First, the positions of the original pixels are obtained and the original pixels are arranged. Then, for each original pixel, R-1 original pixels are copied and filled in the X direction from the right side of the original pixel in order. After the interpolation in the X direction is completed,
S-1 lines are copied in the direction to fill the space between lines. The above operation is repeated for the subsequent pixels in the next row to interpolate the original pixels at a desired magnification R × S.

FIG. 9 is an explanatory diagram of the simple interpolation.
5 and S = 4.

As shown in FIG. 9A, the origin is defined as an original pixel A, and adjacent original pixels A (0, 0) and B (1,
0), C (0,1), and D (1,1) are first enlarged 5 times in the X direction and 4 times in the Y direction, and the positions of the original pixels A, B, C, and D after the enlargement are obtained. Are A (0,0), B (5,
0), C (0,4) and D (5,4) (FIG. 9)
(B)).

Therefore, the original pixels A, B, C and D are copied 5-1 = 4 times in the X direction to fill the interval, and when copying in the X direction is completed, 4-1 = 3 times in the Y direction. Copy a line. By this operation, as shown in FIG.
A double-magnified image is formed on the bitmap.

[0014]

In the above linear interpolation method, 2
Since the gradation of points between points can be obtained and the resolution of the enlarged image can be kept at a certain level, there is an advantage that the image quality can be compared with other methods. When performing the interpolation operation, the values of Rx and Px in the calculation formula are values of 1/2, 1/4, 1/8, ..., respectively (powers of 2 such as 2 times, 4 times, 8 times, etc.) In this case, the calculation result can be obtained by the shift operation, so that the calculation can be performed at a relatively high speed.
If the magnification is double, six times, seven times, nine times, ten times,..., Complicated division is required, and there is a problem that processing takes a long time.

The simple interpolation method is the fastest in comparison with other interpolation methods. However, since the gradation is not adjusted and only copying is performed, the lower the image quality and the larger the enlargement ratio, the better. There is a disadvantage that the quality of the enlarged image is rapidly reduced, and there is a problem that the degree of enlargement is limited depending on the resolution of the output device.

The present invention uses a high-speed pixel interpolation method for an enlarged image, which can stabilize the image quality by maintaining the resolution of the enlarged image at a certain level or more, and can perform interpolation processing at a high speed regardless of the magnification, and the present method. It is intended to provide an image processing device.

[0017]

In order to achieve the above object, a high-speed pixel interpolation method for an enlarged image according to the present invention uses an original pixel developed into an m × n bit map by R times in the X direction and Y directions. Is an interpolation method of an enlarged image of a bitmap enlarged by S times, and when there are powers of two U and V satisfying R-αU ≧ 0 and S-αV ≧ 0, two adjacent pixels are used. The linear interpolation is repeated in the X direction in the X direction with the reciprocal of the power from 2 to U in the Y direction with the reciprocal of the power of 2 from 2 to V, and a new pixel is determined. Is simply interpolated using the original pixels and the pixels obtained by the linear interpolation. In the embodiment, the coefficient α of the power of two U, V
Is preferably 2 or 4.

The image processing apparatus according to the present invention further comprises a control unit,
An image memory, a program storage unit, and a data transmission unit for transmitting data to a printer;
Bitmap enlarging means for further enlarging the bitmap of R in the X direction by R times and in the Y direction by S times, R-αU ≧ 0, S-αV
An interpolation multiple calculating means for obtaining a reciprocal sequence of powers of 2 from 2 to U and V when there are powers of 2 U and / or V satisfying ≧ 0; Linear interpolation is repeated between adjacent pixels in the X direction from 2 to U using the reciprocal of the power of 2 and, if a power V exists, the adjacent pixels are used as the reciprocal of the power of 2 from 2 to V in the adjacent pixels. Linear interpolation is repeated in the Y direction to determine a new pixel, and a new pixel is positioned in the bit map. X is calculated for each of the original pixel and the interpolation pixel obtained by the linear interpolation.
When there is a bitmap portion not interpolated by the linear interpolation means between pixels adjacent in the direction and the Y direction, simple interpolation means for simply interpolating the non-interpolated portion with the respective pixels, And a pixel high-speed interpolation means. In a preferred embodiment, the coefficient α of the power of two U, V is preferably 2 or 4.

[0019]

FIG. 2 is an example of an image processing apparatus to which a high-speed pixel interpolation method for an enlarged image according to the present invention can be applied.

The image processing apparatus 100 shown in FIG. 2 controls the entire apparatus and executes image processing by means of image processing stored in a program storage unit 3. The control unit 1 stores image data and necessary data. A memory 2 for storing, an operating system (hereinafter referred to as OS), an image processing application program, a printing color conversion processing program such as a printer driver, and a ROM such as a PROM or another magnetic medium for storing a pixel high-speed interpolation processing program of the present invention. A storage device 4 such as an FD or a magnetic disk for storing image data and other data and programs, a display device 5 such as a display, an input device 6, and image data subjected to printing processing. It has transmission means 7 for transmitting to the printer.

FIG. 1 is a block diagram showing the structure of the pixel high-speed interpolation means 10, FIG. 1A shows the relationship with other programs, and FIG. .

The high-speed pixel interpolation means 10 is stored in the program storage unit 3 together with the OS 20, the image processing application program 30, the printer driver 40 for performing the color conversion processing for printing, and the support program.

The OS 20 gives the control unit a control procedure such as an operation for realizing the image processing function of the image processing apparatus 100 and a screen display control.
From the application program 30 necessary for its execution
And the programs of the printer driver 40, the pixel high-speed interpolation means 10, etc., are taken out and executed and controlled by the control unit 1 (see the embodiment).

The pixel high-speed interpolation means 10 enlarges a given image at the enlargement magnification and develops it as a bit map on the memory 2 based on image position information, enlargement magnification and the like, and resolution and the like from the application program 30. And then interpolate the pixels into the enlarged image.

In this case, resolution conversion is performed by performing linear interpolation under a certain condition, and simple interpolation (padding of pixels) is performed by using original pixels of the enlarged image and linearly interpolated pixels for portions other than the conditions. Then, the interpolated image data (bitmap) is delivered to the printer driver 40.

Here, the linear interpolation is based on the linear interpolation method described in the description of FIGS. 7 and 8, and the simple interpolation is based on the simple interpolation method described in the description of FIG.

The printer driver 40 includes a rasterizer,
A color conversion module and a halftone module are provided as basic means for obtaining binary data required for drawing, and image data (bitmap) interpolated by a rasterizer is converted into R (red), G (green), B The image data is converted into three primary colors (blue), rasterized for each color, and developed into a predetermined work area of the memory 2 as RGB multi-gradation (for example, 256 gradations) bit image data. Rasterizer Performs a color conversion process on the raster-converted RGB multi-tone bit image data, and outputs a CM for printing.
It is converted to YK gradation bit image data.

Next, CMYK is performed by the halftone module.
The grayscale bit image data is distributed to each of the grayscale data by the grayscale ink table, halftone processing is performed, and the distribution or arrangement of the dark and light colors on the bitmap is determined for each color, and a binary bitmap is created for each color. . When the halftone processing is performed, the processing results are sequentially taken out of the memory 2 and transmitted to the printer via the transmission means 7.

As shown in FIG. 1B, the pixel high-speed interpolation means 10 includes a bitmap enlargement means 11, an interpolation multiple calculation means 12, a linear interpolation means 13, and a simple interpolation means 14. The bitmap enlarging means 11 converts the image on the mxn bitmap as shown in FIG. 3 (a) at given magnifications R and S by R times in the X direction and Y directions as shown in FIG. 3 (b). Is expanded by a factor of S to a predetermined area of the memory 2.

FIG. 3 is an explanatory diagram of the bitmap enlarging means. For the sake of explanation, it is assumed that R = 3, S = 5, and the original pixels A, B, C, D in FIG.

In FIG. 3A, the positions of the original pixels A, B, C, and D on the bitmap 31 are (1, 1), (2, 1),
(1, 2) and (2, 2), and when the original pixels are developed on the enlarged bitmap 32, the parts other than the original pixels (the parts surrounded by the line segment ACDBA) are The pixels become white pixels (or black pixels), and the resolution of the image is reduced.

The interpolation multiple calculation means 12 is a characteristic configuration of the present invention, and obtains a multiple for determining areas for linear interpolation and simple interpolation. If the multiple cannot be determined, simple interpolation by the simple interpolation means 14 is performed. The linear interpolation means 13 performs linear interpolation by the linear interpolation method described in FIGS. 7 and 8, and the simple interpolation means 14 performs simple interpolation by inflating pixels as described in FIG.

<Embodiment 1> FIG. 4 is an explanatory diagram of an interpolation multiple calculating means 12, a linear interpolating means 13, and a simple interpolating means 14. For the sake of explanation, a line segment (one-dimensional) is taken as an example.
The same applies to a plane (two-dimensional). FIG. 4A shows m
FIG. 4B shows an interpolation process using adjacent pixels A and B on a 16 × 16 bip map, and FIG. 4C shows an interpolation process using adjacent pixels A and B on a 16 × 16 bip map. 11x1
1 shows an interpolation process using adjacent pixels A and B on one bitmap.

Now, considering a case where a 16 × 16 magnification command is issued from the application program 30, first, as shown in FIG. 4B, １ ／ points are obtained, and then １ ／ points are obtained. Can be obtained, and further, 1/8 points can be obtained. By performing shift calculation by the linear interpolation means 13, new pixels 3, 5, 7, 9, 11, 1
3,15 can be interpolated and other new pixels 2,4,6,8,1
By performing simple interpolation on 0, 12, 14, and 16, all the points 2 to 16 other than the pixels A and B can be interpolated with new pixels in a short time.

In order to obtain the values 1/2, 1/4, 1/8,... For the enlarged bitmap R × S by the above method, the enlargement magnification is R, S, and the power of 2 is U, V. R-2U ≧ 0 in the X direction and S-2U in the Y direction
U and V satisfying ≧ 0 are obtained, and in the X direction, a sequence of powers of 2 from 2 to U may be obtained, and their reciprocals may be obtained. If the maximum power of 2 (8 in the example of FIG. 4B) is obtained, the power of 2 series 2, 4, 8 and the reciprocal thereof can be easily obtained (may be obtained by calculation or the correspondence table). May be registered).

Next, in the case of 11.times.11 times, if all points are interpolated by the linear interpolation method, a complicated calculation is required and a long time is required. Therefore, as shown in FIG. 4C, when the maximum value of the power of 2 closest to the magnification 11 is obtained,
Since 2U ≧ 0 to U = 4, a sequence of powers 2 and 4 of 2 in the range is obtained, and linear interpolation processing by the linear interpolation means 13 is performed between points 1 to 9 at ２ ， and ４. The other parts are subjected to interpolation processing (pixel inflating) by the simple interpolation means 14. Note that the pixel P (point 9) is interpolated by the linear interpolation processing of the pixels A and B.

Next, for the remaining points 10 and 11, the simple interpolation means 14 copies the pixel at point 9 in the pixel B direction (X direction).

As is apparent from the above description, R-2U = 0
In this case, simple interpolation is performed.

Although only the X direction has been described in the above description, the interpolation process can be performed in the Y direction in the same manner as in the X direction, by regarding the pixels obtained by the interpolation in the X direction as the original pixels. Note that the interpolation process in the Y direction is performed first,
The interpolation processing in the X direction may be performed later.

<Embodiment 2> FIG. 4 shows an embodiment in which as many points as possible are interpolated by linear interpolation and interpolation by simple interpolation is reduced as much as possible.

However, in order to improve the processing speed, it is desirable to reduce the number of linear interpolations and increase the number of simple interpolations.

FIG. 5 is an explanatory diagram of the interpolation multiple calculating means 12, the linear interpolating means 13 and the simple interpolating means 14 when the number of linear interpolations is reduced and simple interpolation is performed between pixels. Although (one-dimensional) is taken as an example, the same applies to a plane (two-dimensional). FIG. 5A shows adjacent pixels A and B on an m × n bipmap, and FIG. 5B shows an interpolation process using adjacent pixels A and B on a 16 × 16 bipmap. FIG. 5C shows an interpolation process using adjacent pixels A and B on an 11 × 11 VIP map.

Now, assuming that a 16 × 16 magnification command is issued from the application program 40, first, as shown in FIG. 4B, の points are obtained, and then 各 points are obtained. , And 1/8 of each point can be obtained. In this case, the number of linear interpolations (1+
(2 + 4 = 7 times), which increases the processing time. Therefore, in the present embodiment, the calculation of the 1/8 point is omitted. Thereby, the number of times of linear interpolation processing can be reduced (1 + 2 =
3 times).

For this purpose, the maximum value of the power of 2 that satisfies R−4U ≧ 0 is obtained by the interpolation multiple calculating means 12. As a specific example, in the example of FIG. 4B, the maximum power of 2 is 8, but in the present embodiment, U = 4 is obtained from 16-4U ≧ 0 and used as the maximum value of the power of 2.

Thus, the power series of 2 in the X direction is 2,
4 and the linear interpolation means 13 obtains an interpolated pixel by performing ２ ， and ４ linear interpolation processing between adjacent pixels A and B, and obtains pixels other than the pixels A and B which are interpolated by linear interpolation. By performing the interpolation processing by the simple interpolation means 14, it is possible to secure the resolution and to speed up the interpolation time.

In FIG. 5B, an arrow means that a simple interpolation process is performed between them. For example, it is indicated that the pixel A is copied at the points 2, 3, and 4, and the pixel 5 is copied at the points 6, 7, and 8.

Next, FIG.
This is similar to the case of FIG. 5C, and as shown in FIG.
First, the interpolation multiplier calculating unit 12 obtains the maximum value 2 of the power of 2 closest to the magnification 11 from 11-4U ≧ 0, obtains a power-of-two sequence 2 in the range, and obtains points 1 to 1 in FIG. The pixel 5 is interpolated by performing the linear interpolation processing by the linear interpolation means 13 by に つ い て in the interval of 9 and the other pixels (2, 3,
For (4, 6, 7, 8), the interpolation processing by the simple interpolation means 14 can be performed to secure the resolution and to speed up the interpolation time.

In this embodiment, the interpolation multiple calculating means 1
2, the maximum value of the power of 2 closest to the magnification is calculated as R-4U ≧ 0, and then the maximum value of the power of 2 smaller than that is obtained. However, the present invention is not limited to this, and R-8U ≧ 0, R-
It is also possible to adopt a configuration in which 16U ≧ 0 and a power-of-two value that satisfies this expression is obtained. Here, how far the maximum value of the power-of-two sequence can be changed, that is, whether the number of times of linear interpolation processing by the linear interpolation means can be reduced depends on how far the range in which simple interpolation can be repeated can be expanded. It is defined by the overall device resolution. The case where the repetition of the simple interpolation is about 3 to 4 corresponds to the present embodiment.

Although only the X direction has been described in the above description, the interpolation process can be performed in the Y direction in the same manner as in the X direction, by regarding pixels obtained by interpolation in the X direction as original pixels. Note that the interpolation process in the Y direction is performed first,
The interpolation processing in the X direction may be performed later.

Further, in the first and second embodiments, the interpolation multiple calculating means 12 is described as a calculating means (program). However, the present invention is not limited to this.

Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment, and it goes without saying that various modifications can be made.

[0052]

FIG. 6 is a diagram showing an example of the high-speed interpolation processing. The method of the first embodiment, that is, R-4U ≧
Example in which the maximum value of the power of two that satisfies 0, S−4V ≧ 0 is calculated for an enlargement factor of 8 × 8, and is doubled by linear interpolation and quadrupled by simple interpolation to obtain a total magnification of eight times. It is.

In FIG. 6, A, B, C, and D are the original pixels.
It has been multiplied to 64 pixels.

In this example, first, the interpolation multiple calculating means 12
Then, the interpolation multiple is reduced to 1/2, then the three points A, A, and C in the figure are calculated by the linear interpolation means 13, and then the original pixel A and the three points A, A, and C are calculated by the simple interpolation means 14. Four of the interpolation pixels
Interpolate between adjacent original pixels A, B, D, and C by copying four times in the lower right direction based on the pixel (actually, first copy in the X direction and then copy in the Y direction). The processing is performed to obtain the result of FIG.

Next, the same processing is repeated between the original pixel B and the adjacent original pixel, and the same processing is repeated for the original pixels C and D, thereby interpolating the enlarged pixel on the bit map. it can.

[0056]

As described above, according to the present invention, after the pixels are interpolated by a linear interpolation process with respect to the magnified image on the bitmap by a power of 2, the remaining pixels on the bitmap are inflated. Since the simple interpolation process is performed, the image quality can be maintained while maintaining the resolution of the enlarged image at a certain level or higher, and high-speed pixel interpolation can be performed regardless of the magnification.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a pixel high-speed interpolation unit.

FIG. 2 is an example of an image processing apparatus to which a high-speed pixel interpolation method for an enlarged image according to the present invention can be applied.

FIG. 3 is an explanatory diagram of a bitmap enlarging unit.

FIG. 4 is an explanatory diagram of pixel interpolation processing by an interpolation multiple calculation unit, a linear interpolation unit, and a simple interpolation unit;

FIG. 5 is an explanatory diagram of high-speed pixel interpolation processing by an interpolation multiple calculation unit, a linear interpolation unit, and a simple interpolation unit.

FIG. 6 is an example of high-speed pixel interpolation processing.

FIG. 7 is an explanatory diagram of a linear interpolation method.

FIG. 8 is an explanatory diagram of a linear interpolation method.

FIG. 9 is an explanatory diagram of a simple interpolation method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Control part 2 Memory 3 Program storage part 6 Input means 7 Data transmission means 10 Pixel high-speed interpolation means 11 Bit map enlargement means 12 Interpolation multiple calculation means 13 Linear interpolation means 14 Simple interpolation means 31, 32 Bit map

──────────────────────────────────────────────────の Continuation of the front page (51) Int.Cl. ⁶ Identification code FI G09G 5/36 520 B41J 3/00 A H04N 1/21 3/12 L 1/387 101 G06F 15/66 355C (72) Inventor Takeo Endo 3-5-5 Yamato, Suwa City, Nagano Prefecture Seiko Epson Corporation (72) Inventor Yuji Matsueda 3-3-5 Yamato Suwa City, Nagano Prefecture Inside Seiko Epson Corporation (72) Kazuya Horii Inventor Kazuya Horii Nagano Prefecture 3-5-5 Yamato, Suwa City Inside Seiko Epson Corporation

Claims (5)

[Claims] 1. An interpolation method for an enlarged image of a bitmap in which original pixels developed into an m × n bitmap are enlarged R times in an X direction and S times in a Y direction, wherein R−αU ≧ 0, S -If there are powers of 2 U and V satisfying αV ≧ 0, the power of 2 from 2 to V in the X direction is the reciprocal of the power of 2 from 2 to U between adjacent pixels. Iteratively repeats linear interpolation in the Y direction with the reciprocal of to determine new pixels, and performs simple interpolation using the original pixels and the pixels obtained by the above linear interpolation for the portions on the bitmap that cannot be interpolated by these linear interpolations. Pixel high-speed interpolation method for enlarged images. 2. A high-speed pixel interpolation method for an enlarged image according to claim 1, wherein the coefficient α of the power of two U, V is 2 or 4.
A high-speed pixel interpolation method for an enlarged image. 3. A control unit, an image memory, a program storage unit, and a means for transmitting data to a printer, wherein the m × n bit map developed on the memory is further converted to X
Bitmap enlarging means for enlarging R times in the direction and S times in the Y direction; and when there are powers of two U and / or V satisfying R−αU ≧ 0 and S−αV ≧ 0, 2 to U, An interpolation multiple calculating means for obtaining a power of 2 up to V; and if a power U is present, adjacent pixels repeat linear interpolation in the X direction with a reciprocal of a power of 2 from 2 to U; When the number V exists, two pixels are used between adjacent pixels.
Linear interpolation is repeated in the Y direction with the reciprocal of a power of 2 from V to V to determine a new pixel and position it in the bitmap; and for each of the original pixel and the interpolation pixel obtained by the linear interpolation When there is a bitmap portion that is not interpolated by the linear interpolation unit between pixels adjacent in the X direction and the Y direction, a simple interpolation unit that simply interpolates the non-interpolated portion with the respective pixels. An image processing apparatus comprising: a pixel high-speed interpolation unit comprising: 4. The image processing apparatus according to claim 3, wherein the coefficient α of the powers of two U, V is 2 or 4. 5. A storage medium storing the pixel high-speed processing means according to claim 3, wherein the control operation of the control unit of the image processing apparatus according to claim 3 functions.