JP3483426B2 - Image reduction method and storage medium storing program - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reduction method for reducing a digital original image to obtain a reduced image and a storage medium recording a program for causing a computer to execute the method.

[0002]

2. Description of the Related Art In the conventional image reduction of this type, as shown in FIG. 17, one pixel row (pixel row 1, pixel row 2, ...) Of an original image OG arranged in a desired reduction direction (X direction in the drawing). , Each pixel row Z) is reduced at a desired reduction ratio n / m (n and m are positive numbers of n <m) and the original image OG is reduced in a one-dimensional reduction direction to obtain a reduced image SG. Trying to get.

In this type of image reduction, the original image OG is usually used.
The processing range SH can be set in the reduction direction of. This is to prevent the unnecessary pixels from being reflected in the reduced image SG when there is an unnecessary portion (unnecessary pixel) unrelated to the image at the end of the pixel row in the reduction direction of the original image OG. . Although FIG. 17 shows the image reduction in the case where the range sandwiched by the dotted lines is the processing range SH, if there is no unnecessary pixel in the original image OG, the entire pixel row in the reduction direction is processed as the processing range SH.

Conventionally, the image reduction for each one pixel example is performed by a weighted average of adjacent pixels by a weight assigned according to the position of the pixel in the pixel row of the original image, as shown in FIG. In FIG. 18, the reduction ratio n / m is set to 2/7. Moreover, each code | symbol in a figure is as follows.

S is a pixel number (s = 1, 2, ..., S _(MAX ) of pixels forming one pixel row (pixel row 1, pixel row 2, ..., Pixel row Z) to be processed in the original image OG. ₎ : S _(MAX) is not shown in the figure, but the number of pixels in the one pixel column to be processed in the original image OG), P (s) is each pixel constituting one pixel column to be processed in the original image OG Density data, t is the image S after reduction
The pixel number (t = 1, 2, ... t _(MAX) : t of the pixels forming one pixel row of G (pixel row 1, pixel row 2, ..., Pixel row Z)
_(MAX) is not shown, but the number of pixels in one pixel column of the reduced image SG), Q (t) is density data of each pixel forming one pixel column of the reduced image SG, s (T) is the pixel number of the pixel of the original image OG that is referred to when calculating Q (t) among the pixels that form one pixel column of the processing target of the original image OG, and s (1) is the original The pixel No and ah of the first pixel of the processing range SH in the one pixel column of the processing target of the image OG are P
Load correction amount of (s (1)) (ah = 0, 1, ..., (n−
1)) and R (t) are calculation target regions of the original image OG used for calculating the Q (t), and are within each pixel of the original image OG used for the load calculation for obtaining the Q (t). , G (t) is a calculation target pixel section of the original image OG used to calculate the Q (t), and is an original image used in the load calculation for obtaining the Q (t). A section showing a range of a plurality of adjacent pixels of OG.

The conventional method can be expressed by the following equation. Q (t) = ((n−a (t)) / m) × P (s (t)) + Σ {(n / m) × P (i)} [i = s (t) +1 to s (t + 1) ) −1] + (a (t + 1) / m) × P (s (t + 1)) (1) However, s (t) = s (1) + int [(ah + (m × (t−
1))) / n] a (t) = (ah + (mx (t-1))) mod n

In the above equation, int [A /
B] is the integer part of the result of dividing A by B in the integerizing function, A
mod B is a function for obtaining a remainder when A is divided by B.

In the processing, t = 1, 2, 3, ... And t are sequentially counted up, and Q (1), Q (2), Q (3) ,.
The density data of each pixel forming one pixel column of the reduced image SG is obtained in this order, and the pixel data is reduced by one pixel column. Each pixel column (pixel column 1, pixel column 2, ...) Arranged in parallel in the direction orthogonal to the reduction direction (Y direction in FIG. 17).
This is performed for each pixel row Z) to reduce the original image OG in the reduction direction at the reduction rate n / m.

As is clear from FIG. 18 and the above equation,
In this conventional method, each pixel forming one pixel column to be processed of the original image OG is conveniently divided into n pieces, and 1 /
The load of m is assigned, the maximum load of one pixel of the original image OG is set to n / m, and the total load is m / m (100%) starting from the (ah + 1) th section of the s (1) th pixel. R (t) is determined so that
Density data of one pixel after assigning a load according to the number of divisions of R (t) to a plurality of adjacent pixels (each pixel forming G (t)) in the pixel row and reducing the weighted average result by the load I am trying. Then, each R (t) for each Q (t) is continuously determined, and the density data of each pixel in one pixel column of the reduced image SG is sequentially obtained.

According to this conventional method, as is clear from the arithmetic expression for determining s (t), for example, the reduced image S
The pixel [s (k)] in the processing target pixel row of the original image OG referred to in calculating the density data [Q (k)] of the k-th pixel in the G processing target pixel row is s (1 ), The pixel becomes an approximately [(m / n) × k] -th pixel. Therefore, as the reference pixel in the original image OG for each pixel of the reduced image SG, a pixel at an appropriate position is selected according to the reduction ratio. Then, since the density data of each pixel of the reduced image SG is obtained by the weighted average of the adjacent pixels in the vicinity of each reference pixel, the entire processing range SH of the original image OG in the reduction direction is n / m.
It is possible to reduce the size approximately twice.

Further, when reducing a two-dimensional digital original image in each of the two-dimensional directions (X direction and Y direction), as shown in FIG. 19, the original image OG is reduced in the X direction by the above method. Then, the original image OG is reduced to 2 by reducing the image SG1 after the reduction operation in the Y direction by the above method.
An image SG2 reduced in the dimension is obtained.

[0012]

However, such a conventional image reduction method has the following problems. Referring to FIG. In FIG. 20A, binarization is performed in black and white (pixels indicated by diagonal lines are black) (here, the density data of black pixels is “0.00”, and the density data of white pixels is “1.00”). FIG. 20B shows the original image OG that has been processed.
The original image OG of 0 (a) is reduced by 2/7 in the XY direction by the conventional method, the density ratio of each pixel of the reduced image SG is reduced to s (1) = 1 and ah = 0. .

Letting P (s _X , s _Y ) be the density data of each pixel of the two-dimensional original image OG and Q (t _X , t _Y ) be the density data of each pixel of the reduced two-dimensional image. For example, Q
(1, 3) is the calculation target pixel section G _X (1) in the X direction and the calculation target pixel section G _Y (3) in the Y direction in FIG.
For each pixel included in a two-dimensional calculation target pixel section surrounded by and, a two-dimensional area surrounded by a calculation target area R _X (1) in the X direction and a calculation target area R _Y (3) in the Y direction It is a weighted average of the loads assigned according to the calculation target region, and is calculated by the following calculation.

[0014]   Q (1, 3) = 2/7 × (2/7 × P (1,  8) + 2/7 × P (2,  8) + 2/7 × P (3,  8) + 1/7 × P (4,  8))         + 2/7 × (2/7 × P (1,  9) + 2/7 × P (2,  9) + 2/7 × P (3,  9) + 1/7 × P (4,  9))         + 2/7 × (2/7 × P (1, 10) + 2/7 × P (2, 10) + 2/7 × P (3, 10) + 1/7 × P (4, Ten))         +1/7 × (2/7 × P (1, 11) + 2/7 × P (2, 11) + 2/7 × P (3, 11) + 1/7 × P (4, 11))         = 2/7 x (2/7 x 1. 00 + 2/7 × 1. 00 + 2/7 × 0. 00 + 1/7 × 0. 00)         +2/7 x (2/7 x 1. 00 + 2/7 × 1. 00 + 2/7 × 0. 00 + 1/7 × 0. 00)         +2/7 x (2/7 x 1. 00 + 2/7 × 0. 00 + 2/7 × 0. 00 + 1/7 × 0. 00)         +1/7 × (2/7 × 1. 00 + 2/7 × 0. 00 + 2/7 × 0. 00 + 1/7 × 0. 00)         = 22/49 ≒ 0. 45

The density data of the other pixels of the reduced image SG are calculated in the same manner. As is clear from FIG. 20, when the density data of the pixel of the reduced image SG is obtained by the weighted average by the load assigned according to the position of the pixel in the original image OG as in the conventional method, for example, Q ( 1, 3),
Like Q (2,1), Q (2,3), Q (2,4),
The original image OG is calculated so that the inside of the two-dimensional calculation target area of the original image OG for calculating the density data of these pixels is roughly divided into two.
When the black-and-white edge passes, the density data of the pixel of the reduced image SG has an intermediate value (gray). Also, Q
There are also cases where the density data decreases (becomes a dark gray) such as (1,4) and Q (2,2). On the other hand, Q
If the black and white edges of the original image OG do not pass within the two-dimensional calculation target area of the original image OG used for calculating the density data of these pixels, such as (3,1) and Q (4,1), the density data May not decrease. In this way, the original image OG
There is a problem that the image quality of the reduced image SG deteriorates because the image is blurred or not blurred according to the passing state of the black and white edge.

Another problem of the conventional method will be described with reference to FIG. In FIG. 21, the reduction ratio n / m is, for example, 9 /
FIG. 10 is a diagram schematically showing a result of image reduction performed by the above-described conventional method when the value is close to “1.0” as in 10.

As is clear from the figure, the weighted average ratio of the adjacent pixels of P (s) to each Q (t) is 9: 1,
The change of 8: 2, ..., 5: 5, ..., 2: 8, 1: 9 will be cyclically repeated.

In this case, although the reduction ratio is close to "1.0", for example, two adjacent pixels of the original image OG are 5: 5 (= 1: 1) as shown in Q (5) of FIG. ) Will be averaged at the same rate. If there is an edge in the original image OG, for example, as shown in Q (1) of FIG. 21, in the pixel portion that is substantially equal to the density data of one pixel of the original image OG, the edge is saved even after reduction and the image is clear, In the pixel portion that averages two adjacent pixels at an equal ratio of 1: 1 as described above, the edge is reduced and the image is blurred because it is an equal average value of two pixels.

Conventionally, the reduction processing for each pixel column is performed by s
(1) Since ah is processed with the same value, the blur of the image appears in the same pattern for each pixel column in the reduction direction, and is orthogonal to the reduction direction as shown by the dotted line in FIG. The linear unevenness Gm appears in the reduced image SG after the reduction, and the quality of the reduced image SG is deteriorated. Particularly, when the reduction ratio is close to “1.0”, the appearance period mT of the linear unevenness Gm that periodically appears in the reduced image SG.
Becomes wider, and this linear unevenness Gm becomes noticeable even by visual observation, which is a problem.

Further, when the two-dimensional digital image is reduced in both the X and Y directions by the conventional method, for the same reason as shown in FIG.
As shown in (b), grid-like unevenness appears.

The present invention has been made in view of the above circumstances, and provides an image reduction method capable of reducing an image without degrading the image quality and a program for causing a computer to execute the method. It is an object to provide a recorded storage medium.

[0022]

The present invention has the following constitution in order to achieve such an object. That is, the invention according to claim 1 provides an image reduction method for reducing an original digital image to obtain an image after reduction, (1)
A step of determining the calculation target pixel section of the original image used for calculating the density data of the calculation target pixel of the reduced image according to the reduction ratio of the image; and (2) the original determination in the step (1). The weight is assigned to each pixel in the calculation target pixel section of the original image according to the density data of each pixel in the calculation target pixel section of the image, and (3) the load is allocated in the step (2). The weighted average of each pixel in the calculation target pixel section of the original image determined in the step (1) by the weight is used as the density data of the calculation target pixel of the reduced image. In the step (2), the reduced image obtained by reducing the original image is obtained by obtaining the density data of each pixel of the image after the reduction , and in the step (2), the process of (1) is performed.
The density of each pixel in the calculation target pixel section of the original image
When arranging in order of magnitude of data, the specified number of
Pixel weight is set to "0%", and other multiple pixels are set.
It is characterized by allocating weights .

[0023]

[0024]

The invention described in claim 2 is the above-mentioned claim 1.
According to the image reduction method described in (1), the two-dimensional digital original image is simultaneously reduced in the two-dimensional direction.

According to a third aspect of the present invention, the image reduction processing is performed by the image reduction method according to the first aspect for each pixel row of the original image in which pixels are arranged in a one-dimensional reduction direction, and the original image is obtained. Is reduced in the reduction direction.

According to a fourth aspect of the invention, there is provided a storage medium recording a program for causing a computer to execute a process of reducing a digital original image to obtain a reduced image. A step of determining a calculation target pixel section of the original image used for calculating the density data of the pixel of the image calculation target according to the reduction ratio of the image, and (2) calculation of the original image determined in the step (1). According to the magnitude order of the density data of each pixel in the target pixel section, a step of assigning a load to each pixel in the calculation target pixel section of the original image, and (3) the load assigned in the step (2), Reducing the result of weighted averaging of each pixel in the calculation target pixel section of the original image determined in the step (1) as density data of the calculation target pixel of the reduced image. Density data for each pixel in the later image The original image is reduced to obtain an image after reduction , and in the step (2),
Within the calculation target pixel section of the original image determined in step (1)
When each pixel of is arranged in order of magnitude of density data,
The weight of a predetermined number of pixels lined up is set to "0%", and
A program for causing a computer to execute a process of assigning a load to a plurality of pixels is recorded in a storage medium.

[0028]

The operation of the invention described in claim 1 is as follows. The density data of each pixel of the reduced image is shown in (1) below.
It is sequentially determined in steps (3) to (3).

In the step (1), the calculation target pixel section of the original image used for calculating the density data of the calculation target pixel of the reduced image is determined according to the reduction ratio of the image. The method of determining the calculation target pixel section of the original image can be determined, for example, by the same calculation as in the conventional art. For example, when the image reduction in the one-dimensional reduction direction is performed for each pixel row of the original image in which pixels are arranged in the reduction direction as in the invention described in claim 3 , the formula (1) of the conventional method is used. The calculation target pixel section can be determined based on Further, when the image is reduced in the two-dimensional direction at the same time as in the invention described in claim 2 , the two-dimensional section surrounded by the calculation target pixel section in each direction determined based on the formula (1) of the conventional method is The calculation target pixel section of the original image determined in the step (1) may be used.

In the step (2), each pixel in the calculation target pixel section of the original image is determined according to the density data of each pixel in the calculation target pixel section of the original image determined in the above step (1). In step (3), the weight is averaged for each pixel in the calculation target pixel section of the original image determined in step (1) above according to the load assigned in step (2) above. , The density data of the pixel to be calculated in the reduced image.

That is, as in the conventional method, the density data of the pixel to be calculated in the image after being reduced by the weighted average by the weight assigned according to the position in the original image of each pixel in the pixel to be calculated in the original image is calculated. Instead of obtaining the density data, the density data of the pixel to be calculated in the image after reduction is obtained by the weighted average by the weights assigned according to the magnitude data of the density data of each pixel in the calculation target pixel section of the original image.

By thus reducing the image, regardless of the position of each pixel in the calculation target pixel section of the original image, the image is reduced by the weighted average according to the distribution state of the density data in the calculation target pixel section of the original image. The density data of each pixel of the subsequent image can be obtained, and the distribution state of the density data of each pixel in the calculation target pixel section of the original image can be reflected in the reduced image, resulting in deterioration of image quality due to image reduction. Can be prevented.

As a method of allocating the load according to the magnitude order of the density data of each pixel in the calculation target pixel section of the original image,
For example, as in the invention described in claims 2 and 3, when each pixel in the calculation target pixel section of the original image is arranged in order of magnitude of density data, the load of one or a plurality of pixels that come to the center (nearby) It is preferable to allocate so as to increase.

[0034]

According to the first aspect of the invention, when the pixels in the calculation target pixel section of the original image determined in the step (1) are arranged in the descending order of the density data (the density data is The weight of a predetermined number of pixels arranged on the large side and the small side is set as “0%” and excluded from the weighted average, and a plurality of other pixels, that is, the pixels when the pixels are arranged in descending order of density data A weight is assigned to a plurality of pixels (pixel groups near the median value) in the vicinity, and the density data of each pixel of the reduced image is obtained by the weighted average of the pixel groups near the median value. By doing so, since the weighted average is obtained in the pixel group near the median value, it is possible to prevent the density data of the pixels of the image after reduction from being blurred or lowered, and on the other hand, the noise component may be present in the median pixel. Even if it is included, the noise component can be reduced. Furthermore, when the pixels are arranged in order of magnitude of density data, a predetermined number of pixels arranged on both ends are excluded to perform weighted averaging.
It is also possible to reduce the image by removing the influence of the noise component contained in the pixels having large or small density data.

According to the second aspect of the invention, the image reduction in the two-dimensional direction of the two-dimensional digital original image can be performed simultaneously and suitably.

According to the third aspect of the invention, similar to the conventional method, image reduction processing is performed for each pixel row of the original image in which pixels are arranged in the one-dimensional reduction direction, and the original image is reduced in the reduction direction. Can be suitably performed.

According to the invention described in claim 4 , by causing the computer to read the program recorded in the storage medium, the computer executes the image reduction processing according to the invention described in claim 1.

[0039]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an example of a hardware configuration for implementing an image reduction method according to the present invention.

FIG. 1 is composed of a computer system. The CPU 10 that executes the image reduction process for reducing the digital original image to obtain the reduced image is the bus line 20.
Is connected to an internal memory 30 configured by a RAM or the like. In this internal memory 30, areas such as a program memory 31 into which a program is ported when executing image reduction processing, an original image memory 32 for storing an original image, and a reduced image memory 33 for storing a reduced image are included. Are set respectively. The CPU 10 also controls the driver 4 for the external storage device via the input / output interface 40.
1, an original image input device 42, an output device 43, a setting device 44, etc.

An external storage device 45 is loaded in the driver 41. An image reduction processing program created in advance is stored in the external storage device 45. External storage device 4
Reference numeral 5 comprises a magneto-optical disk, a CD-ROM, a floppy disk, a hard disk, a magnetic tape and other storage media. When the power of the apparatus is turned on and the image reduction processing according to the present embodiment is activated, the image reduction processing program is read from the external storage device 45 and the program memory 3 is read.
1 and the CPU 10 executes the image reduction processing described later according to the image reduction processing program. The external storage device 45 in which this image reduction processing program is stored is
It corresponds to the storage medium recording the program in the present invention.

The original image input device 42 is a device for inputting an original image, and is, for example, a digital camera, an input scanner, an image capturing device using a line sensor such as a CCD disclosed in Japanese Patent Laid-Open No. 16367/1990. Composed of. The original image input by the original image input device 42 is stored in the original image memory 32, and the original image is subjected to image reduction processing described later, and the reduced image is the reduced memory 33.
Memorized in. It should be noted that the previously input original image is stored in an external storage device, the original image is read from the external storage device, stored in the original image memory 32, and the image reduction processing is performed on the original image. You may comprise.

The original image memory 32, as shown in FIG.
The reduced-size image memory 33 is configured to store a two-dimensional digital original image, as shown in FIG.
It is configured to store a two-dimensional digital reduced image.

The output device 43 is for outputting the reduced image generated in the reduced image memory 33, and is composed of, for example, a display device, a recording scanner, a printer and the like. The reduced image may be stored in an external storage device.

The setting device 44 is for making various settings, designations and instructions, and is composed of a keyboard, a mouse, a switch, a setting monitor and the like.

The first image reduction processing with the above configuration will be described below with reference to the flowchart shown in FIG. Note that FIG. 4 is a flow chart in the case of simultaneously performing image reduction in a two-dimensional direction of a two-dimensional digital original image. Also, the original image is stored in the original image memory 32 when the processing is started.
Be stored in.

In the following description, s, P (s), t,
Q (t), s (t), s (1), ah, R (t), G
(T) basically has the same meaning as in the conventional method, in particular, X
When it is necessary to distinguish the direction and the Y direction, each code is attached with a subscript indicating the direction (X direction [s _X , t _X , s _X (t
_X ), s _X (1), ah _X , R _X (t _X ), G
_X (t _X )], Y direction [s _Y , t _Y , s _Y (t _Y ), s
_Y (1), ah _Y , R _Y (t _Y ), G _Y (t _Y )]),
When each pixel of the original image is shown in two dimensions, two-dimensional coordinates (s
_X , s _Y ), when the density data of each pixel of the original image is shown in two dimensions, it is expressed by two-dimensional notation (P (s _X , s _Y )),
Similarly, in the reduced image, when each pixel is shown in two dimensions, two-dimensional coordinates (t _X , t _Y ) are used and the density data is set to two.
When it is indicated by dimension, it is expressed by two-dimensional notation (Q (t _X , t _Y )). Further, as shown in FIG. 5, R _X (t _X ) and R _X
Let R be the two-dimensional calculation target area surrounded by _Y (t _Y ).
_{Represented by XY} (t _X , t _Y ), G _X (t _X ) and G _Y (t _Y )
A two-dimensional calculation target pixel section surrounded by and is G _XY (t _X ,
t _Y ).

Step S1 of FIG. 4 (corresponding to the step (1) of the present invention): an original used to calculate the density data (Q (t _X , t _Y )) of the pixel to be calculated in the reduced image. Calculation target pixel section of image (two-dimensional) (G _XY (t _X , t _Y ))
Is determined according to the reduction ratio (n / m) of the image.

This calculation target pixel section G_XY(T_X, T_Y)
Can be determined based on the conventional equation (1), for example.
It That is, the calculation target pixel section G in the X direction
_X(T_X), Calculation target pixel section G in the Y direction_Y(T_Y)
Each of these G is decided based on the equation (1)._X(T_X)
And G_Y(T_Y) Two-dimensional calculation target pixel area surrounded by and
To G_XY(T_X, T_Y). Calculation target pixel in X direction
Section G_X(T_X) Is s in the X direction of the original image_X(T
_X) Th pixel to s_X(T_XUp to the +1) th pixel
It becomes a section including pixels. However, na_X(T_X) = 0
In the case of (when the coefficient of the first term on the right side of equation (1) is “0”)
In the X direction of the original image (s)_X(T_X) +1)
S from the th pixel_X(T_XPixels up to +1) th pixel
It will be a section containing. Similarly, the calculation target pixel section in the Y direction
G_Y(T_Y) Is s in the Y direction of the original image _Y(T_Y)
S from the th pixel_Y(T_YPixels up to +1) th pixel
(Including n-a_Y(T_Y) = 0
About the Y direction of the original image (s_Y(T_Y) +1) th image
From s_Y(T_YSection including pixels up to +1) th pixel
Image).

Note that a _X (t _X ) = (ah _X + (m ×
(T _X −1))) mod n, and a _Y (t _Y ) =
(Ah _Y + (m × (t _Y −1))) mod n.

Further, ah _X and ah _Y are uniquely set to “0” (a _X (t _X ) = (m × (t _X −1)) mod.
n, a _Y (t _Y ) = (m × (t _Y −1)) mod
n) determines G _X (t _X ) and G _Y (t _Y ) and G
_XY (t _X , t _Y ) may be determined.

For example, in the case of FIG.
_XY (1,1) is (1 to 4,1 to 4) of the original image, that is, (1,1), (2,1), (3,1), (4,
1), (1, 2), (2, 2), (3, 2), (4
2), (1,3), (2,3), (3,3), (4
3), (1,4), (2,4), (3,4), (4
4) 16 pixels section, G _XY (2,1) is the original image section of (4 to 7,1 to 4) 16 pixels, G _XY (3,1) is the original image , (8 to 11, 1 to 4) consisting of 16 pixels, ..., G _XY (1,2) is a section of (1 to 4, 4 to 7) consisting of 16 pixels of the original image,
Is ...

In this way, the calculation target pixel section G _XY (t _X ,
t _Y ), and the density data Q (t) of the pixel to be calculated in the reduced image is calculated by using the pixels in G _XY (t _X , t _Y ).
_X , t _Y ), the original image OG is surrounded by the processing ranges SH in the X and Y directions, respectively, as shown in FIG.
It is possible to obtain the reduced image SG in which the entire dimensional processing range (the range surrounded by the dotted line) is reduced substantially evenly by n / m times. If there are no unnecessary pixels in the X direction and / or Y direction of the original image OG, the original image OG
The entire processing range SH is processed as the processing range SH. The processing range SH in the X direction (Y direction) is s _X (1), ah _X (s _Y
(1), ah _Y ), etc., but the pixel number (s) of the last pixel of the processing range SH in the X direction (Y direction)
_It may be configured to set _{X (LIM)} (s _{Y (LIM)} ).

The reduction ratio (n / m), s _X (1), a
Necessary data such as h _X , s _Y (1), a h _Y (s _{X (LI M)} , s _{Y (LIM)} ) are set by the setter 44 or the like prior to the process of step S1.

Step S2 in FIG. 4 (corresponding to the step (2) of the present invention): density data of each pixel in the calculation target pixel section G _XY (t _X , t _Y ) of the original image determined in the above step S1 Of the original image calculation target pixel section G _XY
A weight is assigned to each pixel in (t _X , t _Y ).

An example of how to allocate the load is shown in FIG.
You In the figure, the calculation target pixel section G_XY(T _X, T_Y) Each picture
It shows a state in which elements are arranged in order of magnitude of concentration data, and s_CIs before
Note G_XY(T_X, T_Y) Each pixel in descending order of density data
It is one pixel (median value pixel) that comes to the center when arranged,
s_MAX, S_MINIs the above G_XY(T_X, T_Y) Inside the pixel
Then, the density data indicates pixels having maximum and minimum values. Sanawa
S_MAX≧ ... ≧ s_C≧ ... ≧ s_MINIs. The concentration
Normal sort program is used to sort data in order of magnitude.
You can do it with rum. In addition, for example, Q (1,
Calculation target pixel section G of the original image for calculating 3)
_XY(1,3) (G_X(1), G_YSecondary surrounded by (3)
Original calculation target pixel section) is arranged in order of magnitude of density data
Then, it becomes as shown in FIG.

In FIG. 7A, the weight of the median pixel s _C is set to “100%”, and the weight of all the pixels other than the median pixel s _{C is} set to “0%”.

FIGS. 7B and 7C show the above G _XY (t _X ,
t _Y ) when the pixels are arranged in descending order of density data, the load of a predetermined number of pixels arranged on both ends (the side where the density data is large and the side where the density data is small) is “0%”.
In FIG. 7B, the weight is evenly distributed to the remaining plurality of pixels (pixel group near the median value), and FIG.
Then, the weight is assigned to each pixel of the pixel group around the median value in a mountain shape, that is, the weight of the pixel in the center is the largest, and the weight is assigned smaller as the density data approaches the maximum and the minimum.

The load may be assigned in a form other than the above, but when the pixels are arranged in order of magnitude of density data, the load of the pixel near the center is increased, and the load is decreased toward both ends. It is preferable to allocate the load to. By allocating the weight in this way, it is possible to reduce the image by removing the influence of the noise component contained in the pixel having large or small density data.

The above G_XY(T_X, T_YNumber of pixels in
Is an even number, the median pixel s _CIs not decided,
In this case, for example, by the following method,
G_XY(T_X, T_Y), Except for one pixel in_XY(T
_X, T_YThe number of pixels to be processed in () is an odd number.

One pixel at a predetermined position in G _XY (t _X , t _Y ) (for example, one pixel in the upper left corner or one pixel in the lower right corner) is excluded. In the second image reduction processing described later, when reducing in the X direction, for example, G
Of the pixels at both ends in _X (t _X ), 1 on the s _X (1) side
In the case of removing one pixel of either the pixel or the one pixel on the s _{X (MAX)} side and reducing in the Y direction, for example, among pixels at both ends in G _Y (t _Y ), s _Y (1 ) Side pixel or s
One of the pixels on the _{Y (MAX)} side is removed.

P (s (t)) or P in equation (1)
(S (t) +1) load value ((n−a (t)) / m)
Or n / m and the load value of P (s (t + 1)) (a (t +
1) / m). First of all,
The X coordinate is determined as follows. G_X(T_X) Is s_X(T
_X) Pixel to s_X(T_XUp to +1) pixels
Includes ((na)_X(T_X)) / M) and (a_X(T_X+
1) / m) and the size of ((n−a)_X(T_X))
/ M) <(a_X(T_XExclude when +1) / m) 1
The X coordinate of the pixel is s_X(T_X), And ((na)
_X(T_X)) / M)> (a _X(T_XWhen +1) / m)
Is the X coordinate of one pixel to exclude_X(T_X+1).
((Na_X(T_X)) / M) = (a_X(T_X+1) /
In the case of m), the X coordinate of one pixel to be excluded is s_X(T_X)
Do you s_X(T_XIt is decided in advance whether it will be +1).
G_X(T_X) Is (s_X(T_X) +1) pixel to s
_X(T_XN / m ≧ when the number of pixels is up to +1)
(A_X(T_X+1) / m), so one pixel to exclude
X coordinate of s_X(T_X+1). 1 pixel to exclude
The Y coordinate is also determined in the same manner as the X coordinate. By this
Exclude one pixel of the original image with the set X and Y coordinate values. example
For example, G in FIG._XY(1,1) is (4,4 of the original image
Excluding the pixel of 4), G_XY(3,1) is the (1
The pixels 1 and 4) will be excluded. Note that this is
Output target pixel section G_XY(T_X, T_Y) Calculation target area R_XY
(T _X, T_Y), The above R_XY(T_X，
t_Y) With the smallest area having_XY(T_X，
t_Y) Is equivalent to selecting the pixel in. Also described later
In the second image reduction processing, when reducing in the X direction,
G depending on the above-mentioned method of determining the X coordinate of the removed pixel_X(T_X)
If you decide the pixel to be removed and reduce it in the Y direction,
G depending on how the Y coordinate of the pixel is determined_Y(T_YPixels in)
Decide.

In the load allocation shown in FIGS. 7B and 7C, the number of pixels for which the load on both ends is "0%" is, for example, all the pixels in G _XY (t _X , t _Y ). Determined by the ratio to the number. For example, if all the pixels in G _XY (t _X , t _Y ) are 15 pixels and the ratio of the pixels for which the load is “0%” is 1/3 of the total at each end, the density data is 5 pixels from the larger side and 5 pixels from the smaller side are excluded, and the weight is allocated to the remaining 5 pixels. The percentage may be determined by% or the like. In addition, when a decimal number appears in the calculation of the ratio, it is rounded up. For example, if all the pixels in G _XY (t _X , t _Y ) are 15 pixels and the ratio of the pixels for which the load is “0%” is “30%” of each end side, the total is 15 / ( Three
Since 0/100) is 4.5, it is rounded up and 5 pixels are excluded from the side having the larger density data and 5 pixels are excluded from the side having the smaller density data.
In the case of FIGS. 7B and 7C, the above G _XY (t _X ,
When the number of pixels in t _Y ) is an even number, one pixel may be excluded by the above and the number of pixels in G _XY (t _X , t _Y ) may be made an odd number, and then the load may be allocated.

When a weight is assigned to a plurality of pixels,
The weight of each pixel is determined so that the total weight is “100%”.

The form of load allocation may be determined in advance, or may be specified for each original image from the setter 44 or the like.

Step S3 of FIG. 4 (corresponding to the step (3) of the present invention): The calculation target pixel section G of the original image determined in step S1 by the load allocated in step S2.
_The result of weighted averaging of each pixel in _XY (t _X , t _Y ) is the density data Q (t _X , t of the pixel to be calculated in the reduced image.
_Y ).

For example, the load is divided as shown in FIG.
When shaken, the median pixel s_CThe concentration data of
The density data Q of the pixel to be calculated in the image after reduction
(T_X, T _Y). In this way, the image after reduction
The density data Q (t _X, T_Y) Ask
By doing so, weighted averaging of multiple pixels is not performed.
The density data of pixels in the image may be blurred or deteriorated.
There is no. Further, as shown in FIG. 20 (a), it is binarized into black and white.
If the original image is to be reduced, the calculation target pixel section of the original image
If the number of white pixels in the image is greater than the number of black pixels
Is the density data of the pixel to be calculated in the reduced image.
Use density data, and the number of black pixels is greater than the number of white pixels.
If not, the density data of the pixel to be calculated in the reduced image
After reduction, such as selecting as black density data
Calculation of the original image as the density data of the target pixel
The darker one with the larger number of white and black pixels in the output target pixel section
You can also select frequency data. In addition, concentration data
Of the noise component included in the pixel with large or small
The image can be reduced by removing the influence. Also, FIG.
G of the original image of 0 (a)_XYCalculation target image such as (3, 1)
When the density data of all pixels in the elementary section should be the same
If some pixels contain noise, the conventional method will
Will be affected by the noise component of
When the weights are allocated as shown in (c), the noise component
The image can be reduced by removing the influence.

FIG. 9 shows the result of reducing the original image of FIG. 20 (a) by allocating the load as shown in FIG. 7 (a). As is clear from a comparison between FIG. 9A and FIG. 20B, in FIG. 9A, it is possible to prevent the density data of each pixel of the image SG after reduction from being blurred or lowered, which is preferable. The image has been reduced.

Also, s _X (1) = s _Y (1) = 1, ah
_{When X} = ah _Y = 0 and the image reduction processing is performed by 9/10 in each of the X and Y directions by the conventional method, as shown in FIG. 10, Q (5,5) is P (5,5), P (6,5),
P (6,5) and P (6,6) have an equal average of four pixels, and Q (5,5) is blurred. However, if a load is allocated as shown in FIG. Since one of the pixels is selected, Q (5,5) is not blurred, and it is possible to prevent grid-like unevenness from occurring in the reduced image.

Even in the second image reduction processing to be described later, if a load is allocated as shown in FIG. 7A when calculating Q (5) in FIG. Since Q (5) is not blurred, it is possible to prevent linear unevenness from occurring in the reduced image.

When weights are allocated as shown in FIGS. 7B and 7C, the total of the results obtained by multiplying the density data of each pixel in the pixel group around the median by the weight allocated to each pixel, The density data Q (t
_X , t _Y ). In this way, when the density data Q (t _X , t _Y ) of the pixel to be calculated in the reduced image is obtained, it is a weighted average of a plurality of pixels. In this case, the weighted average of a plurality of pixels is the density data. Since the pixels near the center in the order of magnitude are targeted, it is possible to reduce blurring and deterioration of the density data of the pixels of the image after reduction as compared with the conventional method.
On the other hand, even if the median pixel includes a noise component, the noise component can be reduced. Further, when the pixels are arranged in order of magnitude of density data, a predetermined number of pixels arranged on both ends are excluded from the weighted average, so the influence of noise components included in pixels with large or small density data is removed. You can also reduce the image.

11 and 12 show an example of the result of reducing the original image of FIG. 20A by allocating the load according to FIGS. 7B and 7C. In addition, in FIG. 11, the load is "0.
In the case of weight-averaging the remaining 10 pixels with 3 pixels on each side, the weight 1/10 (10%) is evenly allocated to the remaining 10 pixels. As a result of reduction, FIG. 7B shows the result of reduction when the remaining 10 pixels are assigned a mountain-shaped load as shown in the figure. Further, FIG. 12 shows a case where the pixels for which the load is set to “0%” are set to 5 pixels on both end sides and the remaining 6 pixels are weight-averaged.
6 (100/6%) is a reduction result when the weights are evenly allocated, and FIG. 6B shows a reduction result when the remaining 6 pixels are weight-allocated as shown in the figure. 11 and 12
As is apparent from a comparison between the above drawings and FIG. 20 (b), the density data of each pixel of the image after the reduction is less likely to be blurred or reduced as compared with the conventional method.

Returning to FIG. 4, the above steps S1 to S3 are repeated until the following end determination is satisfied, and each Q (t
_X , t _Y ) are sequentially obtained, a reduced image obtained by reducing the original image is generated, and the process ends when the end determination is satisfied (step S4).

Note that Q (t _X , t _Y ) is as shown in FIG.
, The X direction is preferential, that is, Q (1,1),
Q (2,1), ..., Q (t _{X (MAX)} , 1), Q (1,
2), Q (2, 2), ..., Q (t _{X (MAX)} , 2), ..., Q
(1, t _{Y (MAX)} ), Q (2, t _{Y ( MAX)} ), ..., Q (t
_{X (MAX)} , t _{Y (MAX)} ) may be obtained in this order, or FIG.
As shown in (b), the Y direction has priority, that is, Q (1,
1), Q (1, 2), ..., Q (1, t _{Y (MAX)} ), Q
(2,1), Q (2,2), ..., Q (2, t _{Y (MAX)} ),
..., Q (t _{X (MAX)} , 1), Q (t _{X (MAX)} , 2), ..., Q
You may obtain | require in order of _{(tX (MAX)} , tY _(MAX) ).

When the number of pixels in the X and Y directions (t _{X (MAX)} , t _{Y (MAX)} ) of the reduced image is predetermined, t of Q (t _X , t _Y ) _X is in the range of 1 to t _{X (MAX)} ,
Also, each Q (t _X , t _Y ) in the range of t _Y of 1 to t _{Y (MAX} )
Then, the processing is terminated when Q (t _{X (MAX)} , t _{Y (MA X)} ) is obtained.

In the X and Y directions, the processing range SH, s _Y (1 ₎ from the s _X (1) th pixel to the last pixel (s _{X (MAX)} th pixel) of the original image is processed in the X direction. ) th from the pixel to the last pixel (s Y _(MAX) th pixel) in the case of processing a Y direction processing range SH, for the X-direction, is used for the calculation of Q (t _X, t _Y) The processing is performed in a range until the pixels lined up in the X direction of the original image reach the s _{X (MAX)} th pixel, and the Y direction is set to the Y direction of the original image used for calculating Q (t _X , t _Y ). Processing is performed within the range until the lined pixels reach the s _{Y ( MAX)} th pixel, and the pixel of the original image used for calculating Q (t _X , t _Y ) in the X direction and the Y direction is the s _{X (MAX)} th pixel. , S _{Y (MAX)} th pixel is reached, the process ends.

Furthermore, the processing range SH of the X and Y directions is
Pixel No (s) of the rear pixel_{X (LIM)}, S _{Y (LIM)}) Has been decided
In the case of the X direction, Q (t_X, T_Y)
The pixels lined up in the X direction of the original image used for output are s_{X (LIM)}Th
Perform processing in the range until reaching the pixel, and in the Y direction
Is Q (t_X, T_Y) In the Y direction of the original image used to calculate
The pixels lined up are s_{Y (LIM)}Within the range until the second pixel is reached
And Q (t_X, T_Y) Calculation
The pixels of the original image used for_{X (LIM)}Th, s_{Y (LI M)}Th
When the number of pixels is reached, the process ends.

All pixels in the X direction of the original image are processed.
In the case of target SH, s_X(1) = 1 (processing range S
The pixel number of the pixel at the end of H (s_{X (LIM)}) When deciding
Is s _{X (LIM)}= S_{X (MAX)}) And the X direction of the original image
If a predetermined number of pixels at the end of the
_X(1) = [number of truncated pixels on the head side + 1] (processing range
The pixel number (s) of the pixel at the tail of SH_{X (LIM)}) When deciding
Is s_{X (LIM)}= [S_{X (MAX)}-Number of truncated pixels on the tail side-
1]). The processing range SH in the Y direction is also the above X direction
The processing range SH can be determined in the same manner as the processing range SH.

The reduction ratios in the X and Y directions may be the same.
Image with different reduction ratio in X and Y directions
You may reduce it. In this case, the calculation target pixel area of the original image
Picture G _XY(T_X, T_Y) To be calculated in the X direction
Pixel section G_X(T_X) Is determined according to the reduction ratio in the X direction,
Calculation target pixel section G in the Y direction_Y(T_Y) Is reduction in Y direction
You may decide according to the rate.

Next, the second image reduction processing of the present invention will be described with reference to FIG.
4, with reference to the flowchart shown in FIG.
Note that FIG. 14 and FIG. 15 are flowcharts in the case where the image reduction processing is performed for each pixel row of the original image in which pixels are arranged in the one-dimensional reduction direction to reduce the original image in the reduction direction. Also, the original image is stored in the original image memory 32 when the processing is started.
Be stored in.

In this image reduction processing, the image reduction for one pixel column to be processed is performed in steps T2 to T7 of FIG.
This process is sequentially executed for each pixel column arranged in parallel in the direction orthogonal to the reduction direction to reduce the entire two-dimensional original image in the one-dimensional reduction direction. In the flowcharts of FIGS. 14 and 15, the processing range SH in the direction orthogonal to the reduction direction (when the pixel row on the end side is removed to reduce the image) is stored in the original image memory 32. The i-th pixel column in the direction orthogonal to the reduction direction of the original image (i = s _Y (1), s _Y (1) +1, s _Y (1) + when reducing in the X direction.
2, ..., and i = s _X (1) when reducing in the Y direction,
The reduction result of s _X (1) +1, s _X (1) +2, ...
Although it is processed so as to be stored as the j-th pixel column (j = 1, 2, 3, ...) In the direction orthogonal to the reduction direction of the reduced image memory 33, the original image stored in the original image memory 32 is processed. When all the pixel columns in the direction orthogonal to the reduction direction of the image are to be processed, the process may be performed with i = j (1, 2, 3, ...).

For example, in the case of image reduction in which the X direction in FIG. 2 is the reduction direction, each pixel column arranged in parallel in the Y direction is sequentially processed as one pixel column to be processed. In this case, Q in one pixel column (i-th pixel column) to be processed in the original image
Of the pixels included in the calculation target pixel section G _x (t _x ) for calculating (t _x , j), the density data (P (s _x , i)) of the pixel used for weighted averaging is read from the original image memory 32. To reduce the image, and the result is reduced to (t
_X, the density data Q (t _X pixel coordinate values of j), j) will be stored in the reduced after image memory 33 as. On the other hand, FIG.
In the case of image reduction in which the Y direction is the reduction direction, each pixel column arranged in parallel in the X direction is sequentially processed as one pixel column to be processed. In this case, one pixel column (i
Of the pixels included in the calculation target pixel section G _Y (t _Y ) for calculating Q (j, t _Y ) in the (th pixel row), the density data (P (i, s _Y )) of the pixel used for weighted averaging ) to read from the original image memory 32 and image reduction, the result of reduced post-image memory 33 (j, t _Y) density data Q (j pixel coordinate values of the reduced after image memory 33 as t _Y) I will remember.

Step T1 in FIG. 14: Initialization is performed.
Set "1" to j and set s to i for reduction in the X direction.
Set _Y (1) and s _{X to} i for reduction in the Y direction
Set (1). Note that s _Y (1), s _X (1),
Reduction ratio (n / m), ah _X (for reduction in the X direction),
Necessary data such as ah _Y (in the case of reduction in the Y direction) is set by the setter 44 or the like prior to the processing of step T1.

Step T2 in FIG. 14: Initial setting of image reduction processing for one pixel column is performed. T for reduction in the X direction
Set "1" to _X and "1" to t _Y for reduction in the Y direction.

Step T3 of FIG. 14 (corresponding to the step (1) of the present invention): density of the pixel to be calculated in the reduced image (pixel to be calculated in the pixel row to be processed in the reduced image) The calculation target pixel section of the original image (in the pixel column of the processing target of the original image) used for calculating the data is reduced by the image reduction ratio (n /
m).

This calculation target pixel section can be determined based on the conventional equation (1), for example. That is, in the case of reduction in the X direction, the s _X (t _X ) th pixel to the s _X (t _X ) th pixel in the i-th pixel column of the pixel columns arranged in the Y direction of the original image are reduced. Partition up to the pixel of
When n−a _X (t _X ) = 0, (s _X (t _X ) +
The section from the 1) th pixel to the s _X (t _X +1) th pixel) is set as a calculation target pixel section G _X (t _X ). Also,
In the case of reduction in the Y direction, the s _Y (t _Y ) th pixel to the s _Y (t _Y +1) th pixel in the i-th pixel column of the pixel columns arranged in the X direction of the original image are reduced. Partition to pixel (however, if n−a _Y (t _Y ) = 0, then (s
_{The section from the Y} (t _Y ) +1) th pixel to the s _Y (t _Y +1) th pixel) is defined as the calculation target pixel section G _Y (t _Y ).

In the case of FIG. 20A, G _X (1) is a section made up of four pixels (1 to 4, i), G _X (2)
Is a section composed of four pixels of (4 to 7, i), G
_X (3) is a section composed of four pixels (8 to 11, i), ..., And G _Y (1) is a section composed of four pixels of (i, 1 to 4), G _Y ( 2) is a section composed of four pixels of (i, 4 to 7), and G _Y (3) is 4 of (i, 8 to 11).
A section composed of individual pixels, ....

Also in this processing, in the case of reduction in the X direction, ah _X is uniquely set to "0" to determine G _X (t), and in the case of reduction in the Y direction, ah _Y is uniquely determined. "0"
G _Y (t) may be determined as

By thus determining the calculation target pixel section and using the pixels in the calculation target pixel section to obtain the density data of the calculation target pixel of the reduced image, the processing range in the reduction direction of the original image is obtained. The entire SH can be reduced approximately n / m times evenly.

Step T4 (corresponding to the step (2) of the present invention) and step T5 (corresponding to the step (3) of the present invention) of FIG. 14 are steps S2 and S of the first image reduction processing.
The process is the same as 3. That is, in the second image reduction processing, in step T3, the processing target 1 of the original image is processed.
A calculation target pixel section is determined for the pixel row, and, for example, as shown in FIG. 7, each of the calculation target pixel sections of the original image is determined according to the magnitude order of the density data of each pixel in the calculation target pixel section. The weight is assigned to the pixels, and the density data of the pixels to be calculated in the reduced image is obtained by the weighted average of the assigned weights. Thereby, in the image reduction in the one-dimensional reduction direction, the same effect as that of the first image reduction process can be obtained.

In step T6 of FIG. 14, it is determined whether or not the image reduction processing for the current one pixel column to be processed is completed. If not completed, t _X is set at the time of reduction in the X direction.
Is counted up, and at the time of reduction in the Y direction, t _Y is counted up (step T7), the process returns to step T3, and the density data of the next pixel in the pixel row of the current processing target of the reduced image is obtained. Ask. Thereafter, steps T3 to T are performed until the image reduction process for the current one pixel column to be processed is completed.
In the case of reduction in the X direction by repeating the process of 7, t _X
= 1, 2, 3, ...
j), Q (2, j), Q (3, j), ... In order, the density data of each pixel forming the j-th pixel column in the Y direction of the reduced image is converted to the i-th image in the Y direction of the original image. In the case of reduction in the Y direction, the count is sequentially incremented to t _Y = 1, 2, 3, ..., And Q (j, 1), Q (j,
2), Q (j, 3), ... In order, the density data of each pixel forming the j-th pixel column in the X direction of the reduced image is obtained from the i-th pixel column in the X direction of the original image. . When the image reduction processing for the current one pixel column to be processed is completed, it is determined in step T8 in FIG. 15 whether or not the image reduction processing has been completed for all pixel rows to be processed. If not, i, Each j is counted up (step T9), the process returns to step T2 in FIG. 15, the reduction process is performed on the next pixel column, and the process is terminated when the image reduction process on all the pixel columns to be processed is completed.

In the case of reduction in the X direction (reduction in the Y direction), the end determination of step T6 is performed as follows. The number of pixels in the X direction (Y direction) of the reduced image (t _{X (MAX)} (t
_{Y (MAX)} ) is predetermined, Q (t
_{When X (MAX)} , j) (Q (j, t _{Y (MAX)} )) is obtained, the image reduction processing of the one pixel column to be processed is terminated.

Further, in the X direction (Y direction), from the s _X (1) (s _Y (1)) th pixel of the original image to the last pixel (s _{X (MAX)} th pixel (s _{Y (MAX)).} When processing up to the (th pixel)) as the processing range SH in the X direction (Y direction), the processing target of the original image used for calculation of Q (t _X , j) (Q (j, t _Y )) is processed. Image reduction processing of the one pixel row to be processed is performed within the range until the pixels in the one pixel row reach the s _{X (MAX)} th pixel (s _{Y (MAX)} th pixel), and Q (t _X , j )
When the pixel in the one pixel column of the processing target of the original image used for the calculation of (Q (i, t _Y )) reaches the s _{X (MAX)} th pixel (s _{Y (MAX)} th pixel), The image reduction processing for one pixel column is completed.

Further, the processing range SH in the X direction (Y direction)
Hara used to calculate the pixels of the tail of the pixels _{No (s X (LIM) (} s Y (LIM))) if they decide _{is, Q (t X, j)} (Q (i, t Y)) to The pixel in the one pixel column of the image processing target is s
Image reduction processing is performed on the 1-pixel row to be processed in the range up to the _{X (LIM)} th pixel (s _{Y (LIM)} th pixel), and Q
The pixel in the one pixel column of the processing target of the original image used for the calculation of (t _X , j) (Q (i, t _Y )) is the s _{X (LIM)} th pixel (s _{Y (LIM)} th pixel) Is reached, the image reduction processing of the one pixel column to be processed ends.

In the case of reduction in the X direction (reduction in the Y direction), the end determination of step T8 is performed as follows. If the number of pixel columns (t _{Y (MAX)} (t _{X (MAX)} ) in the Y direction (X direction) of the reduced image is predetermined, j = t
_When the reduction processing of the pixel row of _{Y (MAX)} (j = t _{X (MAX)} ) is finished, the processing is finished.

Further, in the Y direction (X direction), from the s _Y (1) (s _X (1)) th pixel of the original image to the last pixel (s _{Y (MAX)} th pixel (s _{X (MAX))} Up to the (th pixel)) as the processing range SH in the Y direction (X direction), when the reduction processing of the pixel row of i = s _{Y (MAX)} (i = s _{X ( MAX)} ) is finished. Ends the process.

Further, the processing range SH in the Y direction (X direction)
If the pixel number (s _{Y (LIM)} (s _{X (LIM)} ) of the tail pixel is determined in, the pixel row of i = s _{Y (LIM)} (i = s _{X (LIM)} ) is reduced. The process ends when the process is completed.

The processing range SH in the X and Y directions is
As described in the first image reduction processing, s _X (1)
(S _{X (LIM)} ), s _Y (1), and (s _{Y (LIM)} ) may be determined.

Further, for example, when an image capturing apparatus using a line sensor captures a pixel column in each main scanning direction and immediately reduces the image of the pixel column, the original image memory 32 is set as shown in FIG. , The data for one pixel column in the main scanning direction can be stored, and each time the image capturing device captures data for one pixel column in the main scanning direction, the data is stored in the original image memory 32 in FIG. Figure 1
4, the image may be successively reduced by the processing of FIG. 15 and stored in the reduced image memory 33.

[0100]

As is apparent from the above description, according to the first aspect of the present invention, the weighted average by the weights assigned according to the magnitude order of the density data of each pixel of the calculation target pixel section of the original image is used. Since the density data of the calculation target pixel of the reduced image is obtained, the weighted average of the density data of the original image is used to calculate the density of the image after the reduction regardless of the position of each pixel in the calculation target pixel section. The density data of the pixel to be calculated can be calculated, and the distribution state of the density data of each pixel of the pixel block to be calculated of the original image can be reflected in the reduced image, preventing the deterioration of the image quality due to the image reduction. You can

[0101]

According to the invention described in claim 1 ,
When the pixels of the calculation target pixel section of the original image are arranged in order of magnitude of the density data, the weight of a predetermined number of pixels arranged at both ends is set to "0%" and the weighted average is removed, and each pixel is compared with the density data. When the pixels are arranged in order, the density data of the pixel to be calculated in the reduced image is calculated by the weighted average of a plurality of pixels (pixel group near the median value) near the center. density data of pixels while can be reduced to decrease or blurring, it can be contained noise components to the medium median pixel to reduce the noise component. Further, when the pixels are arranged in order of magnitude of density data, a predetermined number of pixels arranged on both ends are removed from the weighted average, so that the influence of noise components contained in pixels with large or small density data is eliminated. The image can be reduced by removing it.

According to the second aspect of the invention, the image reduction in the two-dimensional direction of the original image can be performed simultaneously and suitably.

According to the third aspect of the present invention, the image reduction process is preferably performed for each pixel row of the original image in which pixels are arranged in the one-dimensional reduction direction, and the original image is preferably reduced in the reduction direction. You can

According to the invention described in claim 4 , it is possible to cause a computer to execute the image reduction processing by the method of the invention described in claim 1.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a hardware configuration for implementing an image reduction method according to the present invention.

FIG. 2 is a diagram showing a configuration of an original image memory.

FIG. 3 is a diagram showing a configuration of a reduced image memory.

FIG. 4 is a flowchart showing a procedure of a first image reduction process according to the present invention, which simultaneously performs image reduction in a two-dimensional direction of a two-dimensional digital original image.

FIG. 5 is a diagram showing a two-dimensional calculation target area and a two-dimensional calculation target pixel section.

FIG. 6 is a diagram showing a relationship between an original image and a reduced image.

FIG. 7 is a diagram showing an example of a method of allocating weights according to the magnitude data of the density data of each pixel in the calculation target pixel section.

FIG. 8 is a diagram showing a state in which each pixel in a calculation target pixel section of a black-and-white binarized original image is arranged in descending order of density data.

FIG. 9 is a diagram showing a result of reducing the original image of FIG. 20 (a) by the load allocation of FIG. 7 (a).

FIG. 10 is a diagram for comparing and explaining a conventional method and a method according to the present invention in image reduction when the reduction rate is close to “1.0”.

FIG. 11 shows the distribution of the loads shown in FIGS.
It is a figure which shows an example of the result of having reduced the original image of (a).

FIG. 12 shows the distribution of the load shown in FIGS.
It is a figure which shows another example of the result of having reduced the original image of (a).

FIG. 13 is a diagram showing a case where the processing procedure is prioritized in the X direction and a case where it is prioritized in the Y direction.

FIG. 14 shows a procedure of a second image reduction process according to the present invention, in which an image reduction process is performed for each pixel column of an original image in which pixels are arranged in a one-dimensional reduction direction to reduce the original image in the reduction direction. It is a flowchart.

FIG. 15 is a flowchart showing a procedure of second image reduction processing according to the present invention.

FIG. 16 is a diagram showing a configuration of a modified example of the original image memory.

FIG. 17 is a diagram for explaining a conventional image reduction method.

FIG. 18 is a diagram showing a conventional image reduction method for each pixel column.

FIG. 19 is a diagram showing a procedure for reducing a two-dimensional original image in a two-dimensional direction by a conventional method.

FIG. 20 is a diagram for explaining a problem of the conventional method.

FIG. 21 is a diagram for explaining another problem of the conventional method.

FIG. 22 is a diagram showing another problem of the conventional method.

[Explanation of symbols]

10: CPU 31: Program memory 32: Original image memory 33: Reduced image memory 45: External storage device P (s _X , s _Y ): Original image pixel density data Q (t _X , t _Y ): Reduced Data of pixels of the image G _XY (t _X , t _Y ), G _X (t _X ), G _Y (t _Y ): Calculation target pixel section s _C : Median pixel

Front page continuation (58) Fields surveyed (Int.Cl. ⁷ , DB name) H04N 1/38-1/393 G06T 3/40

Claims (4)

(57) [Claims] 1. An image reduction method for reducing a digital original image to obtain a reduced image, comprising: (1) a calculation target of an original image used for calculating density data of pixels to be calculated in the reduced image. A step of determining the pixel section according to the reduction ratio of the image; and (2) a step of determining the density data of each pixel in the calculation target pixel section of the original image determined in the step (1) A step of assigning a weight to each pixel in the calculation target pixel section, and (3) Each pixel in the calculation target pixel section of the original image determined in the step (1) according to the load assigned in the step (2). The result of weighted averaging is used as the density data of the pixel to be calculated in the reduced image, and the density data of each pixel of the reduced image is obtained in the step consisting of to obtain an image, and Engineering of the (2) So determined in step (1)
Each pixel in the calculation target pixel section of the original image
When arranging in order of size, load of a specified number of pixels on both ends
The weight is set to "0%", and the load is divided to other pixels.
An image reduction method characterized by shaking . 2. The image reduction method according to claim 1 ,
An image reduction method characterized by simultaneously reducing a two-dimensional image of a two-dimensional digital original image. 3. An image reduction method according to claim 1 is performed for each pixel row of the original image in which pixels are arranged in a one-dimensional reduction direction to reduce the original image in the reduction direction. Image reduction method characterized by. 4. A storage medium recording a program for causing a computer to execute a process of reducing a digital original image to obtain a reduced image, comprising: (1) a calculation target pixel of the reduced image. A step of determining a calculation target pixel section of the original image used for calculating the density data according to a reduction ratio of the image; and (2) each pixel in the calculation target pixel section of the original image determined in the step (1). According to the magnitude order of the density data of, the load is assigned to each pixel in the calculation target pixel section of the original image, and (3) the load assigned in the process of (2) The result of weighted averaging of each pixel in the determined calculation target pixel section of the original image is used as density data of the calculation target pixel of the reduced image, and a step of Original image by obtaining density data To obtain a reduced image obtained by reducing, and, in the step (2), determined in step (1)
Each pixel in the calculation target pixel section of the original image
When arranging in order of size, load of a specified number of pixels on both ends
The weight is set to "0%", and the load is divided to other pixels.
A storage medium that records a program that causes a computer to execute a swing process.