JPH07210669A - Image processor and its control method - Google Patents

Abstract

PURPOSE:To obtain a satisfactory output image independently of the classification of the image even in the case where a gradation image is expanded or has the resolution increased. CONSTITUTION:When an image 101 having a low resolution is inputted, the input image is divided according to respective gradation values of the image (102), and edges are extracted from each divided image (103). Vector data is generated from edge images of gradation values (105). Extracted vectors are variably magnified based on the increase of the resolution, and the outlines are drawn (106). The insides of outlines are painted out (107), and they are finally synthesized (108).

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 a control method therefor, and more particularly to an image processing apparatus and a control method therefor when input image information is enlarged or resolution is increased.

[0002]

2. Description of the Related Art Conventionally, various methods have been proposed as a method for converting input low resolution information into high resolution information. The proposed conventional method is based on a target image type (for example, a multi-valued image having gradation information for each pixel, a binary image binarized by pseudo halftone, and a binarized by a fixed threshold). The conversion processing method differs depending on the binary image, character image, etc.).

The image targeted by the present invention is a multi-valued image such as a natural image having gradation information for each pixel. In the conventional interpolation method, the same pixel value closest to the interpolation point is arranged. Closest interpolation method or 4 points of the original image surrounding the interpolation points (interpolation pixels) (pixel values at 4 points are A, B, C and D)
A co-temporal interpolation method or the like is generally used that determines the pixel value E to be interpolated by the following calculation according to the distance.

E = (1-i). (1-j) .A + i. (1-j) .B + j. (1-i) .C + i.j.D (however, the distance between pixels When 1 is set to 1, the pixel to be interpolated is located at a distance of i from the pixel A in the horizontal direction and j in the vertical direction (hence, i ≦ 1, j ≦ 1).

[0005]

However, the above-mentioned conventional example has the following drawbacks.

The closest interpolating method in which the same pixel value closest to the interpolation point is arranged has the advantage that the configuration is simple, but since the pixel value is determined for each block to be enlarged,
The presence of blocks is visually conspicuous and the image quality is poor.

The co-linear interpolation method, which is calculated by the distances of four points surrounding the interpolation point, is an effective method that is generally used for enlarging a natural image and is averaged.
The image quality is smoothed. However, in the edge portion and the portion where sharp image quality is required, the image quality becomes blurred. Further, in the case of an image obtained by scanning a map or the like or an image such as a natural image including a character portion, there is a problem that important information is not transmitted to the receiver side due to blurring due to interpolation.

[0008]

[Means for Solving the Problems] and

The present invention has been made in view of the above problems, and it is possible to obtain a good output image regardless of the kind of the image even when the gradation image is enlarged or the resolution is increased. It is an object of the present invention to provide an image processing apparatus and a control method thereof that enable the image processing apparatus.

In order to solve this problem, the image processing apparatus of the present invention has the following configuration. That is, an image processing device for enlarging or increasing the resolution of image data, in which the input image data is divided based on the density distribution,
After expanding or increasing the resolution of each divided image, the respective images are combined and output.

Further, the control method of the image processing apparatus of the present invention includes the following steps. That is, a method of controlling an image processing apparatus for enlarging or increasing the resolution of image data, in which input image data is divided based on the density distribution, and each divided image is enlarged or increased in resolution. After that, the respective images are combined and output.

[0011]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows an example of the configuration of a resolution conversion system in the embodiment. This time, as an example, a, b, ...
The case of the resolution conversion system in 8 gradations of h is shown, but of course, the gradation may be any gradation.

In the figure, reference numeral 101 is an original low-resolution multivalued image. 102 is an image decomposition unit. 103 is an edge extraction unit. Reference numeral 104 is a representative point setting unit. Reference numeral 105 is a vector information creation unit. Reference numeral 106 is a drawing unit. 1
Reference numeral 07 is a filled portion. Reference numeral 108 is an image combining unit.

The operation of the above configuration will be described below.

First, low resolution image data 1 to be input
It is assumed that 01 is image data of 8 gradations (pixel values 0 to 7). At this time, in the screen division unit 102, the image data is divided into eight stages of a, b, ... H. Next, the edge extraction unit 103 extracts an edge portion from each of the eight plates divided by the screen division unit 102. Next, the control point extraction unit 104 extracts a representative point on each side from the edge information obtained by the edge extraction unit 103. Based on the representative points obtained here, the vector information generation unit 105
Vectorization is performed, and the drawing unit 106 enlarges to an arbitrary magnification (enlargement ratio set by the user) and draws.

Next, the inside of the edge portion drawn by the drawing unit 106 is filled with each gradation value by the filling unit 107. Finally, the planes divided by the image combination unit are superimposed on each other to generate a high-resolution multivalued image.

Next, each of the above components will be described.

The image dividing unit 102 divides the input image information as shown in FIG. 2 into each gradation information. here,
For simplification, the image of each density value of the input image is 2
It is digitized and gradation information is provided separately. Thereby, the image of each gradation may be stored as binary data, so that the memory efficiency can be improved. However, it may have information in multivalued form.

Next, in the edge extraction unit 103, the binary image for each of the divided gradation images is enlarged three times vertically and horizontally as shown in FIG. Here, the reason for expanding to 3 times is that, in this example, since the control point is generated in dot units, it is the minimum magnification necessary to obtain the contour information of a dot of 1 dot × 1 dot. Is.

Next, as shown in FIG. 5, when the values of the target pixel in the four directions of up, down, left, and right match (four-direction match), the process of removing the target pixel is performed, and the edge of Extract parts. This time, the four-direction matching is used to reduce the number of patterns in some cases (to improve the processing speed), and other methods such as eight-direction matching may be used.

Next, the control point extraction unit 104 sets the information in three patterns shown in FIG. 6 based on the edge information extracted by the edge extraction unit.

6 (a) is judged to be folded back, and (b) is shown.
Is determined to be a corner portion, and the portion like (c) is determined to be the other portion. FIG. 7 shows that the representative points are set by performing these three cases (the point positions that are the features of each pattern are
It is assumed to be predetermined for each pattern).
The white portion in FIG. 7 is the representative point. 6A and 6C are the same as those in FIG. 6B, the difference between them is determined by whether several dots or more are continuous after being bent. I shall. Since the number of dots serving as the threshold value in this case depends on the resolution of the input image and the like, it is possible to appropriately set these determination criteria. In addition, the respective patterns shown in FIG. 6 are considered including the case where each pattern is rotated, and are not necessarily shown in the illustrated direction (for example, FIG.
In (a), it does not mean only that it protrudes or is depressed to the right.

The vector information creating section 105 scales the representative point information in order to create an image of output resolution (of course, taking into account that it has already been tripled).

In the drawing unit 106, the scaled representative points are connected by a short vector, and thereby the contour line of the output image is generated.

The filling section 107 fills the inside of the outline obtained by the drawing section 106 to create an image with a smooth resolution to be output.

By performing the above steps 102 to 107 for each of the divided images and combining (logical sum) by the image combining unit 108, a multi-valued image having a resolution to be resolved and output can be obtained.

The operation based on FIG. 1 has been described above. Desirably, as shown in FIG. 1, it is desirable to perform the image division-filling processing in parallel for each gradation value, but as an example of sequentially performing each processing, the apparatus (or system) shown in FIG. 15 is used. The specific configuration of is shown.

In FIG. 15, 1 is a CPU that controls the entire apparatus, 2 is a ROM that stores a program for causing the CPU 1 to function as a processing unit described later, and 3 is a C
RAM 4 used for the work area of PU 1 and the like is an external storage device such as a hard disk device. 5 is 20
Reference numeral 6 denotes an image scanner for reading a document image at about 0 dpi, and reference numeral 6 denotes a scanner I / F for converting image data output from the image scanner 5 into digital data (for example, 8 gradations) and incorporating it into the apparatus main body. Reference numeral 7 is an image memory for storing image data, and 8 is a printer I.
Reference characters / F and 9 are printers for inputting image data and recording images. Note that this printer is a laser beam printer having a resolution of 600 dpi (three times the resolution of a scanner), and for the sake of simplicity, it is possible to record an image of 8 gradations. It is assumed that a well-known PWM (pulse width modulation) method is adopted as a technique for recording a multi-gradation image with a laser beam printer, and a description thereof will be omitted here.

The operation of FIG. 15 will be described with reference to the flowchart of FIG. The program based on the flowchart is stored in the ROM 2.

In step S1, a gradation image is input from the image scanner 5, and the image is temporarily stored in the image memory 7 after being vertically and horizontally tripled. Here, since the input image has 8 gradations, it is stored as 3 bits per pixel. Next, the variable n (preserved in the RAM 3) is set to "0". Next, in step S3, the variable n is incremented by "1", and it is determined in step S4 whether vector extraction of all gradation values has been completed.

When it is determined that the image is incomplete, the process proceeds to step S5, and the image having the gradation value n ("1" when the process is first executed in this step) in the image memory 7 is separated. In step S6, the edges of the separated image of the gradation value n are extracted by performing the four-direction matching process described above. Actually, it is not necessary to separate and extract the edge portion because it is sufficient to focus only on the pixel of the gradation value n in the image data stored in the image memory 7 to extract the edge portion.

Next, in step S7, each pattern is fitted, feature points are selected, vector data is extracted, and the extracted data is stored in the external storage device 4 in a file format. The reason why the external storage device 4 is stored as a file is that the external storage device 4 is more advantageous in terms of cost performance. However, if the memory has a margin or the input image size is small, it is better to positively use the memory chip in terms of speed. At this time, since the edge portion having the gradation value n remains in the image memory 7, this portion is deleted. Then, thereafter, the process returns to step S3 to extract the vector of the next gradation value. Therefore, the determination of step S4 being "Yes" means that the image memory 7 is initialized to "0".

When the extraction of the vector data for all gradation values is completed in this way, the process proceeds to step S8,
The variable n is initialized again with "0", and its value is incremented by "1" in step S9.

In step S10, it is determined based on the value of the variable n whether or not the drawing of images of all gradation values has been completed. When it is determined that the processing is not completed, the following steps S11 and S12 are performed. In step S11,
Vector data is sequentially taken out from the file of the gradation value n stored in the external storage device 4, scaled based on the resolution of the input image and the resolution of the output image, and drawn in the image memory 7. However, in this embodiment, the resolution of the input image is 200 dpi, the output resolution of the output image (printer 10) is 600 dpi, and the scaling processing is not performed because the vertical and horizontal triple processing has already been performed at the time of vector extraction. When this drawing is completed, the process proceeds to step S12,
The inside of the drawn contour is filled with a gradation value n. After that, the process returns to step S9 to draw the next gradation value n to be processed.

In this way, when the drawing processing of all the gradation values is completed, since the images of all the gradations have been developed in the image memory 7, the contents are sequentially output to the printer 9 in step S13. To do.

As a result, according to the present embodiment, it is possible to convert the resolution of a multi-valued image or a color image like a binary image, and output without depending on the resolution of the input device. It will be possible.

In the above embodiment, an example of converting low resolution image data into image data of higher resolution has been described, but the present invention is not limited to this. This is because, for example, the resolutions of the input image and the output image are the same, and even when the input image is enlarged, the same processing can be implemented.

Further, as shown in FIG. 15, the apparatus of the embodiment can be applied to a copying machine, but at the same time, it can be applied to a system including an image scanner and a printer, and even if an input image is a read image from a scanner, a floppy disk is used. It may be an image stored on a disk or an image received via a line, and the output target is a printer,
It goes without saying that it may be a display device or a line (a line for transmitting to a destination terminal).

<Description of Other Embodiments> Next, other embodiments of the present invention will be described.

[Second Embodiment] FIG. 10 is an overall view of the second embodiment. In the figure, 1001 is a low-resolution multivalued image, 1002 is an image dividing unit, 1003 is an edge extracting unit,
Reference numeral 1004 is a representative point setting unit, 1005 is a vector information creating unit, 1006 is a drawing unit, 1007 is a filling unit, 10
Reference numeral 08 is an image combining unit.

This time, as input images, 0x10, 0x3
0,0x50,0x70,0x90,0xb0,0xd
8 of 0,0xf0 (“0x” indicates a hexadecimal number)
It is assumed that the image has gradation. Here, the case of the super-resolution conversion system with 8 gradations is shown, but of course, the gradation may be any gradation.

In the second embodiment, the image division unit 1002
In (1), the input low-resolution multi-valued image is not divided by dividing only the gradation value information as in the first embodiment, but is cut out as information below the gradation value to be extracted.
That is, 0x00-0x10, 0x00-0x30,
0x00-0x50, 0x00-0x70, 0x00-
0x90, 0x00-0xb0, 0x00-0xd0,
There are eight numbers from 0x00 to 0xf0. The features of this dividing method will be described below with reference to the drawings.

Reference numeral 1101 in FIG. 11 is low resolution image information. In the case of such an image, the images 1102 and 1103 are obtained when binarized by using the division method of the first embodiment, and the images 1104 and 1105 are obtained when the contour information is extracted.

On the other hand, when the division method of the second embodiment is used, when binarized, the images 1106 and 1107 are obtained,
When the contour information is extracted, the images 1108 and 1109 are obtained.

When the dividing method of the first embodiment is used, 1
Although the inner contour of 104 and the contour of 1105 have the same contour, vectorization needs to be performed respectively, but the division method of the second embodiment does not require this. Therefore, the ratio of the complicated processing related to vector extraction is reduced, and as a result, the execution time can be shortened. However, in the drawing process, for example, when drawing an image of n gradations, "1" is added to the portion which is on the already drawn n-1 gradation image.

Further, in the contour information generation method used this time, since the inner contour of 1104 and the contour of 1105 are generated from different edge information, vectorization is performed by using the division method of the first embodiment. When this occurs, the edge portion may be displaced, but by using this division method, the edge portion is always generated from the same contour information, so that the displacement does not occur.

[Third Embodiment] FIG. 12 is an overall view of the third embodiment. In the figure, 1201 is a low-resolution multivalued input image, 1202 is an image dividing unit, 1203 is an edge extracting unit,
1204 is a representative point setting unit, 1205 is a vector information generation unit, 1206 is a drawing unit, 1207 is a painting unit, and 12 is a drawing unit.
Reference numeral 08 is an image combining unit, and 1209 is an image division detecting unit.

The third embodiment is an apparatus in which the system of the second embodiment is improved so that it can be applied to the case where an image is inputted with a gradation value whose gradation value is unknown.

Here, as an example, an image of 5 gradations is taken. When the input image 1201 is input, the image division value detection unit 1209 calculates the frequency of the density based on the input low resolution image data. As a result, for example, FIG.
As shown in 3, it is possible to obtain a distribution map.

After that, as shown in FIG. 13, the input frequency for the gradation value is checked. Therefore, when there are some gradation values determined to be present, an intermediate value (value at the center of the peak) of each gradation value is calculated and used as a threshold for division. By using this method, resolution conversion can be performed even when each gradation value is not a single value but spreads to the left and right as shown in FIG.

As a result of the above, it is not necessary to have gradation information for the input image, and it is possible to convert the resolution of any image.

[Fourth Embodiment] FIG. 14 is an overall view of the fourth embodiment. In the figure, 1401 is a low-resolution color input image, 1402 is an image dividing unit, 1403 is an edge extracting unit, 1404 is a representative point setting unit, 1405 is a vector information generating unit, 1406 is a drawing unit, 1407 is a filling unit, and
1408 is an image combining unit, 1409 is an image division value detecting unit,
Reference numeral 1410 is a YUV dividing unit, 1411 is a high resolution color output image, and 1412 is a YUV combining unit.

The fourth embodiment is an example in which the third embodiment is applied to a color image. Image division value detection 1 of the third embodiment
Prior to 402, the color image is decomposed into Y, U and V into three multi-valued images. Here, the YUV space is used as the color space, but of course, the RGB space or another color space may be used. Further, when the color gradation value is already known, the division method of the first embodiment may be used. The second to fourth embodiments are the same as the first embodiment.
As described in the above embodiment, the invention may be applied to a system including a plurality of devices or an apparatus including one device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0054]

As described above, according to the present invention, it is possible to obtain a good output image regardless of the type of the image even when the gradation image is enlarged or the resolution is increased. become.

[0055]

[Brief description of drawings]

FIG. 1 is an overall view of an image processing apparatus in this embodiment.

FIG. 2 is an example of gradation division in the present embodiment.

FIG. 3 is an example of image enlargement in the present embodiment.

FIG. 4 is an example of edge extraction in the present embodiment.

FIG. 5 is a pattern used in edge extraction in this embodiment.

FIG. 6 is a pattern used for setting representative points in the present embodiment.

FIG. 7 is an example of setting representative points in the present embodiment.

FIG. 8 is an example of contour line generation in the present embodiment.

FIG. 9 is an example of painting in this embodiment.

FIG. 10 is an overall diagram of an image processing apparatus according to a second embodiment.

FIG. 11 is a diagram comparing the division methods in the second embodiment.

FIG. 12 is an overall diagram of an image processing apparatus according to a third embodiment.

FIG. 13 is an example of an image in which each input gradation value is widened in the third embodiment.

FIG. 14 is an overall diagram of an image processing apparatus according to a fourth embodiment.

FIG. 15 is a block configuration diagram in the case of being applied to a specific device in the first embodiment.

16 is a flowchart for explaining the operation of FIG.

[Explanation of symbols]

 101 low-resolution multi-valued input image 102 image division unit 103 edge extraction unit 104 representative point setting unit 105 vector information generation unit 106 drawing unit 107 filling unit 108 high-resolution multi-valued output image 1209 image division value detection unit 1410 YUV Split part 1412 YUV coupling part.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H04N 1/387 101

Claims (8)

[Claims] 1. An image processing apparatus for enlarging or increasing the resolution of image data, wherein input image data is divided based on a density distribution, and each divided image is enlarged or increased in resolution. After that, the image processing apparatus is characterized by combining and outputting the respective images. 2. The image processing apparatus according to claim 1, wherein the division based on the density distribution is performed based on the density value of the pixel. 3. The division based on the density distribution is such that a threshold for division is estimated based on the distribution of density values of pixels, and the input image is divided according to the estimated threshold. The image processing device according to item. 4. Before the input image is divided, a separating means for separating the input image in a predetermined color space is provided, and the dividing is performed for each image data of each color component separated by the means. The image processing apparatus according to claim 1. 5. A method of controlling an image processing apparatus for enlarging or increasing the resolution of image data, comprising dividing input image data based on a density distribution, and enlarging or increasing each divided image. A method for controlling an image processing apparatus, comprising: combining the images and outputting them after resolution conversion. 6. The method of controlling an image processing apparatus according to claim 5, wherein the division based on the density distribution is performed by using the density value of the pixel as a reference. 7. The division based on the density distribution is such that a threshold for division is estimated based on the distribution of density values of pixels and the input image is divided according to the estimated threshold. A method for controlling an image processing apparatus according to item. 8. A method for separating an input image in a predetermined color space before dividing the input image, wherein the division is performed for each image data of each color component separated in the step. The control method of the image processing apparatus according to claim 5.