JP3070128B2 - Method and apparatus for binarizing a gradation image - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for binarizing multi-valued image data in a scanner, a display, a printer, a facsimile or the like.

[0002]

2. Description of the Related Art Conventionally, as one of binarization methods of a gradation image, an error diffusion method (R.
Floyd & L. Steinberg: "An
Adaptive Algorithm for Sp
atial Grayscale ", SID Symp
osium Digest of Papers, pa
ge36 (1975)) is known.

FIG. 5 is an explanatory diagram of one example.
In the figure, each square is one pixel, the cross-hatched pixel 52 is a pixel being processed, and the hatched pixel 53 is a pixel that has been binarized. Other unhatched pixels 54 are unprocessed pixels. The pixel 52 being processed is binarized by a certain threshold value, that is, converted into 1 or 0. The error generated at that time is multiplied by, for example, weighting factors A to D as shown in FIG. 3B and distributed to surrounding unprocessed pixels. FIG. 6 is a schematic configuration diagram of an example that embodies the above-described error diffusion method. In the figure, 1 is an image input, 2 is an image buffer memory, 3 is an adder, 4 is a threshold value setting circuit, 5 is a comparator,
6 is an arithmetic unit, 7 is a weighting circuit, 8 is an error buffer memory, 9 is an output buffer memory, and 10 is a binarized output.

An image input 1 is temporarily stored in an input buffer memory 2 and is read out pixel by pixel. After the adder 3 adds the error data of the processed pixels, the comparator 5
Is input to The comparator 5 compares the threshold value with a constant threshold value from the threshold value setting circuit 4.
Or, it is binarized to 1 and stored in the output buffer memory 9. The difference between the data input to the comparator 5 and the output data is calculated by the calculator 6, and the obtained error is multiplied by the weighting coefficient shown in FIG. 5 in the weighting circuit 7 and stored in the error buffer memory 8. . The error data stored in the error buffer memory 8 is stored as a cumulative value for unprocessed pixels, and is added to the corresponding image input in the adder 3.

[0005]

However, while the error diffusion method is excellent in gradation and resolution, it has to have an error buffer memory for several lines and has a large amount of calculation, so that the apparatus is complicated and expensive. There was a disadvantage that it was.

[0006] In order to simplify the apparatus there is a method of diffusing an error as shown in FIG. 7 in a one-dimensional main scanning direction only. Since the error is distributed only in the main scanning direction, a memory capacity of several pixels is sufficient, so that the apparatus is simplified. However , as shown in FIG. 8 , the dots 81 are likely to appear with regularity perpendicular to the processing direction. , we have problems that the main scanning direction resolution not practical and not significantly higher.

[0007] Therefore, the present invention is in the error diffusion method diffuses an error an error only in the main scanning direction, the main scanning direction without occurs perpendicular stripes, high speed and low cost binarization method that can configure The task is to provide

[0008]

According to the present invention, there is provided a method for binarizing a gradation image in which an error is distributed only in a main scanning direction, wherein a threshold value is changed for each main scanning line. Equalize the difference between the thresholds of adjacent main scanning lines.
Error that is further distributed,
First for each fixed pixel or with a random number of pixels
This is a binarization method of a gradation image to be initialized .

Here, the difference between the threshold values of adjacent main scanning lines is described.
In order to make
(1)
This can be realized by changing with the function shown in (2). Th (1) = a / N (a is an arbitrary integer from 1 to N) (1) Th (l) = ((Th (l−1) × N + m) mod (N)) / N · ··· (2) N = m × n-1 (N is an integer of 2 or more) where l represents a main scanning line number, and Th (l)
Represents the threshold value of the l-th main scanning line. mod is remainder
A function representing the remainder, and therefore Th (l) is a period of period N
In the function, the numerator (Th (l-1) × N
+ M) mod (N) is any integer from 0 to N-1
Therefore, Th (l) is a number between 0 and 1.

Further, the present invention relates to an apparatus for realizing the above-described method for binarizing a gradation image, wherein the gradation is binarized by distributing an error only in the main scanning direction. In the image binarization device, the threshold between adjacent main scanning lines
Storage means for storing threshold data in which the difference between the threshold values is equal, and line synchronization control means for controlling reading of the threshold data from the storage means in synchronization with the line. . Further, the present invention relates to a grayscale image binarization device, characterized in that an initialization control means for initializing a distributed error for each predetermined pixel or with a random number of pixels is provided.

[0011]

In the present invention, the pixel arrangement is shifted for each scanning line by changing the threshold value for each scanning line, thereby preventing the occurrence of a stripe pattern. In addition, the adjacent main run
By making the difference between the thresholds of the scanning lines equal
The pixel arrangement in the halftone is noticeable low frequency noise
Is prevented , and smooth gradation expression is enabled. Also,
Since the error is diffused one-dimensionally only in the main scanning direction, the circuit can be configured with delay elements, and the circuit can be configured at high speed and at low cost.

The threshold differences between adjacent main scanning lines are equal.
To explain in more detail the effect of having
When an image with 1 / N gradation is input, the grid
It is filled up soon, and it is easy to see a cycle with the same number of gradations.
Frequency noise is prevented, and the noise is at least
It falls in the high frequency region with a smaller cycle.

[0013]

FIG. 1 is a schematic diagram showing an embodiment of the above-described error diffusion method according to the present invention. Parts similar to those in the conventional example of FIG. 6 are denoted by the same reference numerals, and description thereof is omitted. The difference from FIG. 6 is that the delay circuit 1
1 and a point that a threshold memory 12 for controlling the threshold setting circuit 4 and a line synchronization control circuit 13 are provided.

The image input 1 is temporarily stored in the input buffer memory 2 and is read out pixel by pixel. Then, after adding the error data of the processed pixels by the adder 3,
It is input to the comparator 5. The threshold is read out from the threshold memory 12 for each line by the line synchronization control circuit 13 and set. The threshold memory 12, adjacent
Stores in advance the threshold value was set to respectively equal the difference between the threshold of the main scanning lines. Threshold memory 1
The threshold value set by 2 is compared with the input data in the comparator 5, and the output binarized to 0 or 1 is stored in the output buffer memory 9 according to the magnitude.
At this time, the difference between the input data and the output data in the comparator 5 is calculated by the calculator 6 as an error. After the error is delayed by one pixel by the delay circuit 11, the error is multiplied by a predetermined weighting coefficient by the weighting circuit 7,
It is input to the adder 3. At this time, since the error is delayed by one pixel, the adder 3 adds the error to the next pixel. For distributing error only the main scanning direction, delayed by several pixels the errors, so may be added after multiplied by the weighting factor by the weighting circuit 7, necessary to provide the error buffer memory 8, such as FIG. 6 Instead, it can be configured with a delay circuit, and the circuit and operation become extremely simple.

FIG. 2 shows an embodiment of the threshold function according to the present invention. In the equation (1), m = 5, n = 4 and a = 1.
It is. This threshold function has a circumference of N = 5 × 4-1 = 19.
, Which is not described in FIG.
Is a repetition of a numerical value of l = 1 to 19. Also next to
The threshold difference between the combined lines is m / N = 5/19
It has the property of becoming a constant value.

FIG . 3 shows an embodiment of the present invention using the threshold values shown in FIG.
The result of binarizing the halftone by the light binarization method,
This is an example of a case where a 19-level halftone is input. Figure
Inside, the part shown by the broken line is one unit of repetition,
In the direction (sub-scanning direction), the threshold value is repeated at a cycle N.
Has become. In the horizontal direction (main scanning direction), the darkness of the input image
Pixels appear in a cycle that matches the degree, and in the case of FIG. 3,
1/19 the input concentration of the, to 1 pixel printed on 19 pixels
You. According to the present invention, the threshold difference between adjacent lines
Is constant everywhere, so the images placed are also at regular intervals
Smooth halftone reproduction is done.

The method of selecting m and n is arbitrary,
Make sure that the distance between the printing pixels is
Good value. The cycle N of the threshold value change depends on the number of input gradations.
Therefore, an appropriate selection is made in the vicinity of where N × N becomes the number of gradations. error
Because the diffusion is performed, the gradation of N
It can be reproduced, but if you take it small, it will be vertical stripes, if you take it large it will be
Prevent horizontal stripes from approaching, and arrange them as evenly as possible
This is so that For example, in the case of 256 gradations
N = 16, that is, the combination of m × n is 3 × 5, 4 ×
About 4, 4 × 5 is good. When the main and sub scanning resolutions are different
In this case, N may be determined according to the ratio. For example, Lord
When the resolution in the scanning direction is twice the resolution in the sub-scanning direction
Is that N × 2N is near the resolution, and for that N, m = 2N
The combination is set so that N = m × n-1 near 2n.
Measure.

FIG. 4 is a schematic structural diagram for embodying another embodiment of the binarization method of the present invention. 1 and 6 are denoted by the same reference numerals and description thereof is omitted.

This embodiment differs from FIG. 1 in that an initialization circuit 14 controlled by an initialization control circuit 15 is provided on the output side of the weighting circuit. In the apparatus shown in FIG. 1, there is no problem in the early stage of the process, that is, in the early stage of the main scanning line. However, as the process progresses, as the error accumulates, the effect of changing the threshold decreases. Vertical stripes may gradually appear. Therefore, in this embodiment, for example, the error accumulated every several pixels is cleared to 0, and the initial state where no error is accumulated is returned. Thereby, the effect of the threshold value change can be maintained in a more effective initial state. The initialization control circuit 15 controls how many pixels are initialized. Initialization period is about several input gradation may be a standard. Because
The error that is discarded without being distributed for initialization is 1 /
This is because the ratio is about (the number of gradations), and the resolution per pixel is lower than the minimum resolution of the density.

According to the present invention, an error distribution destination is simplified,
Since it can be performed only in the line direction, a binarizing device can be constituted by a delay circuit instead of an expensive error line buffer.

[0021]

As described above, according to the present invention, the processing
Processing speed and simplification of the circuit. In particular, a high-speed and inexpensive configuration can be achieved without deteriorating halftone images.
There is an effect that a value conversion method can be provided.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram that embodies an embodiment of a binarization method of the present invention.

FIG. 2 is a specific example of a threshold function in one embodiment of the binarization method of the present invention.

FIG. 3 is an output example of a halftone image according to the binarization method of the present invention;
It is.

FIG. 4 is a schematic configuration diagram embodying another embodiment of the present invention ;
It is.

FIG. 5A is an explanatory diagram of a weight coefficient of a conventional error diffusion method .

FIG. 5B is a specific example of a weight coefficient of the conventional error diffusion method .

FIG. 6 is a schematic configuration diagram of a conventional error diffusion method .

FIG. 7 is an explanatory diagram of another weight coefficient of the conventional error diffusion method .

FIG. 8 is an output example of a halftone image by a conventional error diffusion method .

[Explanation of symbols]

1 image input 2 input buffer memory 3 adder 4 threshold value setting circuit 5 comparator 6 calculator 7 weighting circuit 8 error buffer memory 9 output buffer memory 10 binary output 11 delay circuit 12 Threshold memory 13 line synchronizing control Circuit 14 Initialization circuit 15 Initialization control circuit

Claims (4)

(57) [Claims] 1. A method for binarizing a gradation image in which an error is distributed only in a main scanning direction, wherein a threshold value is changed for each main scanning line , and a difference between threshold values of adjacent main scanning lines is equalized.
Binarization method of the gradation image, characterized in that the properly. 2. The distributed error is calculated for each predetermined pixel.
2. The method according to claim 1, wherein initialization is performed with a random number of pixels. 3. A gradation image binarization device for binarizing a gradation image by distributing an error only in the main scanning direction.
Make the threshold difference between adjacent main scanning lines equal
Binarization of the gray scale image, wherein a storage means for storing the threshold data, the synchronization to and a line synchronization control means for controlling reading of the threshold data from the storage means in line with apparatus. 4. The distributed error is calculated for each predetermined pixel.
4. The binary image binarizing apparatus according to claim 3 , further comprising initialization control means for initializing with a random number of pixels.