JPS6226632B2 - - Google Patents

Description

[Detailed description of the invention] The present invention relates to a binarization device for grayscale images.

A special expression method is required to display a grayscale image using binary values of white and black. The halftone method used in photolithography achieves this by using a display method that changes the size of the dots depending on the density of the pixels. In an electronic version of this method, one output dot corresponds to one pixel of the original image, and the gray value is quantized using a threshold value predetermined according to the pixel position. There is a dither method that determines white, and a density pattern method that constructs a binary image by associating each pixel of an original image with an n×n dot matrix that pseudo-expresses its gradation value. The present invention relates to the latter device.

FIG. 1 explains this density pattern method. 100 is an original image, and 101 represents pixels of the original image. Pixels have gradation values, and the original image is expressed by these gradations. 102 is a dot matrix corresponding to this pixel. 10
4 is a black dot, and 103 is a white dot. This example is a 4 x 4 dot matrix, and the ratio of the black level to the total number of dots is 4, so assuming that the black shading value is 1, a shading value of 4/16 is displayed. In this example, 17 types of dot matrices are prepared for the dot pattern, each representing a gradation value of i/16, and a corresponding dot matrix is embedded at the position of the pixel according to the gradation value of the original image. It is good to think that it will be possible. Of course, dot matrix 17
There is no need to limit it to seeds, and there are various ways to make them. The halftone image forming device described in Japanese Patent Publication No. 54-1410 is one of this type of device.
It is a seed.

This device outputs n types of dot matrices representing the shading using n-level values obtained by quantizing the shading values of the original picture.

In the conventional method, only one point is referred to and the dots are
In order to form a matrix, it was not possible to reflect the gradient of values at that point, that is, the change in shading, to the dot matrix. For that purpose, Fig. 2
The originally smoothly changing image shown in 00 is changed to 2
When displaying the values, a border line like 201 that does not originally exist appears on the screen. Although this boundary line is originally a binary display where the degree of change is small, the dot matrix is created only from the gray values of that point. An inconsistency occurs,
This is what occurs. With conventional methods, this problem cannot be avoided, and it is also not possible to distinguish between areas of the original image with large changes and areas with small changes.

The present invention eliminates such drawbacks and improves gradation images in two ways.
The purpose is to provide a value conversion device. That is, in order to avoid this problem, the present invention makes it possible to refer not only to the shading value of one point but also to the shading values of surrounding points, and by reflecting the degree of change at that point in the dot matrix. , to remove mismatch between dots and matrices and form a smooth dot pattern as shown at 200.

The present invention includes a memory that stores the value of a pixel to be binarized and the values of surrounding pixels, a device that calculates the amount of change in shading from those values, and a device that calculates the amount of change in density based on the results.・Output a smooth dot pattern using a pattern generating device.

FIG. 3 shows an embodiment of the present invention. An input signal 301 of an original image is temporarily stored in a memory 302. The control unit 306 accesses the memory for the values of the pixel to be processed and its surrounding pixels, and the result is sent to the calculation unit 303. The arithmetic unit calculates the amount of change in the value of the peripheral portion and sends this to the dot matrix generation unit 304. Dot matrix generation unit 304
creates a dot matrix based on this quantity and outputs it to 305. In this embodiment, the dot matrix generation section 304 consists of a memory that accommodates a predetermined dot pattern, and the calculation section 303 calculates an address for accessing this memory.

To explain in detail with reference to FIG. 4, the memory 302 in FIG.
22, 323, and the values 301 of each pixel (gradation values, here digital values) obtained by scanning the original image are sequentially supplied. Each shift register has a number of bits equal to the number of pixels corresponding to the scanning width of the original image. The controller 309 sequentially stores adjacent 9 bits of the contents of these shift registers into the register 330 in the arithmetic unit 303. When the values of these pixels are indicated by a, b, c, ..., h, i, e in the center is the pixel to be binarized, b is the pixel above pixel e, and d is the pixel above pixel e. This is the pixel to the left of .

The calculation unit 303 further includes change amount calculation units 331 and 33.
2 and an address generation section 333.

The change amount calculation unit 331 calculates the horizontal gradient in the vicinity of pixel e, and calculates the value X expressed by the following formula.
seek.

X=(a-c)+(d-f)+(g-i) Also, the change amount calculation unit 332 calculates the gradient in the vertical direction near the pixel e, and calculates the value Y shown by the following formula. demand.

Y=(a-g)+(b-h)+(c-i) The obtained values X and Y are supplied to the address generation section 333.

The address generation section 333 uses the value e in the buffer 330 and X, Y to generate address data for the dot matrix generation section 304 consisting of a dot pattern memory. The simplest method is to create data in which the value e is the upper value, X is the middle value, and Y is the lower value. The generated address data is supplied to the dot matrix generation section 304 under the control of the controller 306. Therefore, dot matrix generation section 304 can output dot matrix data corresponding to this address onto line 305.

The dot matrix generating section 304 has a plurality of types of halftone dot matrices as dot matrices of the same density. These halftone dot matrices of the same density differ depending on the values of X and Y mentioned above. For example, FIGS. 5a and 5b are dot matrices representing halftones at a density level of 0.5. Of these, Figure 5a is a halftone with zero gradients in the horizontal and vertical directions, and Figure 5b is a halftone with a gradient in the horizontal direction (turns black to the right) and no gradient in the vertical direction. be. In this way, the dot matrix generation unit 304 has a plurality of types of halftones depending on the gradients in the horizontal and vertical directions, and which dot matrix to select representing halftones of the same density depends on the above X and Y. It is determined accordingly.

As described above, according to the present invention, binarization is performed not only based on the value of the pixel of interest, but also based on the values of surrounding pixels. The image becomes smoother and the resolution can be improved.

In the above embodiment, values of peripheral pixels adjacent to the pixel to be processed are used, but values of pixels that are not necessarily adjacent may also be used. In addition to the horizontal and vertical gradients, other variations may also be used.

[Brief explanation of the drawing]

FIGS. 1 and 2 are diagrams showing the prior art, FIGS. 3 and 4 are diagrams showing an embodiment of the present invention, and FIGS. 5a and 5 are diagrams showing dot matrix data representing halftones. be. 302...Memory, 303...Arithmetic unit, 304
. . . dot matrix generation section, 306 . . . controller.

Claims (1)

[Claims] 1. An input means for scanning a grayscale image and sequentially inputting a signal representing the grayscale value of each pixel; and input means for sequentially inputting a signal representing the grayscale value of each pixel; Calculating means for calculating the amount of change in density near the specific pixel from the gray value; and dot matrix generating means for outputting a binarized dot matrix corresponding to the gray value of the specific pixel input from the input means. The dot matrix generating means stores in advance a plurality of types of halftone dot matrices that are pseudo-expressed according to the amount of change in density for the same gray value, and corresponds to the gray value of the specific pixel. A binarization device for a grayscale image, characterized in that a halftone dot matrix corresponding to the amount of change in density calculated by the calculation means is selected from among a plurality of types of halftone dot matrices.