JPH01147779A - Method for processing picture data - Google Patents

Abstract

PURPOSE:To convert the number of scanning lines by generating intermediate picture data by dividing original picture data scanned by N-lines of the scanning lines by the scanning lines of M-lines per one scanning line, and further, converting it into final picture data by applying NM-lines of the scanning lines for L-lines of the scanning lines. CONSTITUTION:The original picture data 20 formed of 6X6=36 dots is divided again into 4 times both in height and width, and the intermediate picture data 30 of 24X24=576 dots is generated. Further, 3 lines of the scanning lines of the intermediate picture data 30 are lumped into one new scanning line. Thus, 9 dots of the intermediate picture data 30 are converted into one dot of the final picture data, and the final picture data 40 of total 64 dots is obtained. Then, optional one dot at every area corresponding to one dot of the data 40 and LXL dots of the data 30 is selected by using a random number. Besides, when L used for converting the data 30 into the data 40 does not go to an integer, the scanning lines of the number nearest to a calculated value are selected. Thus, the number of the scanning lines can be converted while reducing the deterioration of quality.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method of processing image data when creating image data such as figures and characters using an overall raster scanning method. Specifically, the original image data created with a certain number of scanning lines can be arbitrarily changed to final image data with a different number of scanning lines without degrading the quality of the image.

There are various ways to use the raster scanning method to create a copy of an image used for printing, etc., but one of the most commonly used methods is a method that scans each character one by one and records them on a medium. This method is widely used in what is known as a CRT phototypesetting machine. Another method is to scan and record the entire page at once, like a scanner.

This method is used in laser phototypesetting machines that have recently started to be used, and has the great advantage of being able to treat pictures and figures as if they were text. However, it also has the disadvantage that the number of scanning lines of the created image data cannot be changed freely.

The present invention aims to significantly improve the usability of image data by making it possible to freely change the scanning line density without impairing image quality in an image data processing method that scans the entire page at once. It is.

[Prior Art] The problem with the full-screen/batch scanning method is that when one image data is input to output devices having different scanning line densities as described above, images of the same size cannot be obtained. Laser printers are generally used in phototypesetting machines using the full-page scanning method, but the scanning line pitch of laser printers is unique to each machine and cannot be freely changed. Therefore, if there are two printers with different scan line pitches, inputting the same data to each printer will produce output of different sizes. For example, if you give the same image data to a laser printer with a scanning line density of 300 dots/inch and a printer with a scanning line density of 600 dots/inch, the output size of the latter will be half that of the former, and if left as it is, it will remain the same. It is not possible to output an image of this size.

Conversely, printers with the same scanning line density cannot output images of different sizes using the same image data. As described above, it is basically impossible to enlarge or reduce the figure in the full-batch scanning method.

[Problems to be Solved by the Invention] The reason for the above is that in the full-page scanning method, every part of the page to be output is scanned at the same scanning line density. On the other hand, in the case of a partial scanning type CRT typesetting machine, any part of the same page can be
Because it can be scanned at a different line density than the rest of the page, the scaling problem described above does not occur.

If the number of scanning lines of image data can be converted in this way in the full-batch scanning method, the above-mentioned problems can be solved.

Conventionally, in such cases, a method known as ``thinning'' has been used, in which some scanning lines or dots are appropriately discarded depending on the rate of reduction. When enlarging an image, a method has been adopted in which appropriate scanning lines are copied and duplicated depending on the enlargement ratio.

However, when attempting to deal with the problems of the full-batch scanning method in this manner, the following quality problems arise.

1. Simple thinning causes part of the data to be lost, resulting in significant quality deterioration.

2. If data is simply added repeatedly, so-called "jaggies" in the image become noticeable.

3. Mechanical data thinning or repeated addition methods may distort the image or make the size inaccurate, unless the number of scanning lines is changed to an integral multiple or fraction of an integral number. .

4. In halftone dot images WA (flat dots and photographs), noticeable moiré occurs when thinning is performed.

[Means for Solving the Problems] The present invention was made to solve the above problems, and it is possible to arbitrarily increase the number of scanning lines of an image without causing quality deterioration as described above. The object of the present invention is to make it possible to change the number of scanning lines, that is, to make it possible to convert the number of scanning lines while using the full-batch scanning method.

In order to achieve this object, the present invention subdivides the data of an original image scanned with N scanning lines into M scanning lines per scanning line, thereby creating an intermediate image of NM scanning lines. Create data, convert the NM scanning lines of the intermediate image data into final image data having NM/L scanning lines by applying L scanning lines to each scanning line, and separate the original image data and another scanning line. It is characterized by being able to reproduce final image data having numbers.

Furthermore, for each area corresponding to one dot of the final image data and the L×L dot of the corresponding intermediate image data, there is a means for selecting any one dot therein using a random number, and the selected one dot is The method is characterized in that the corresponding dot in the converted final image is determined to be white or black depending on whether the dot is white or black.

Another feature is that when L when converting intermediate image data to final image data is not an integer, the number of scanning lines closest to the calculated value is selected.

[Example] FIG. 3 shows a state in which a certain image 20 is scanned.

For ease of explanation, this image is scanned with six scanning lines R1 to R6 in each of the X and Y force directions, but actual image data is scanned with a large number of thousands of scanning lines.

Now, each scanning line R represents the entire image by dividing it into 6×6=36 dots. In this image 20, the hatched portion is an image portion that becomes a black image.

Next, a method of converting the number of scanning lines according to the present invention will be explained based on the embodiment shown in the drawings.

FIG. 4 is a diagram for explaining the present invention in detail, in which the image data as explained in FIG. 3 is used as original image data 31,
By temporarily creating intermediate image data 32 from this data and processing this intermediate image data by a method described in detail later, final image data 33 having a different number of scanning lines from the original image data 31 is created.

In order to convert the number of scanning lines, first, the dot plane of the original image as shown in FIG. 3 is simply subdivided into multiples in both the vertical and horizontal directions. FIG. 5 shows intermediate image data 30 of 24×24=576 dots obtained by redividing the original image 20 of FIG. 3, which is made of 6×6=36 dots, into four times the size in both the vertical and horizontal directions. Since this intermediate image has been subdivided four times in both scanning line density and dot density, each one dot in the original image 20 is replaced by 16 dots in this subdivided intermediate image.

As is clear from comparing Figures 3 and 5, the original image 20 and intermediate image 30 differ only in the number of scanning lines; the images themselves have not changed at all, and both are different in terms of quality. are the same. Therefore, the original picture (1120
If a dot is white, all 16 dots in the corresponding intermediate image 30 in FIG. 5 will be white; if a dot in the original image is black, all 16 dots in the corresponding intermediate image
All dots also turn black. In this way, no quality degradation occurs in the intermediate image 30.

Final image data with a target number of scanning lines is generated from this intermediate image data, and this process is called scanning line number conversion.

Figure 6 is an example of the final image obtained by converting the number of scanning lines.
Original image 20 consisting of X6 = 36 dots
On the other hand, the final side a40 has 8×8=64 dots, which is about twice the number of dots. This conversion converts image data created for a printer with fewer scanning lines (original image data) to a printer with 4/3 times the scanning line density (final image data). This is an example of performing a conversion.

In order to convert the intermediate image shown in FIG. 5 into this final image, several scanning lines of the intermediate image are combined into one scanning line. In FIG. 6, the three scanning lines of the intermediate image data in FIG. 5 are combined into one new scanning line. As a result, 9 dots of the intermediate image data are converted to 1 dot of the final image, resulting in a final image of 64 dots in total.

The black and white of each dot in the final image newly created by this conversion is determined by the black and white of the nine dots of intermediate image data that make up that dot. This conversion from intermediate image data to final image data is performed based on the following principle.

First, if the nine dots in the original intermediate image that make up the new dot after scanning line number conversion are all white or all black, then
It is uniquely determined as either white or black. The dots shaded to the right and the white dots in FIG. 6 are all black or white dots in the intermediate image, and are determined to be black or white as they are. For example, the intermediate image data (Lxl, Ly4) of dot Al in FIG.
.

. . . The 9 dots of b9 are all black, so the dot A1 in FIG. 6 is determined to be black with diagonal lines to the right. Similarly, the nine dots d and -d9 in FIG. 5 that constitute dot A2 (LXl, Lyl) in FIG. 6 are all white, so they are determined to be white.

However, the other dots indicated by diagonal lines on the left in Figure 6,
The 26 dots C1 to G2G are gray dots.

A processing method for reproducing such gray dots will be explained. First, if the printer can handle halftone densities, it can simply represent them using nine different gradations. Second, if the printer can only handle binary values, it must be decided between white and black. The simplest and most widely used method is to determine whether the original constituent dots, that is, the dots in the intermediate image data, are white or black, whichever is more. For example, dot C9 in Figure 6 is LXl in Figure 5.
, 'Since it corresponds to LV3, e! ~e6 white 6 pieces, 0
The black dots 7 to e9 are made up of three intermediate image dots, and since there are more white dots, white is determined. Similarly, dot C4 in FIG. 6 is Lx3. Since it corresponds to Ly2, it is composed of three white intermediate images f and -f3 and six black intermediate images f4 to f9, and since there are more black images, black is determined.

However, with this majority vote method, quality is likely to deteriorate during the process of deciding between black and white. For example, even if the ratio of black and white is almost the same, but only one black dot is missing, that dot will be determined to be white. If such cases continue to occur, a series of consecutive dots will all turn white even though half of them are actually black, resulting in a lack of fidelity to the original image.

In order to avoid such defects, the present invention uses a probability-oriented method to process gray data. FIG. 7 explains the principle, and the nine dots from el to C9 represent dot C9 in FIG. Among these, C7
, ef3, and eg are black in the intermediate image of FIG. The remaining six dots el to e6 are white. According to the majority voting method described above, this 09 dot is naturally determined to be white, which has the largest number, but in the method of the present invention, any one dot is selected from among the nine dots without making a judgment as to which dot has the largest number. If the dot is white, it is determined to be white, and if it is black, it is determined to be black. That is, one of the sample dots among the constituent dots of the intermediate image is made to represent the entire image, and the dots of the final image are determined by that dot. The problem is which dot should be used as a representative sample dot, but in the present invention, an arbitrary dot is selected using random numbers.

FIG. 1 is an example of a flow diagram showing the method according to the present invention described above, and FIG. 2 shows an example of an image data processing apparatus suitable for implementing the present invention.

In FIG. 2, the original image data 1 as shown in FIG. Step 1 in Figure 1). In the example in Figure 5, the original image data of 6x6 dots is
This is an intermediate image that is re-divided and expanded into 4x4x24 dots.

CPtJ2 transmits a signal instructing the size of the final image to the X-side latch 6 and Y-side latch 7 via the signal line 5 (step 2 in FIG. 1). That is, a command is given as to how many scanning lines of the intermediate image are to be combined into one line. This command instructs conversion of the number of scanning lines, and is performed either by being included in the original image data 1 in advance or by direct command to the CPU 2. In the example in Figure 5, L=3 for both X and Y.
In this case, the two latches 6.7 each latch 3. Adders 8 and 9 accumulate the data latched in the latches 6 and 7, and generate the origin address of each dot that makes up the final image data (Step 3 in Figure 1).
.

FIG. 8 explains this origin address. The square mark 70 shown in the figure is the origin address of each dot that makes up the final image data expressed by lx and LV, and the circle mark 71 is the origin address indicated by a random number. It shows the address of the sample dot. As mentioned above, the first stage adders 8 and 9 are connected to the latch 6.
．． The data ``3'' latched in ``7'' are sequentially accumulated to calculate the origin address of each dot constituting the final image.

The result is transmitted to addition 4110.11. This addition 11
At 10.11, the random number generator 12 receives a signal from the latch 6.7 and transmits a random number (Step 4 in Figure 1).
), whereby the adder 110.11 calculates the address of the sample dot indicated by the circle 71 position in FIG.

For example, in FIG. 8, the random number generator 12 has xn and y.
By specifying the value of n, relative addresses within the 9 dots that form the final image are randomly generated, and the address of the sample dot is designated. Therefore, the absolute address of the sample dot is determined by the coordinates of the origin address 70 and the random number (xn, yn) from the random number generator 12.

Depending on whether the designated sample dot is white or black, the black and white of the corresponding dot in the final image is determined. For example, gray dot C4 in FIG. 6 corresponds to LX3. Consists of 9 dots, f and -f9, which correspond to LV2, of which f
1 to f3 are white, and f4 to f9 are black. On the other hand, the 8th
At lx3゜lj/2 in the figure, the sample dot is at position 71p.
This position corresponds to f8 in Figure 5, and f8
Since C4 is a black dot, the entire dot C4 is determined to be black.

As described above, each gray dot 01 to 026 is judged, and the result is serial/parallel converted by the S/P converter 15 in Fig. 1 according to the data read signal (step 6 in Fig. 1). The data is assembled into code units (first
Figure step 7) will be output by the printer 16.

In the examples described above, the ratio of the scanning lines of the original image to the converted scanning lines is an integer, but generally this ratio is often not an integer.

For example, the dot in Figure 6 is made up of 8 scan lines that are correctly partitioned into the 3x3 dot in Figure 5, but if we try to convert this to a final image with 10 scan lines, we get
One scanning line in the final image corresponds to 2-04 lines in the intermediate image. Therefore, even if the conversion is performed so that there is one scanning line in the final image for every two scanning lines in the intermediate image, a fraction of 0.4 will appear in the calculated value. In the present invention, such a case is handled by adding up the fractions, although this is not shown in FIG. First keep 0.4 of the first scanline, then again scanline 2 of the intermediate image.
If the main part is converted into one line in the final image, the error will be 0.8. Furthermore, when the following conversion is performed, the error becomes 1.2. If the error exceeds one intermediate image, correction is performed. That is, the conversion ratio of the scanning lines of the intermediate image and the final image is 2:1 in this case.
Set to 1. Then, the accumulated error of fraction 1.2 is 0.
2 and is corrected.

From now on, we will add a fraction to these 0 and 2 and perform the same process, but if there is an error of more than 1 line between the number of lines converted by calculation, we can increase the number of scanning lines by 1. .

[Effects of the Invention] According to the present invention, quality deterioration is small even when the number of scanning lines is converted, and even when the conversion ratio is not an integer as in thinning or repeated addition methods, It can be converted easily and accurately.

Another excellent feature of the present invention is that moiré is less likely to occur because random numbers are used for conversion, and the quality can be maintained even if a photograph with halftone dots is converted to the number of scanning lines.

[Brief explanation of the drawing]

FIG. 1 is an example of a flow diagram illustrating the method according to the invention;
The figure shows an example of an image data processing device suitable for carrying out the present invention, FIG. 3 is an explanatory diagram showing an original image of 6×6 dots, and FIG. 4 is a diagram illustrating the present invention in detail. ,
Figure 5 is an explanatory diagram of a 24x24 dot intermediate image created based on Figure 3, and Figure 6 is an 8x8 intermediate image of Figure 5.
An explanatory diagram of the final image created by converting the number of scanning lines to dots,
FIG. 7 and FIG. 8 are explanatory diagrams. 1 Original image data 2 CPU 4 Memory 6 Latch
7 Latch 8 Adder ell 9 Adder 10 Adder m 11 Adder 12 Random number generator
13 Multiplexer 14 Multiplexer 15
S/P converter 16 Printer

Claims (3)

[Claims] (1), the original image data scanned with N scanning lines is redivided into M scanning lines per scanning line, and N
Create intermediate image data with M scanning lines, convert NM scanning lines of the intermediate image data into final image data with NM/L scanning lines by applying each L scanning line to one scanning line, A method for processing image data, characterized in that final image data having a different number of scanning lines from original image data is reproduced. (2) For each area corresponding to one dot of final image data and L×L dots of corresponding intermediate image data,
It has means for selecting any one dot from among them using random numbers, and depending on whether the selected one dot is white or black, the corresponding dot in the converted final image data is changed to white or black. The image data processing method according to claim 1, wherein the image data processing method is determined. (3) When converting intermediate image data into final image data, if L is not an integer, the number of scanning lines closest to the calculated value is selected. Processing method.