SU1356256A1 - Method of transmitting and receiving raster images - Google Patents

Abstract

The invention relates to the technique of facsimile transmission of images via the communication channels. The purpose of the invention is to improve the quality of the adopted raster invention by reducing the intensive re-screening moire. After counting the averaged number of black zl-ts K, the numbers K 1 K, black el-s, which are located at a distance from the corresponding raster cell border in the direction perpendicular to the raster cell boundary, are counted at 1 K. The number K is black. el-tov determined by f-le K K + 1/4 .. 6 Il. SAE SP 05 to ate 05

Description

The invention relates to telecommunications, in particular to the technique of facsimile transmission of images over communication channels, and can be used, for example, in the transmission of newspaper illustrations.

The aim of the invention is to improve the quality of the received raster image by reducing the intensity of the moire re-rasterization.

FIG. Figure 1 shows a fragment of the registry image; figure 2 - raster cell; in fig. 3 shows a video signal of a raster image; FIG. 4 - line of the raster image, content; and the main and secondary raster cells; in fig. 5 - a set of main raster cells (solid lines) and side raster cells in the vicinity of one of them (dashed lines); in fig. 6 is a structural electrical circuit of the device implementing the method for transmitting and receiving raster images.

SRI.

Accounting for black elements containing c in side raster cells allows the proposed method to almost completely compensate for the re-screening moire while maintaining the same amount of information transmitted over the communication channel.

To clarify the essence of the proposed method, we will consider the CTpyKTvov raster image. Its enlarged fragment is shown in FIG. 1. The raster image is a collection of raster black dots on a white background (Fig. 1a). The halftone difference is achieved by changing the size of the raster dots, and the perception of the halftone depends on the shape of the dots. The regularity of the raster structure makes it possible to divide the entire image into raster cells (Fig. 2a), each of which contains one raster dot. To reproduce on the receiving side of the copy of the raster image, it is sufficient to transmit through the communication channel the area of each raster point and to build at the reception a point of the required area and suitable shape. To compensate for possible loss of sharpness due to the inaccuracy of the selection of raster cells, it is necessary to change the area of each raster point taking into account the ploidy of the neighboring points. This is the essence

method of transmitting and receiving bitmap images, which is a prototype of the proposed method.

However, under real conditions, the dimensions of the raster cell of the original raster and the cell artificially extracted in the signal practically never coincide, which leads to appearance of copies of moirés on the copy.

Consider a one-dimensional model of the moire effect on a single scan line. Let d be the raster step in the sweep direction, and d be the length of black dashed sections of raster points with the sweep line (Fig. 3),

After the image we assume homogeneous - in such a field the moire is most noticeable. Then the scan line can be represented in the form of a Fourier series (Fig. For) - ..

 D - sin -2 - S (X) Z ;, n, -oo ndo d

e (l)

The discretization of the function s (x) can be represented as multiplying it by a sequence of α functions that are separated from each other by the magnitude of the discretization step p (the size of the resolution element). Denote this function f (x) (Fig. Зб)

daso

f (x) ZIc / (x-l) - -Z e P. (2)

e .-- co r s

The sampled scan line can be written as

"Nd about sin ---

G

but

. p-1

 -with. BSl d

t- H

: -with

U)

Sampled signal spectrum

contains the r-armonics of the original signal at a discretizing function f (x) at and harmonics with combination spatial frequencies

n ,, 2,. ,,, (4)

(five)

their corresponding wavelengths, i

 -: j :: r

When where p is an integer, the frequencies of all combinational harmonics coincide with the harmonic frequencies of the original signal. When d - (N + r) p5 where 0

may cause harmonics with

one

That's what it is

frequency lower than g

but

moire harmonic sampling. Most noticeable and corresponds to them, and has a wavelength

d

N, -1 G.

;

(6)

Its relative amplitude (with respect to the constant component) is equal to

, Fndo

sin y

but

n, l / nd.

(7)

flNdo d

At r4) there is no moire, but it is impossible to accurately fulfill this condition, and small deviations give a moire with a sufficiently large, according to formula (6), wavelength, clearly visible to the eye. The moire of this type can be suppressed in a known manner, consisting in the choice of the sampling frequency — in such a way that the size of the raster cell is oddly divided to half the size of the scanning beam aperture. At the same time, the moire wavelength is equal to 2d, and since d is rather small, the moire is effectively integrated with the eye, which is used to reduce the sampling rate of moire.

Re-rasterization can be represented as a sequence of two operations. The first is to calculate the average of the raster cell (in the one-dimensional model over the period of the secondary raster d) of the black area; the second is in the formation of the corresponding raster point. Since d 4 lp, the moire can appear only during the first operation.

After the execution of the first operation (in the one-dimensional model), we obtain a sequence of values of the area of

() c /

K (ha) - j s. (X) dx, (8)

mc (

which can be expressed as a function of the continuous x x + d + coordinate

It can be seen from formulas (6), (P-14) that the re-rarring moirh wavelength is N times (where N is the number of exploitation elements that fit the length of the raster cell) longer than the moiré sampling wave; and the choice that reduces the wavelength of the moire discretization and its amplitudes leads to the fact that the wavelength of the repeated rasterization wave makes it well noticeable.

Select in line for each races

  s square cell with origin coordinate

K (x) t J Sj (x) (x-md) (9) 55 md (fig. Yes) side raster

(nr-cft

cell in (d +) (Fig. 4b). The corresponding kpCx function (taking into account the fact that the функ -functions are taken at previous points

The integral in the formula (9) is calculated by the age-old integration of the series (3) j the sum of the c -functions is represented by

1356256.

similar to the expression tate can be obtained

(2). As a result

ten

"Sin. sinFCBn- i) d

    d p

K (x) „l. / - rТ -

.iH,

f 1) t.: t; 1 114,1.4.) x

, e - - d P d

dpdP-- at

irndc d

,

(ten)

15

the function K (x) - a sequence of black area values in raster cells - contains harmonics with

frequencies

J + 1 + ™

 t T -

d p d

n, l ,, ± l, ± 2 ..

(P)

20

25

. Among them are the following long-wave (D 7 d) moire harmonics:

.,. - moire sampling with a wavelength of (6) {its relative amplitude is equal to,

thirty

1g

(12)

- the moire of repeated screening, its wavelength

  B

and,--

oh -1

and relative amplitude dr

S

1,0, - TTNd

(13)

(14)

35

It can be seen from formulas (6), (P-14) that the repeated-rasterization moire wavelength is N times (where N is the number of 45 exploration elements placed in the length of the raster cell) greater than the sampling moire wavelength; and the choice that reduces the moiré sampling wavelength and its amplitude causes the re-rasterization moire wavelength to make it well noticeable.

Select in a row for each plant

50

cell in (d +) (Fig. 4b). The corresponding kpCx function) (taking into account the fact that the функ-functions are taken at previous points

1356256

md to simplify further operations)

  sin (B + i) d t (v) 2 dp

4) dnd-z TTnd, :: r: :: rh -CD

P- H (S + i) d d p

The amplitudes of the harmonics of the functions K (x) and, (x) are the same, the phases differ by 10

 jt (n, l) - (2 -t- i) d. For lower moire harmonics

4 (/ d

Therefore, the function

sin

II nd about co-jIT

d iii

.-w 7 (2 + l) d

d P

e

(W - b | 1i (d

d p

From this it can be seen that the amplitude of the moiré harmonic K, j (x) differs from the amplitude of this harmonic K (x) by the factor

k - lower side cell (fig. 5d).

A device for implementing method 30 for transmitting and receiving raster images) and a scientific research institute contains a block of 1 RAM, a block 2 for sampling and counting black (18), an arithmetic unit 3, a block 4

sampling and recording 4, block 5 constant i. the moire amplitude of the moire decreases to 35 memory, block 6 of the RAM

G 2cos || (2.i) (f I -i d p J 2. a

 1H, 1 N

is coming.

A similar conclusion can be repeated by highlighting the side raster cell on the other side of the main one. Such arbitrariness in the choice of direction is not justified by anything, and for symmetry the value of K should be calculated taking into account the side cells on both sides

K, i (K + ISiiilSil) ,. (nineteen)

The transition from one-dimensional model to the image, we obtain, taking into account both coordinates

1 n / k i (K + -1-,

+ kn, + k.

N

 /

(20)

k - the left side of the cell

 (Fig. 5a); kn - right collateral cell

one

. (Fig. 56);

kf, is the upper by-pass cell (Fig. 5c);

-2 ()

e. (15)

 -co

K (x). (sixteen)

It has a small amplitude of the re-rasterized mu harmonic. The constant component, carrying information about the image, remains unchanged.

From expressions (10) and (15) you can get

d

d p

2cos

(17)

k - lower side cell (fig. 5d).

(Fig. 6). .

Discretized raster image signal from facsimile

40 of the apparatus is written line by line into block 1 of the RAM with a volume of 3 n scan lines, where n is the number of scan lines in the raster one. Thus, in RAM

45 information about three rows of raster cells is placed. The black sampling and counting unit 2 selects from the RAM the value of each raster cell element located in

5Q is the average (of the three recorded in the memory) a row, of the main raster cells surrounding it, as well as of the side raster cells, as shown in FIG. 5. Calculate the values of K.

55

for all major raster cells and k for side. Arithmetic unit 3 calculates the K value of the center raster cell. according to the formula (20) for transmission over the communication channel. On

On the receiving side, the K code goes to block 4 of the sample and record, which selects from the permanent memory block 5 the values of all elements of the raster cell containing K black points, and records the main memory block 6, from which the signal goes to the fax machine.

Claims (1)

Invention Formula A method of transmitting and receiving raster images, in which the transmitted raster image is line by line divided into elements with a size equal to the scan beam aperture, each element is converted into a binary symbol; image, and the horizontal size is equal to the horizontal step of the image raster, count the number K of black elements in each raster cell and for each raster cell count yvayut average number of black elements adjacent to it by raster cells formula   S 0 five 0 where is the K- (i, 2, ..., N) bond the number of black elements in N countable raster cells, the number of black elements five transmit to channel K elements, and on the receiving side, the bitmap image is restored using the received binary symbols, I differ by the fact that, in order to improve the quality of the bitmap image due to a decrease in the moiré rate of the repeated rasterization, K, Kj, K, K of black elements, spaced from the corresponding border of the raster cell at a distance not exceeding one half of the raster step in the direction perpendicular to the corresponding border of the raster cell, and the number K of black elements is given by the formula Definition K + | (K, + K2 + Kj + K) -K ® in ® M W S-a-t-vg siB & SBBissm, ssmBsse @ "si evtvtyue smmmssss se isBsmmssBms. and t is s s s s s s easssisssm. mamBsmesm em No. Av1ttvtshsh sssessBBsm t  sm msmmm - t t t in s shtttash . j in & 0s s in t s & mmmssBA mss msBS-- seimssaas s n msmsBmsB tatvshvvshvs ssmmssmmismm tvtvaaev u KBBSSSsas a Its-BS amssse amsgg siSseB 1 in aa s 0 in "shchwashvvvv e S in t S and S S A figure 1 ffjue.Z soo " Uiiiiiiiiiiiiiiiiiniiiiiiniiiiiiiiniiiiniiniiiiiiiiiiiiiiiimiiimiiniinin dzie.Z .but V / R one a-g one-. .J FIG. five Compiled by A.Ivanov Editor T.Parfenova Tehred M.Hodanich Order 5814/56 Circulation 636 Subscription VNIIPI USSR State Committee on cases of inventions and discoveries 113035, Moscow, Zh-35, 4/5 Raushsk nab. Production and printing company, Uzhgorod, st. Project, 4 x jieL Proofreader L. Patay