DE2720079A1 - PROCESS FOR IMPROVING IMAGE QUALITY IN FACSIMILE TRANSMISSION - Google Patents

Description

Dr.-Ing. Rudolf Hell GmbH Kiel, March 23, 1977

Grenzstrasse 1-5, 2300 Kiel 14 Lf / Uck

Jl

Patent application No. 77/446 27 20079

Password: impulse formation

Process for improving the image quality in facsimile transmission

The present invention relates to a method for improving the image quality in facsimile transmission according to the method described in the preamble of claim 1 mentioned Art.

In facsimile transmission technology, e.g. for office fax machines, weather map transmission or transmission of other documents via Lines or radio, the image quality of the re-recorded image suffers due to imperfections in the transmission channel and the limited available bandwidth.

A common case is that only black and white information be transferred, what with documents, line drawings or the like. Is given. The image signal obtained from the original by electro-optical scanning is a so-called one 2-level signal, i.e. the signal only assumes 2 states, which signal in black or white depending on the agreement. If there are short black and white stretches within the template, see above the scanner also delivers very small, short pulses, which, however, are falsified in the course of the transmission on the transmission channel and no longer recognizable on the recording side Amplitude arrive, which is why their information is lost.

80984S / 0389

If you make a corresponding reinforcement on the receiving side, In order to raise these signals above the detection level, a number of errors occur, echoes or other disturbance impulses falsify the image information, which is why this measure is not suitable.

In today's transmission technology, transmission bandwidth is often used To save or to increase the transmission speed, a method for bandwidth reduction applied, in which the 2-level signal in converted into a 3-level signal and transmitted as a 3-level signal will. This method is described, for example, in US patent 3,303,284 or in German patent 1762644.

The 3-level signal is obtained from the 2-level signal by when the information changes, i.e. each time the 2-level signal changes, the signal is flipped over, resulting in 3 levels. So the signal values - 1 and O occur, depending on the agreement - 1 can be assigned to the status black and 0 to the status white of the original or vice versa.

Due to the halving of the basic frequency achieved with the 3-level signal, the resulting signal can either be halved Bandwidth or twice as fast.

On the receiving side, the signal is converted back into a 2-level signal recovered and recorded.

809845/0389

However, it has also been shown with this method that short black / white lines of an original that are passed through the scanning diaphragm have already been reduced in their amplitude, or lines that are broad, but only one that is below the threshold Have amplitude, are lost because the amplitudes of these signals are insufficient, a flip of the 2-level signal in to effect the 3-level signal.

The present invention is therefore based on the object of eliminating these disadvantages and improving the image quality of the transmission to improve.

The invention achieves this through the features specified in the characterizing part of claim 1.

Advantageous further developments are given in subclaims 2 and 3.

The present invention admittedly differs from that in transmission technology use of the known means of twofold differentiation of the image signal for contour division, but a decisive one The difference is that the aim here is not to sharpen the contours of the signal during transmission, but rather to that the complete information of the original is transmitted, can be clearly recognized and recorded at the receiving location.

In the known method of edge pitching, as e.g. is described in German patent specification 2 114 149, goes during the transmission from the scanning to the recording side the information of very short black / white lines or gray lines - as described at the beginning - is lost, which is why the information in this patent solution described is not sufficient to achieve the aim of the present invention. 809845/0389

The invention is illustrated below with reference to FIGS. 1-11 explained in more detail.

Show it:

FIG. 1 shows a basic circuit diagram for the transmission of a facsimile signal.

FIG. 2 shows an example of the application of the method according to the invention with 3-level modulation.

FIG. 3 shows a circuit for the pulse division.

FIG. 4 shows a schematic representation of the signals occurring in FIG.

FIG. 5 shows a circuit for converting a 2-level signal into a 3-level signal.

FIG. 6 curves for the known conversion of the 2-level into a 3-level signal.

FIG. 7 shows a schematic representation of signals as they occur in the transmission according to the invention.

FIG. 8 shows an example of the transmission by means of amplitude modulation.

B09845 / 0389

FIG. 9 shows an example of the transmission by means of frequency modulation.

FIG. 10 shows a circuit for demodulating the signal at the receiver.

FIG. 11 shows a schematic representation of the signals on the demodulator and at the demodulator output.

Figure 1 shows schematically a template 1, which is from a light source 2 is illuminated and by a schematically illustrated scanning element, which consists of an optical system 3, a scanning diaphragm 3a and an opto-electrical converter 4 is scanned. The generation of the relative movement between the original and the scanning element was not shown in the present case because it is not the subject of the invention, but the general state of the art.

The image signal passes through a pulse shaper 5 and a Modulation unit 6 to a transmission channel 7, which can be a line or a radio transmission link. A demodulator 8 is provided at the receiving location, which receives a video signal the line 9 which is recorded as a facsimile signal.

FIG. 2 shows an example of the application of the method in 3-level modulation. The signal coming from the scanner 4 arrives to a pulse shaper stage 5, which corresponds to stage 5 in FIG. 1, and is transferred from there to a 2-level / 3-level converter 10 given, from which it arrives at a modulator 6. Depending on the application, this modulator can be an amplitude or

80984B / 0389

- ir-

be a frequency modulator. The signal arrives at the transmission link 7 from the modulator 6.

As in Figure 1, a demodulator 8 is provided on the receiving side, which is followed by a 3-level / 2-level converter 11, which supplies the facsimile signal on line 9.

Figure 3 shows an example of the internal structure of the pulse former 5. In the present case, use is made of the two-fold differentiation of the image signal, which is known per se, wherein the twice differentiated signal with the opposite sign is fed back to the image signal.

From the scanner 4, the image signal passes via an RC element, which consists of the capacitor 12 and the operational amplifier 13 with the feedback resistor RK ₁₃ , to a second RC element, which consists of the capacitor 14, the operational amplifier 15 and the feedback resistor RK ₁ - composed, via a resistor 16 to the negative input of an operational amplifier 17, the image signal is differentiated once at the first RC element and a second time at the second RC element. At the operational amplifier 17, which is connected to the feedback _{resistor RK 17} , the negative, twice differentiated image signal and the image signal arriving via line 18 are added and sent to the 2-level / 3-level converter 6 shown in more detail in FIG.

FIG. 4 shows the curves a and b, the curve a being the signal supplied by the scanner 4 and the curve b the image signal generated with the aid of double differentiation at the output of the operational amplifier 17 represents.

Instead of the double differentiation, the known method can be used for scanning the surroundings, which also results in a division at edges that are parallel to the scanning direction, results, which leads to the same signal curve b, in which the weak signal trains are raised to such an extent that they are unambiguous are recognized as lying above the threshold value black, which is representative of a change in the information in the original.

FIG. 5 shows a circuit structure for a 2-level / 3-level converter with a downstream modulator 6, from which the modulated signal is passed on to the transmission link 9 will. As described above, both an FM and an AM modulator can be used as the modulator. Such Modulators can be purchased and are therefore not shown in detail as a circuit.

The 2-level signal emitted by the pulse shaper stage 5 arrives Via the resistors 20a to a comparator 20, which consists of a differential amplifier at the input and an output stage which emits only 2 states OV or + 5V as output signals, there is one at the other input that determines the decision threshold Voltage of, for example, 4.5 V is. The voltage divider from the resistors 20 a and 20 b has the

S098A5 / 0389

Purpose of generating a switching hysteresis so that in the special case when an input signal happens to be present, the corresponds exactly to the threshold signal of 4.5V, no permanent undesired switching back and forth of the Comparator is caused. The output of the comparator is connected to a flip-flop 21, to whose outputs OR gates 22 and 30 are connected. The second entrances to the OR gates 22 and 30 are connected to the output of the comparator 20. The outputs of the OR gates 22 and 30 form the Switching inputs of analog switches 23 and 29, to which the positive and an inverted negative image signal are applied.

A module of the type LM 311 from National Semiconductors, for example, can be used as a comparator.

At the switch 23 is via a resistor 24, the positive image signal, which is either directly from the scanner 4 or from the Pulse shaper 5 is removed, or it is in the case that a black / white transmission is made, a constant positive potential at switch 23. The image signal fed in at terminal 25 is also sent to an inverter given, which consists of 'an operational amplifier 26 with its feedback resistor RK -, - and a resistor 27 at the input of the operational amplifier. The output of this operational amplifier is via a resistor 28 at the input of the Analog switch 29. In the case of a pure black / white transmission, there is a negative constant potential of the same level, as in this case the switch 23 would be in positive polarity.

809845/0389

The switches 23 and 29 can be designed as simple analog switches, e.g. as transistors or as digitally controllable switches MOS switches, for which e.g. the type IH 5011 module from Intersil can be used.

The operation of this circuit is described in more detail below. The comparator 20, which comes from the operational amplifier 20 exists, may for example have a voltage of 4.5 volts at its minus input. Is the input voltage of the Comparator 20 at the plus input, which is derived from the input signal of the Impalsformerstufe 5, the output voltage Ua of the Comparator 20 and the voltage divider from the resistors 20 a and 20 b forms, less than 4.5 volts, i.e. less than that Black level, the following applies.

Let the flip-flop 21 not be set, i.e. at its Q output the value is zero and the value 1 at its «Y output. There the output of the comparator is also zero, the following applies to OR gates 2 2 and 30:

Gate 22 has the combination 0/0,
Gate 30 the combination 1/0.

That is, the switch 23 blocks and the switch 29 is permeable. As a result, the negative signal at switch 29 is let through.

809845/0389

exceeds the voltage at the positive input of the comparator 20 the value of 4.5 volts, the output voltage of the comparator U, rises to + 5V, and it becomes the input of the Flip-flops positive, but its outputs remain unchanged because it is only switched by a negative edge. To the Changes to the inputs of the OR gates 22 and 30 occur:

Gate 22 has the combination 0/1,
Gate 30 the combination 1/1.

That is, both gates switch through and also switches 23 and 29, which has the effect that at the addition point 31 an equal value positive and negative signals are present, which leads to an overall signal of zero.

If the input voltage at the positive input of the comparator 20 falls again below the black-and-white threshold of 4.5 volts, that will The output signal of the comparator is zero, i.e. a negative edge occurs which switches the flip-flop 21. this leads to a combination 1/0 at the output of the flip-flop and to the following combination at gates 22 and 30:

Gate 22 has the combination 1/0,

Gate 30 the combination 0/0,

i.e., switch 23 passes, i.e. a positive signal appears at the output.

If the input signal at the operational amplifier 20 again exceeds the threshold value, the flip-flop 21 remains unchanged, but other signal combinations occur at the OR gates 22 and 30.

809845/0389

options on:

Gate 22 has the combination 1/1,

Gate 30 the combination 1/1,

i.e. both switches 23 and 29 open, i.e. their sum signal is again zero.

The processes just described can easily be checked using the following FIGS. 6a and 6b as well as 7a and 7b.

In Figure 5, the switches 23 and 29 is a further operational amplifier 32 connected downstream, whose feedback resistor RK23 as well as resistors 24 and 28 are dimensioned so that this stage works as a simple adder. As already mentioned, the output of the operational amplifier 32 is the Modulator 6 connected, from which the signal is sent to the transmission line got.

In order to express more clearly where the difference between the present invention and the known methods of 3-level modulation is is due to a simple case in which only black / white transmission takes place in a comparison Figures 6 and 7 shown. FIG. 6 shows the transmission without the impulse setting for the flip and FIG. 7 shows the transmission with pulse splitting for the flip .

Case a) in FIG. 6 shows a signal supplied by the scanner, which corresponds approximately to signal a in FIG. It has Signal sequence WS, SW, WS, SW, WS, SW, WS, SW, WS, where WS means white and SW black.

809845/0389

In Figure 6b, the signal sequence of the 3-level signal, WS, SW, WS, SW, WS specified. As can be easily seen, the information of the WS and SW values is in the middle of the curve lost according to FIG. 6a, since their amplitudes were insufficient to exceed the threshold value for black, which resulted in all values being transmitted as white.

FIG. 7a shows the same signal train of FIG. 6a, but with pulse splitting. It is easy to see here that the mean black values of the curve exceed the threshold value for black, which means that even with the 3-level signal in FIG. b these values are built into the 3-level signal as black / white jumps and transmitted.

However, the present invention is not limited to the pure transmission of black / white pulses, as in FIG. 7 but it can also be used to transmit semitones. When transmitting using amplitude modulation It has been shown that the input signal for the comparator 20 of FIG. 5 is advantageously the one set up Image signal and as an input signal for terminal 25, i.e. for switches 23 and 29, the image signal that is not set up selects what is shown in Figures 8a, 8b and 8c.

In the case of frequency modulation, it is preferable to use the established image signal in both cases, so that the terminal can be connected to the input of the comparator 20. The corresponding curves are shown in FIGS. 8a and b as well as 9 shown.

809845/0389

FIG. 8a shows the signal delivered by the scanner 4, FIG. 8b the signal delivered by the pulse shaper stage 5, and FIG. 8c the signal at the output of the 2-level / 3-level converter of Figure 5. The intersection of the established image signal of Figure 8b with the threshold, which represents the black level, results in the fold-over points for the 2-level / 3-level converter, which are shown as vertical lines in FIGS. 8a to 8c. The curve of FIG. 8c, which supplies the signal supplied by the converter, results from the mode of operation of the switches 23 and 29 as a function from the control of the comparator and the downstream OR gates of Figure 5. At the points where the 3-level signal assumes the value zero, both switches 23 and 29 are open, so that a sum signal of zero results. In the positive Areas, the switch 23 is open and in the negative areas of the switch 29. FIG. 8c shows the case that the Analog switches 23 and 29, the non-displayed image signal of Figure 8a, be it in positive or negative polarity, switched through will.

FIG. 9 shows an output signal of the 2-level / 3-level converter, in which the established signal of Figure 8b is sent to the analog switches 23 and 29, which is, for example, when using the Frequency modulation is beneficial.

FIG. 10 shows the receiver circuit, consisting of a demodulator 8 and a 3-level / 2-level converter 11. A commercially available circuit can be used for the demodulator, as they are e.g. in the textbook "Electrical communications engineering" Volumes I and II, Dr. Ing.Heinrich Schröder, Publishing House for Radio-Photo-Cinema Technology, Berlin-Borsigwalde, is described.

809845/0389

The 3-level / 2-level converter 11 consists of 2 parts, a full-wave rectifier consisting of an operational amplifier 40, in whose feedback branch the two diodes 41 and 42 are connected anti-parallel to the feedback resistor RK ₄₀ , and a simple one Adding circuit consisting of an operational amplifier 44 with its feedback resistor RK ₄₄ . The resistors of the operational amplifier RK ₄₀ and the series resistors 44, 45 and 46 are dimensioned so that half the non-rectified input signal is added to the full amplitude of the rectified signal, whereby a 2-level signal is formed from the 3-level signal . The mode of operation of this circuit is shown in more detail in FIG. Curve I is the non-rectified signal of half the amplitude, curve II is the rectified signal of full amplitude and curve III is the sum signal that results from the sum of signal I and signal II. By inverting the operational amplifier 40, the resulting signal is curve IV. As can easily be seen, this signal is a 2-level signal. For the sake of better understanding, a pure sinusoidal oscillation was used as the basis in FIG. FIG. 12 shows the same relationships for a signal as it is transmitted over the line in the present case. FIG. 12a shows the 3-level signal transmitted and received via the line, which originates, for example, from the transmitted signal in FIG. 9 and has been blurred a little by the band-limited line. FIG. 12b shows the 2-level signal which represents the states black and white after the rectification.

809845/0389

Read 11 e

Claims (6)

Dr.-Ing. Rudolf Hell GnbH Kiel, March 23, 1977 Grenzstrasse 1-5, 23OO Kiel 14 Lf / Uck Patent application No. 77/446 2 '20079 Keyword: Impulse list of patent claims 1. A method for improving the image quality in the transmission of facsimile signals which are derived from an image signal which is essentially a signal of one polarity with a minimum amplitude for one information state (white or black) and a maximum amplitude for another information state (Black or white), whereby flip-over pulses for the formation of a signal of alternating polarity are obtained from the changes in the information of this signal, the opposite polarities of which represent one information state (black or white) and whose zero position represents the second information state (white or white) . Schwarz) and the signal of alternating polarity is formed by successively running through the operations of switching on a signal of one polarity, switching to the zero position, switching on a signal of another polarity, switching to the zero position with each flip pulse, characterized in that that the image signal before it becomes The formation of the flip-over pulses is used, an amplitude division is subjected and compared with a threshold value for obtaining the flip-over pulses. 809845/0389 2. The method according to claim 1, characterized in that the formation of the signal of alternating polarity by means of the flip-over pulses is obtained by switching to 2 constant voltages of opposite polarity. 3. The method according to claim 1, characterized in that the formation of the signal of alternating polarity by means of the flip pulses by switching to the image signal and to a signal obtained from the image signal by inversion is obtained. 4. The method according to claim 1, characterized in that the formation of the signal of alternating polarity by means of the flip-over pulses by switching to the established image signal and to a signal obtained from the established image signal by inversion is obtained. 5. The method according to any one of claims 1-4, characterized in that the image signal is differentiated twice, is added to the original image signal with opposite polarity and is compared to form the flip-over pulses with a threshold value. 6. The method according to any one of claims 1-4, characterized in that an ambient signal is superimposed on the image signal, and that the sum signal for forming the flip-over pulses is compared with a threshold value.