DE978033C - Method for encrypting and decrypting halftone images in image telegraphy - Google Patents

Description

3 4

have been built, so they have punched themselves in practice, which later as impulse combinations were not able to enforce as before, because they are far removed from the usual Telegraphic over an economic solution, the transmission channel are sent,
In the Bartlane process, 1 mm ² image area available in the trade is largely broken down into about 6 square image points, so use makes it. that a pixel density of. 2.45 pixel-completely different possibilities arise, on the other hand, ten / mm, which results in a dimension of 0.41 X 0.41 mm, which has been known for a long time, and an area of mixed processes and image telegraphy processes-0.166 mm ² for corresponds to a pixel. The line density, which both borrowed from teletype technology, is also 2.45 lines / mm. This image is, raster is rather coarse in today's terms was. The image to be transmitted is transmitted in the usual pulse combinations, with this combination of codes being scanned photoelectrically in helical lines by a rotating drum and an ingenious switching process. The inions are ambiguous, so that one can display fluctuating Photo-64 characters triggered by the photocell. Controlling currents according to a primitive quantization To encode the code impulse combinations by means of threshold value circuits, five translators are used, especially in the case of commercial relays that operate a hole punch, which has long been a very quantized and coded image amplitude image simple method, which is used in this exists that punches point by image point into a punched tape, which one transports the individual clear pulses of the impulse combination with the image point scanning speed with clocked by a key pulse generator. The punched tape is z. B. sent in pairs 25 high-speed telegraphs by means of a key pulses supplied immediately and mixed in in the receiver, in such a way that reproduced according to Vernam's goblet of a punched tape receiver. The character rules, the mixing or superimposing decoding happens by means of a very simple impulses of the same type (present or absent) the one, optical back-translator. Due to the number and unequal impulses (present and absent) the position of the holes of the individual hole combination results in a different type of impulse, for which there are two equivalent functions, the exit brightness of one of the hole possibilities. The resulting encrypted combinations of light source radiating through the bottom pulses and their combinations go as indirectly controlled, and the controlled concentrated secret pulses or secret combinations via the exit light beam to the photographic transmission channel. The decryption, that is, the recording of the blackening quanta on light emitting back mixing, is done in the same way as the sensitive paper or film is used directly. Encryption using the same key by combining the well-known Vernamschen selimpulse according to the same sign rules in front of you. Teletype encryption method with the mixing of the individual Bartlane image coding methods in pairs now results in pulses which, based on the group, sometimes use a very effective encryption and decryption theory as a symbolic multiplication, which according to the invention is therein If there is that one draws and notes, there is also a mixed law defining a certain initial assignment between the (or two) for the complete pulse combinations quantized image amplitudes and the code pulses, which converges remarkable symmetry combinations in the sender and receiver over properties such as commutativity and involuteness Way has running after a key and also enjoys the group features. 45 is changed.

The key impulses are generated by a peri- According to one embodiment of the inventive concept odic or non-periodic, meaningful or meaningful, the clear image amplitudes are first delivered with keyless keyhole stripes, the sampled amplitudes are mixed and then the secret and either coded by pressing the remote amplitude buttons.

typewriter is transported further, or the 50. According to a further embodiment of the invention Punched tape operation synchronized with the clear-hole idea, first the quantized clear-image strips is transported. The mixture is amplitude-coded and then the clear impulse com is through a so-called mixing gate, which is made up of relay combinations with key pulse combinations or transistors, and that as “twilight.

lingskopf "known, built into many teleprinters 55 According to a further embodiment of the invention is. First of all, the clear image amplitudes are thought. Bartholomew had a long time ago with key amplitudes according to a first law and McFarla.ne a coding method for the mixed, then the secret amplitudes and coded Halftone image telcgraphy created, which under the then the secret pulse combinations with The name "Bartlane" system became known 60 key pulse combinations after a second is. The continuous scale of the brightness or law mixed again.

Gradations of blackening are determined by quantifying the telex encryption.

sizing into 32 discrete brightness or blackening known mixing methods according to Vernam's

levels divided so that the continuous gray wedge sign rules has proven to be very effective and

is approximated by a stepped gray wedge. The 32 65 has been proven to be secure against unauthorized decryption,

discrete hue amplitudes are converted into a five- when generalized to the effect that the

digit binary code and the code combinations - mixture of the pulse combinations not

tinnen as combinations of holes in a punched tape according to Vernam's sign rules, but

which takes place in combination according to any uniquely reversible mixing function. This Generalized mixing methods can be generalized even further by adding the up to then fixed Mixing function is continuously changed over time. For the generation of temporally variable mixing functions suitable methods and devices have been proposed.

As for the number of blackening levels required for good image reproduction, psychological Tests on a large number of test persons show that the limit of the resolving power of the human eye for discontinuous gray wedges at around 25 to 30 levels of blackness lies, d. that is, at this maximum number of levels between deepest black and lightest Knows the boundaries of two successive levels can just be distinguished, and that with an even larger number of steps, the discontinuous gray wedge is perceived as continuous will. The ability to discriminate depends on the lighting, and when trying is a defined normal (daylight) lighting was used as a basis. It has also been shown that the resolving power of the human eye i: i the bright parts of the gray wedge are greater than in the dark parts, so that there are more steps in the part of the gray wedge that is above 50% gray than can be distinguished below this limit.

For a satisfactory image reproduction, 32 levels are generally quite sufficient. For the highest demands, on the other hand, have to be passed to 64 levels of blackening. This applies e.g. B. if very gradual tone value transitions are reproduced with a pronounced lack of structure should.

The advantages of the pulse code modulation transmission method (PCM) as time division multiplex compared to the classic carrier frequency transmission method are known to the person skilled in the art, especially in the case of shortwave connections, in which it is known Due to the strong signs of shrinkage, a true-to-amplitude transmission, at least temporarily not possible. With PCM all that is needed is the presence and absence of pulses of the same, im but the rest of the amplitude to be differentiated in the receiver because of the message carrier is no longer the amplitude, but the pulse configuration. The signal-to-noise ratio is essential lower than with the carrier-frequency method, and it is still possible with a signal-to-noise ratio of about 2 to carry out a flawless transfer.

While with conventional carrier frequency technology, multiple use occurs by staggering the individual messages according to frequency, with PCM multiple use is achieved by staggering the individual messages (e.g. long-distance calls). In contrast to the carrier method, in which all transmission channels are occupied with messages at the same time, with the PCM there is only a single message element in a single transmission channel at any given moment. Therefore, the appearance of the cross-calculation is less critical here. Since a regeneration of the impulses, i.e. a new formation of the message content, can take place at any point along the route, be it in an amplifier office or in a relay point, there is no addition of the interference amplitudes over the entire transmission path, i.e. no reduction in the transmission quality due to stringing together of amplifiers.
The disadvantages of PCM are the greater equipment required compared to carrier frequency technology and the greater bandwidth requirement. which for an r-digit code is r times that for the carrier-frequency method and

ίο is referred to as bandwidth luxury. With m messages (channels) each with the frequency bandwidth J, the bandwidth required for PCM is r f.

The drawings show two examples

games of the invention shown in block diagrams.

Fig.l and 2 show a transmitting and receiving system for encrypted image telegraphy for one channel each, whereby the mix between the clear code

characters and the key code characters as in de: Teletype encryption is carried out in pulses.

F i g. 3 and 4 show a transmitting and receiving system for encrypted picture telegraphy for one channel each, whereby the mix between the Klaras impulse combinations and the key impulse combinations according to any mixing law in combination he follows.

Fig. 5 and 6 each show a code hole plate eini Coding tube, namely Fig. 5 for the normal

Binary code and F i g. 6 for the Goodall code.

In Fig. 1, 1 is an image transmitter for halftone images which, in a known manner, consists of a rotating drum on which the original image to be transferred, usually a black and white photograph, is recorded.

is tensioned, soft by a photoelectric scanning device point by point and line by line in a helix or in equally spaced circles is scanned. For this purpose, the scanning system runs parallel to the drum axis along the drum

Melmcintellines a continuous or after eachei Drum rotation from a step-shaped feed movement relative to the drum. Depending on The selected module results in a scan line density of 5 to 10 lines per millimeter. Correspond to this!

an equally large pixel density within an image line of 5 to 10 pixels per millimeter. The scanning speed for the small modules is around 40 cm / sec and for the large modules about 20 cm / s: ec. This corresponds to an almost constant transmission speed of around 200C Pixels per second (2000 baud) or a frame rate of 1000 Hz, if this is assumed that a grid consisting alternately of black and white picture elements with a detent point spacing from 0.1 to 0.2 nm is scanned at the speeds given above. The low frequency The signal bandwidth is therefore runc for the commercially available image transmission devices 1000 Hz.

The surface discharge speed is constant! Frame rate is inversely proportional to the square of the scan line density «. It moves about between 0.2 and 0.8 cm ² / sec. The transmission time of the usual picture format of 16.5 X 20 cm varies jf

~ 5 according to the Abiasil line density between 7 and 28 minutes.

In the low-pass filter 2, the image signal is fixed to the Bandwidth / = 1000 Hz cut and after ver

Strengthening in the image intensifier 3 in the tone value reversal stage 4 is subject to a tone width reversal required because the photocell of the scanning system determines the brightness (transparencies) of the pixels, consequently the image amplitude is proportional to the brightness, while the writing lamp of the image receiver, which is controlled by the image signal, records blackening. Therefore, apart from of the cases where a negative image, or vice versa, is recorded from a positive image should, the scanned brightnesses of the image points are converted into blackenings. this happens in a simple manner in that the positive image signal from a constant. Tension equal to that White level is subtracted. The effect is the same as if instead of the positive image its negative (more precisely: its complementary image) would be scanned. The inverted tone level can be used instead of in the Transmitter can also be relocated to the receiver.

The tonal value reversal stage is now followed by those devices that use pulse code modulation known and necessary for their implementation.

First is by the rotating or electronic Scanner 5 briefly and periodically the image signal with double the highest signal frequency, so with 2 / = 2000 Hz, sampled. This periodic signal sampling makes the continuous Replaced image signal by a finite number of temporally equidistant discrete amplitude values. To According to Shannon's sampling theorem in telecommunications, this sequence of discrete amplitude values replaces the continuous signal without any loss of information complete, provided that the signal is restricted to a certain frequency bandwidth (/) is, which is practically always the case, and that the signal is interrogated at a frequency that is at least twice (2 /) the signal bandwidth. This was sent from the discrete amplitude values existing signal through a low-pass filter with the cutoff frequency /, so you would at his Output to get the continuous original signal back true to the amplitude.

In most cases, the signal samples are then subjected to logarithmic compression according to the law y = In (I + vx) , where ν is the distortion coefficient. This makes the distortion factor almost independent of the amplitude. Since the approximation of the image signal curve cannot be subdivided as finely as desired by a stepped curve, either no amplitude is transmitted below a certain amplitude in the range of small amplitudes, or the transmission is only incomplete. It follows from this that the distortion factor increases as the amplitude becomes smaller. If the amplitudes are now logarithmically kGrnpnrn; crt, the small amplitudes are no longer disadvantaged and thus involved in the transmission, while the large amplitudes are compressed.

The holding circuit 6 follows the sampler 5, which essentially consists of a capacitor, which holds the interrogated pulse-shaped amplitude value over the entire interrogation period T = Vi / and which is discharged again by a discharge diffuser at the end of the interrogation period. The queried discrete instantaneous signal values are not yet suitable for coding in a binary code, since they can have any possible value within the signal level, but the coding only requires a finite number of different amplitude values. The queried amplitude values are therefore quantized in the quantization device 9. This means that the signal level is broken down into a finite number of levels, which is equal to the number / 1 = 2 'of the code characters of an r-digit binary code. The amplitude values, which will generally lie somewhere between the quantum levels, are replaced by the amplitude quanta lying next to them. In the case of amplitude values that lie exactly in the middle between two successive quantum levels, it is left to chance of the quantization device, which in these cases becomes unstable, whether they are replaced by the next lower or next higher amplitude quantum. The quantization results in a distortion of the amplitude quanta compared to the original amplitude values, which is known as quantization distortion, in that an error of at most half a quantum level is committed for each amplitude value that is not a quantum by chance. The greater the number of quantum levels, the smaller this error is. After the quantization, the amplitude quanta are coded in the coding device 10. If the η amplitude quanta are ordered according to increasing values, they can be numbered (counted) by the first / 1 natural numbers, a power of 2 being expediently chosen for η. If you write these natural numbers h with the help of the two digits 0 and 1 in dyadic notation, i.e. as a sum of consecutive powers of 2:

h = a "2o + β, · 2i + ... 4- a _e ■ 2 _e + ... + a _r _ , · 2 ^,

where the prefixes a _L , nu can have the values 0 or 1, and r - ^{2 is} log η , if the digit 0 is represented by the absence and the digit 1 by the presence of an impulse always of the same amplitude, then the the first η natural numbers and thus the η amplitude quanta as pulse combinations of an r-digit code and transmit them electrically.

A number of quantization and coding devices are known. Are most elegant the electronic coding tubes, since they do the quantization and coding simultaneously in a single one Allow device to be carried out. A coding tube is a cathode ray tube, which apart from the more usual Deflection devices a catch grid, a quantization grid, a perforated plate for the coding and has an anode designed as a collecting plate instead of the screen. On the perforated plate (Fig. 5, 6) the binary character combinations are represented by punched holes and spaces. The interrogated signal instantaneous voltages which are received by the hold circuit 6 over the interrogation period ( are held, are attached to the pair of plates for the vertical deflection (y) of the cathode ray placed, causing the cathode ray to deflect upward linki of the perforated part of the perforated plate by a distance proportional to the applied voltage. Then the beam will with the help of the quantization grid guided exactly into the row of holes that correspond to the deflected position Closest to the cathode ray. According to the Quanti

The cathode ray, which is now exactly at the level of a wire of the quantization grid, is deflected horizontally from left to right by a voltage from the sawtooth generator 11 that is applied to the pair of plates for the horizontal deflection (x) of the cathode ray. The beam is then made ineffective by a blanking control and tilted back into its starting position to the left of the hole gaps. At the same time, the capacitor of the holding circuit 6 is discharged by a discharge pulse, whereby the beam is tilted vertically downwards back into its zero position. During the horizontal deflection over a row of holes, electrons pass through the holes in the perforated plate, which are captured by the anode and deliver a binary combination of pulses. Since all values within the signal level can still occur under the queried signal voltages and since the cathode ray has a finite cross-section, it will usually happen that the ray hits between two rows of holes in the perforated plate and in this way scans both rows of holes at the same time, which means that there is no clear one or a wrong combination of impulses would be triggered. When the electron beam hits a wire of the quantization grid, secondary electrons are knocked out of it and sucked out through the catching grid in front of it. The grid current is greater when the center of the beam hits the wire directly and less when the center of the beam is between two adjacent wires. The grid currents are fed back to the deflection amplifier for the vertical beam deflection, with the result that the electron beam is held by the grid wires and only the deflection paths along these grid wires are stable. All queried signal voltages that differ from a quantum level by less than the difference between two successive quantum levels are converted in this way into the quantum voltage corresponding to this quantum level and give rise to a pulse combination that is assigned to this quantum level.

In the case of pulse code modulation, it is customary to select the assignment of the pulse combinations to the amplitude quanta in such a way that those from the occupation of the r pulse positions of the binary code are added by adding their place values

(2 », 2i, 22, ..., 2'-i)

50

resulting sum increases and decreases in the same direction as the quantized amplitude value. The reason for this is that the pulse combinations assigned in the aforementioned manner can be decoded particularly easily. In and of itself, however, other assignments are also conceivable in that one assigns some fixed permutation of the pulse combinations sorted according to increasing values to the amplitude quanta, which are ordered according to increasing values. However, the decoding of the pulse combinations becomes more complicated with such an arbitrary assignment. If one arranges the pulse combinations in a natural way according to the first η = 2 ^r natural numbers which they represent, one notices that two successive pulse combinations generally differ from one another in more than one place. Arrives at a pulse combination due to some disturbance

At a point where a pulse change takes place in relation to the following or the preceding pulse combination, if a pulse is incorrectly added, or if a pulse fails, the combination changed by the disturbance generally represents a completely different numerical value than the natural one at a completely different point Order. The addition or omission of a pulse at the (i'-l) · position of an r-digit code pulse combination [the first position on the left is the 0., the last position on the right is the (r- 1). Digit], in which a pulse change takes place, results in a change in the assigned numerical value by one unit of the digit 2 "-i this (« - 1). Digit. This change can at most at the digit with the highest digit n / 2 = 272 = 2 '"» units.

According to a code proposed by Good all, the impulse combinations can be arranged by a certain permutation of their natural order in such a way that two successive combinations only differ from one another at a single point, so that the erroneous addition or omission of an impulse at the point where the change from the preceding or following pulse combination always only causes a change in the associated numerical value by one unit of the digit with the lowest value (20 - = 1), regardless of where the change took place. In addition, one can obtain further codes from the Goodal! Code by cyclic permutations n- \ = 2'-1 which have the same property.

Because of this property of the Goodall codes can be m a Kodierröhre which has a corresponding to the Goodall code punched Kodelochplatte, quantization and thus the guards, the Quantisierungsgitter and the feedback omitted Ünn _who n of the vertically deflected cathode ray a Mellung assumes that is close to the border of two successive quantum levels, the uncertainty of the pulse combination triggered by the horizontal deflection of the cathode ray through it is at most one pulse at only one point, the uncertainty of the assigned amplitude quantum, i.e. only at most one quantum mare, regardless of where could UN ^Sl T "" i ^uu """'while possibly at a die-cut after the normal Binarkode Kodelochplatte without quantization by the horizontally deflected cathode ray a completely wrong Impulskomb nat, one ^from Selöst be, which would be associated with an amplitude Quant which around Γη T ^{= 272} Units of the to be coded

AU η ^mp ! '. ^tude could distinguish. However, encoding the Goodall code is not as easy as the usual binary code.

^ e Commercially available coding tubes for the transmission of speech in the PCM method, which are available for both rLnv ^normal binary code and for the 7,? ir S ^de , use a 2 Λ " ^Bl " ^arkGde with η == 2 ' = 128 amplitude levels (including amplitude zero), i.e. with / hole gaps and 128 hole rows. The first, well-known, very elegant method of encoding, which can be carried out in the coding tubes itself without additional effort, and an ongoing change is due to the lucidation of the code holes the assignment of the

Code pulse combinations for the image amplitudes are permitted. For this purpose, key amplitudes are additively superimposed on the image amplitudes, which come from the same level range as the image amplitudes and which are taken from a special key amplitude generator. The sum amplitudes are applied to the plates for the vertical deflection of the cathode ray, and an addition modulo 2 ^r is thereby obtained. The secret amplitudes belong to the same signal level range as the clear amplitudes, but generally no longer supply a continuously variable message signal.

For a better understanding, FIGS. 5 and 6 serve, in each of which a perforated plate for the five code with η = 2 ' = 32 amplitude levels (including the amplitude zero), namely in FIG. 5 for the normal binary code and in F i G. 6 for the Goodall code. With regard to the implementation of the decoding, which will be discussed later, the hole gaps are arranged in such a way that they provide pulse combinations which are reversed compared to the dyadic notation (lowest value on the right, highest value on the left), so that the first pulse position on the left has the lowest value ( 2 ° = 1) and the last pulse point on the right has the highest value (2 ^r ~>), whereby it is assumed that the cathode ray is ^{deflected vertically 1} ^ n bottom left to top left and horizontally from left to right. The holes in each following column are always twice as high as those in the previous column and are twice as far from each other. Using the digits 0 and 1 (0 = non-hole = non-pulse, 1 = hole = pulse), the corresponding reversed dyadic notations of the individual pulse combinations are entered. The assigned amplitude quanta, numbered from 0 to 31, are written between the two perforated code plates. Since, as already mentioned, generally 32, but at most 64 amplitude levels (including zero amplitude) are sufficient for image transmission, the last two hole gaps on the right or the right of a coding tube with a perforated plate for the seven code (128 amplitude levels) are used. the last hole column on the right is not required. This is achieved by giving the sawtooth amplitude for the horizontal deflection a value which only leads to a deflection of the cathode ray up to and including the fifth or sixth hole gap. If one imagines the sixth and seventh or the seventh hole column to be omitted, then in FIG. 5 the binary code reduced by two digits or by one digit four times or twice on top of each other. If the sum level is limited in such a way that it delivers a maximum vertical deflection over 64 or all 128 rows of holes, with both the image level and the key level extending over 32 or 64 rows of holes, one obtains with the perforated plate according to FIG. 5 an addition of image and key amplitudes modulo 32 and 64, respectively.If K is the quantized clear amplitude (image amplitude), S is the quantized key amplitude and G is the quantized secret amplitude, then one obtains, if Ätis'i'- i fails,

so generally

G ₁ = K + S mod. 2 ^r ,

d. H. the secret amplitude is a sum amplitude that corresponds to the level range of the clear or key amplitudes is reduced.

The situation is somewhat different with a perforated code plate (Goodall code) according to FIG. 6. If you ignore the sixth and seventh or the seventh hole column here, the Goodall code, reduced by two places or one place, appears in both cases mirrored on the horizontal center line between the 63rd and 64th perforated disc In addition, the Goodall code reduced by the last two digits, the two parts of the code hole plate separated by the center line appear again mirrored on their center lines between the 31st and 32nd and between the 95th and 96th row of holes. If the total level is limited here too in such a way that it delivers a maximum vertical deflection over 64 or all 128 rows of holes, with both the image level and the key level extending over 32 or 64 rows of holes, then one obtains when adding image- and key amplitudes H = "Prnjin7ii _n n (fa * Komolement) of the sum amplitude for the amplitude 2 ^r ~ ⁱ - 1, i.e. in the present case. ^. 1 i ^ Vispitio ^ uncn (r = 5 or 6) for the amplitude 63 or 127. With the terms already used, written in formulas, one obtains, if K + S <^ 2 ^r -l

G ₂ = K + S,

G ₂ = 2 ^{r +} il-

G ₁ = K + S,

ie as the secret amplitude a sum amplitude which is again reduced to the level range of the clear or key _{amplitudes, but is complementary to the sum amplitude G 1} with respect to the number of amplitude levels of the code used (31 or 63), which is reduced by 1.

A numerical example should make the situation even clearer.

If a five-code with 32 amplitude levels is used and is z. B. the clear amplitude 9, the Key amplitude is 13 units, so the secret amplitude is in both cases

G = K + S = 9 + 13 = 22, since 22 <31.

But is z. B. the clear amplitude 23 and the key amplitude 29 units, the sum amplitude A is

A = K + S = 23 + 29 = 52> 31.

In the ordinary binary code, this is reduced to the secret amplitude G ₁ = 52-32 = 20 by subtracting 32, while in the Goodall code it is reduced to the secret amplitude G ₂ = 63-52 = 11 by forming the addition to 63. The two different secret warnings obtained in this way, however, always add up to 3i. This is also shown by the general formulas when one forms G ₁ + G ₂ :

G ₁ =

2 ^r ,

The addition of the key to the clear amplitudes is done in the addition stage 7, in the z. B. the both voltages can be connected in series.

The key amplitude generator 8 can be designed in very different ways. So he can provide a continuous, or a discrete non-quantized, or a discrete quantized key signal.

A keyhole strip is expediently used to generate the key amplitudes, However, because of the high scanning speed of 2000 hole combinations per Second no longer directly electro-mechanically or electro-optically, but after the preceding one Programming in the form of contact potential configurations can only be scanned electronically can. The sampled potential configurations are decoded and provide the discrete quantized ones Key amplitudes which are added to the image amplitudes in addition stage 7. Details This will be discussed in the following description of the key pulse generator 16, the does not differ significantly from the key amplitude generator 8.

After the quantization and coding in the devices 9 and 10, the pulses generated in the coding tube, which are unsuitable for further use, have to be shaped or regenerated. For this purpose, after sufficient amplification, they are sent through a double-sided amplitude filter 12, which cuts out thin slice pulses of constant amplitude near the zero line. The extraction of these window pulses is a combined pulse-length (PLM) and pulse-phase modulation (PPM). The window pulses open gate 13 for their duration, which is a number of normalized square pulses from the code pulse generator proportional to the opening duration 14 passes, which generates the code pulses at the frequency 2 rf and in phase with the pulses coming from the coding tube.

Instead of the encryption already described, a reduced addition of key to the Image amplitudes, or in addition to this, the PCM method allows a very effective further encryption, which is known in principle from telex encryption and also a Continuous change in the assignment of the code pulse combinations to the key amplitudes is permitted.

This is the purpose of the mixer 15 and the key pulse generator 16, which according to any Program a periodic or non-periodic, meaningful or meaningless sequence of key pulses from of the same type and in the same text as the code pulses. The mixture is done z. B. after Vernam's sign rules, which state that if the presence of an impulse with (-I-) and the absence in a clock interval is denoted by (-), the meeting of the same signs of image and key one, unequal sign the other as a secret sign results in what there are two different equivalent possibilities, namely

This rule is commutative, that is, you can swap the clear pulse and key pulse, and get the same secret pulse in both cases: G = KS = SK. Furthermore, the assignment is involutive. The clear pulse is obtained from the secret pulse using the same commutative mixing rule:

K = GS = SG.

The mixture is not mixed by character combination, but by character elements in pairs. Therefore, it can be carried out one after the other, just as the clear and final impulses are delivered one after the other. The impulsive mixing, however, corresponds to a very specific mixing rule of complete impulse combinations, which has some remarkable symmetry properties. This rule of mixing can be expressed as follows. If one writes the impulse combination;: n with the help of the two digits 0 (absence) and 1 (presence) in (right-sided) dyadic notation, the mixture is the addition of two five-dimensional vectors, the components of which can now take the words 0 and 1, Modulo 2. If one has, for example, the two vectors (1 1011) and (M101) mii: to add the numbers 27 and 29 modulo 2, the sum vector results in

(11011)
(11 101)

Löö 11 ο),

= -JC - = -

and

■ i-x- = ■■ -x +

to which the number 6 would be assigned, while the arithmetic addition of the two numbers 27 and 29 to

27 = 1.1011

29 = 1 .1101

56 = 111,000

leads. This reduced addition is not to be confused with the addition modulo 2 ^r discussed on the occasion of the encoding of the image amplitude. This would be the case with both of them. Deliver numbers for the binary code 56 - 32 == 24 and for the Goodall code 63 -56 = 7 instead of 6.

As is known, the Mischs; tuf (! 15 consists of a so-called mixing gate, a combination of an AND and an OR gate with the effect that, if at both control inputs at the same time Voltage or no voltage, the output carries voltage, and if only at one of the two Control inputs voltage is present, the output is de-energized, or vice versa.

The key pulse generator 16 can be implemented in many ways. In the simplest case it consists of a non-periodic keyhole strip in the radio code and five scanning devices, wherein the keyhole strip with randomly distributed hole combinations by means of a random generator will be produced. The key impulse generation is in this primitive form with the image encryption not feasible. Because 2000 hole combinations are required per second, what a Strip transport speed of 5 m / sec would require. At most, this could be done with magnetic tapes or drums, but the tape or drum consumption would be enormous. if such lanyards should last for some time,

the encryption partners would have to be given thick rolls, the value of which would be very questionable if they got lost or fell into unauthorized hands. Therefore, key extension devices were found a long time ago conceived by a relatively short, non-periodic original key streak (so-called »I-worm«) with only a few hundred randomly distributed hole combinations. This Original perforated strip has different perforated columns from the number of levels of the binary code used and the number of comparison combinations, e.g. B. 32: 8 = 4, which means that on average every fourth Query step leads to a hit and thus to a key pulse, but this does not mean that hit interval 4 dominates. The smaller hit intervals predominate. The probability for the occurrence of larger hit intervals it becomes smaller and smaller and decreases asymptotically to zero.

lengths (parallel to the longitudinal extent of the strip), io Since there are 2000 key impulse comuins every second number of prime characters (perforations, i.e. 10 000 key codes for the five-digit code) and non-hole), so that he must be available at his one impulse, must be the end is stepped stepped. The hole gap question frequency in the example is at least 40 kHz will amount to a corresponding number of sampling. Because of the irregular occurrence of the ein.ichiungei scanned synchronously and periodically, but for safety reasons the Abdaf through the different. Lengths of the frequency of questions must be at least five times greater, i.e. a: e! rif.n columns i; n over time ei 3S increasing be about 200 kHz. Here is one of the five Ve shifting the columns to each other results in each of the hit pulses delivered per clock interval alternating impulse combinations across because only the first one is used. The key impulse They mature in the scanning facilities, 20 generator is immediately after the occurrence of a first with a period that is equal to the product of the Zei pulse and switched off in the next clock interval ch <nelemenleannummer of all hole gaps is. Is - - - ·· · .... · ·· .: 1

z. ] $. c. the length of the hole gaps on average 100 hole combinations, what a stripe length of about

25 cm, the period is around 25, the one for the odd, 100 ⁵ == 10 ¹⁰ sampling steps. This period is sufficient for the other for the even number of key pulses.

a sampling frequency of 2000 hole combinations ... j j

per second and with uninterrupted operation about

58 days or around 2 months, with eight hours a day.

So happy to operate for half a year. Since the periodic 30 e 'ments. If this known scanning of the hole gaps is used with the large scanning travels of the "cross-comparison query", further speeds can no longer be carried out reliably electromechanically or electro-optically,
it is carried out electronically by means of a ring counter, a contact plate being established beforehand by placing the hole 35.

Stripe electrical in the form of contact potential- In the superimposition stage 17 the code

configurations are programmed that are scanned instead of the hole combinations (see German

Patents 10 95 312 and 1101492). The Ab-,

scanning of the individual contact columns takes place in each case 40 a rigid synchronization of the various loca not exactly at the same time, but with the five-digit code len pulse generator with the corresponding one on the

Bring about the sending side. For the synchronization, ϊ .. B. a further clock interval can be provided, i.e. a sixth for the five-code code and a seventh for the six-code code, in which the synchronization pulse is accommodated. This clock interval is placed at the beginning of each clock period, roughly like the start step in teletype operation, so that each time a new one begins

equal cross query «(see German patents 50 clock period is synchronized again. The syn-10 12 635 and 10 74 630). Here the chronizing impulses are not made, with the same amplitude scanned hole combinations used, like the code pulses, somewhat longer_than these and but these are merged with one or more them again in the receiver through a time filter different (but a maximum of 31 for the five-digit code). out by about through a KC limb with fit-fixed Hole combinations compared continuously. 55 a voltage is built up with the time constant If a sampled value matches any threshold value that is determined by an amplitude of a of the fixed combinations, which filter is separated while the shorter code suitability is referred to as a "hit", an impulse * is not used to build up this threshold value in the same cycle from another generator is sufficient. By adding another delivered alternating pulse train, the alternating 60 clock interval, however, occurs a slight band positive and negative impulses or from image enlargement by »/. or. /"a and 'there are absent impulses, which in the hit- There are, however, also rigid synchro-

The moment at which the pulse of the alternating measures is known, selected without a bandwidth sequence and enlarged for encryption, if the code is reversed. The "hits" occur here completely in terms of time as synchronization pulses themselves, so occur irregularly, so that they can appear very sparsely or very sparsely depending on the composition of the individual clusters. The middle successive ^Im P ^u ^ i? "\ ^Bi ; ^at '°" ^e """""lere" frequency of hits "is equal to the quotient regularly on, but at intervals, which are always

switched on again briefly, i.e. operated intermittently. There are also two removable storage elements required for the key impulses that alternate

While one memory element is being filled, the other is being read. Repeated in the next section of the measure playing with exchanged memory

ICtill WHO UVl "ν Vl giviwiiu ^ v ... * .w ~ _o - - ■..

Clock generators for the clock frequencies 40 kHz or 200 kHz are required, which are made in frequency multipliers derived from the clock frequencies of 2 kHz and 10 kHz

3 -

Pulse combinations mixed in synchronization pulses that originate from the clock generator 22 and which are detached again in the receiver in order to

with a time offset of '/ 10000 see so that the Pulses of the individual pulse combinations in succession in time with the cycle of the code pulse generator 14 be available.

It is now a more complicated one, but much more secure in terms of encryption technology Key renewal procedure has become known, which is based on the principle of the so-called »

are an integral multiple of the small clock period τ = V «r / between two successive pulses of the code pulse generator 14. If you now have a master clock generator with a natural frequency of 2 rf stabilized by a tuning fork or a quartz on the receiving side, you can pull this with the irregular, but always in-phase code pulses at the respective times of their arrival. The other lower clock frequencies are obtained by frequency division. The latter type of synchronization is preferable, since the first type could provide indications of unauthorized decryption by fading in special synchronization pulses.

The synchronization stage 17 is followed by the low-pass filter 18, which cuts the low-frequency frequency band to rf and frees the encrypted signal function from its direct current components, so that the code pulses are not transmitted sharply, which would require a very wide frequency band, but rounded off so that they are continuous secondary Signal function leave the transmitter. In the transmission amplifier 19, the signal is amplified to the required transmission level and then sent to the trunk line 20. In the case of wireless transmission, the encrypted secondary signal function is modulated onto a high-frequency carrier; 20 then means the antenna.

Are interleaved in time when time division multiplex transmission method in which a plurality of transmission channels for m messages, the image transmitter 1 is thinking replaced by m image transmitter and the sensor 5 by a Mehrfachabtaster for m image signals of the m image signals sequentially periodically and briefly mi <a m- times higher frequency samples. Instead of a quantizing and coding tube 9/10, two are required, one for the odd and the other for the even number of channels. The periodic switching from one tube to the other is done on the input and output side by two rotating or electronic switches. In the case of time division multiplexing, the bandwidth r · / for a channel is multiplied by the number m of channels; the required bandwidth is therefore mrf, and the bandwidth luxury is r compared to the carrier-frequency transmission method.

The clock-correct control of all parts of the transmission and encryption system requires a central clock generator 22 with the frequency 2 / = 2000 Hz, from which the frequency multiplier 23 derives the clock frequency 2 rf = 10 kHz and possibly other higher clock frequencies, depending on the nature of the key amplitude generator 8 and the key pulse generator 16. Via the line 25, the scanner 5, the hold circuit 6, the key amplitude generator 8, the sawtooth generator 11 for the coding tube 9/10, the code pulse generator 14, the key pulse generator 16 and the synchronizing device 17 are controlled with the clock frequency 2 / . The code pulse generator 14 and the key pulse generator 16 are also controlled via the line 26 with the clock frequency 2 rf.

The signal generator 21 is provided for starting up the transmitter and the receiver from the transmitter. Starting takes place in the usual known manner that a start button is pressed for about 1 second, which triggers a start signal which turns on the drive motor of the image transmitter 1 via line 27 and the mother clock generator 22 via line 28 the line 29 to the carrier frequency generator 24, where depending on the type of transmission a tone or high frequency alternating voltage keys and via the transmission channel 20 to the remote receiver, in which the drive motor of the image receiver and the receiver-side mother clock generator is switched on at the same time. The transmitter and receiver image drums are rotated into the phase position in which they are held by the holding devices. After releasing the start button it disappears; Start signal, which automatically triggers a short after-signal after a short time, which simultaneously releases the image drums in the transmitter and receiver and at the same time via line 29 the key amplitude generator 8, the sawtooth generator 11, the code pulse generator 14 and the key pulse generator 16 in the transmitter and receiver of switches on the same agreed and set initial positions, which means that the encrypted transmission operation begins synchronously and concurrently in both positions.

In Fig. 2 shows the receiving system. After any demodulation (in the case of wireless transmission), the encrypted image signal passes from the transmission channel 20 (long-distance line or antenna) through the low-pass filter 30, where it is cut to the signal bandwidth rf (mrf with m channels). After amplification in the receiving amplifier 31, the synchronization pulses are separated in the subsequent time filter 32, which essentially consists of an RC-GYied , which synchronize the local master clock generator 33 with the frequency 2 / with the corresponding clock generator 22 of the transmitter (FIG. 1). The rounded impulses of the continuous encrypted reception signal function now have to be regenerated. For this purpose, the bilateral amplitude filter 34 is used, which, after adding a direct current component, cuts thin slice pulses from the signal in the vicinity of the zero line. These disk pulses each open gate 35 for their duration, which gate 35 lets through a number of normalized square-wave code pulses of frequency 2rf proportional to the opening duration from the local code pulse generator. The pulse regeneration is now carried out. In the following mixer stage 37, which works exactly like the transmitter-side mixer 15 (FIG. I), the encrypted pulses are separated in pairs one after the other with the key pulses supplied from the local key pulse generator 38. The law of backmixing according to Vernam's sign rules is the same as the law of mixing, since this is invoicing. The key pulse generator 38 is the same as the corresponding one (16) on the Scndcscite, and the key pulse generation is started using the same (original keyhole strip as on the transmitter side from the same predefined and set starting position The pulse combinations must now be decoded. This is done according to an ingenious method by Shannon in the two decoders of the same type 39 and 40, which are to be used alternately, one of which is intended for the odd-numbered and the other for the even-numbered pulse combinations.

\ L

19 20

dische switching is done by the two electronic devices for decoding the

Two pole rotating or electronic switch Goodall codes indicated, also a procedure

11 and 42 with the switching frequency f. Two decoders and an electronic device for coding

are necessary because the reconstructed ampli in Goodall code without the use of a coding

tude of a clock period T only at the beginning of the post-5 tube.

th clock period is available, the restore If the transmitter end an additional encryption estellte amplitude for the current clock period, the amplitudes by addition of Schlüsselamplidaher by decoding the pulse combination amplitudes modulo 2 ^Γ is made, the Her must preceding clock period encrypted image amplitudes are obtained in the Empfangsiß which cannot be decrypted again at the same time in the same pre-IO system. For this purpose, preparation is possible. next the addition stage 43, in which the decoders 39 and 40 essentially consist of secret amplitudes, the fixed quantum amplitude of • _e from a KC parallel circuit, which is added by means of a 2 ^r = 32 voltage units, which first gate, which by the incoming from the constant voltage source 44 with which the EntImpulse is opened for its duration and as 15 speaking voltage U _{0 is} supplied. In the nachein switch works, with a constant span following subtraction stage 45 is _ver from the sum source is connected, where z. B. subtracted the key amplitude by a menamplitude, the very ° large series resistor or a pentode generated by the key amplitude generator 46 for constant charging current of the capacitor, which is provided to the corresponding generator 8 RC circuit during the pulse duration, 20 (F i g. 1) is completely the same on the transmitter side, but regardless of the already existing compared to this by a clock period T in the charging of the capacitor. It is assumed phase is lagging behind. Because the recovery of the men that the capacitor at the beginning of the pulse amplitudes by the decoders 39 and 40 on the folee of a combination has no charge and that the receiving side takes place with the Shannon Dekoäie pulse combination 01001 = 18 (in the side-converting process by exactly one clock period Γ later reversed spelling) is to be decoded. Further than the sampling or the beginning of the coding, it is assumed that the capacitor is on the transmitter side during the image amplitudes. The clearing pulse duration is always around 2 ', i.e. in the example image amplitudes are increased according to their anti-arithmic (Me by 32 voltage units, expansion then through the low-pass filter 4 / lTden, regardless of which charge it 30 with the cutoff frequency / = 1000 Hz In the first clock interval τ = V ₂ / - / ge through the signal bandwidth again relates nothing to IkHz since this is not occupied by any pulse. The result is the restoration during the duration of the pulse in the second phase of the continuous Image signal, which is the clock interval, is connected to the capacitor via the gate with the power source by scanning the transmitted image in the Her and is briefly charged to 35 image transmitters of the transmission point (Fig. 1) after the lon-X) solar units. During the value reversal, apart from the minor error, the free time interval between two successive ones is the same as a result of the quantization. The vibrating impulses are generated by the KC circuit from the current, the clear image signal is then separated in the ouelle, and the capacitor discharges, amplified image intensifier 48 and the image receiver Sam via the resistance of the KC circuit. The 40 49 supplied, in which the by the image signal ZeCstante is selected so that the capacitor controlled write lamp records the image on phctogra- "lannune exactly to the half in each Taktinter- phischem paper or film VAU absfnkt must then RC = Vr / ln2 be. The same generators as au de ^ n up to the middle of the third, further to 8 units 45 transmission side are required for the clock-correct control of the described Daheffänt the capacitor voltage to 16 units. From the clock frequency 2 / des up to the middle of the fourth clock interval, the two mother clocks 33 are occupied in the frequency multiplication with pulses, and finally on fan 50 the clock frequency 2rf and in the Fre-4Knhdten Ws Lr in the middle of the fifth clock interval, quenzteiler 51 the Clock frequency / S ™ "; ^ teaches is again occupied with a pulse. While the line 52 is the ^» PV? ^ «^ For the duration of this pulse in the last clock inter- 50 the key pulse generator 38, the two decoders vailist the capacitor again Connected to the power source 39 and 40 and the Schlüsselampl.tuaengenerato and is controlled by a further 32 voltage inputs with the clock frequency 2 / via ä.z Uu ~ * Γ d so that it is now 4 + 32 = 36 Span- 53 are also the code pulse generator 00,

_S5

Rator 38 and the key amplitude generator 46 switched on, namely at the same time and from the same agreed and set initial positions as in the transmitter, the key amplitude generator 46 but with a phase lag of one clock period T compared to the corresponding generator 8 on the transmitting side.

In the case of time division multiplexing with m channels, the two switches 41 and 42 should be replaced by multiple distribution switches for m image signals, which periodically distribute the odd-numbered channels to the decoder 39 and the even-numbered channels to the decoder 40.

The F i g. 3 and 4 each show a block diagram of the transmitting and receiving systems for encrypted messages Image telegraphy with a variant of encryption of the code pulse combinations.

While in Fig. 1 and 2 the mixture of the code pulse combinations with the associated key pulse combinations impulsively and successively in time according to Vernam's sign rules takes place, whereby a very specific mixing law for the complete pulse combinations results, the code pulse combinations with the key pulse combinations can also after mixed with any other prescribed law of mixing. However, this works in this one general case no longer impulsively and sequentially, but both the Code pulse combinations and the key pulse combinations at the same time as potential configurations are present.

To set up any mixed assignment, a square table in matrix form is created with a vertical input column for the clear combinations, which can be noted by the assigned amplitude quantum numbers, and with a horizontal input row for the key combinations, which are also expediently noted by their numbers. At the crossing points of the columns and rows, the assigned secret combination G is entered for each ordered combination pair K, S by its corresponding number. The assignment in such a table, which mathematically amounts to a definition of a tabulated discontinuous function of two variables, with the value set for the two independent variables and the dependent variable coming from one and the same limited set, does not need to be commutative or involutive

The only condition for unambiguous decryption, ie back mixing, is that the table does not have two identical columns and that each combination occurs only once in each column, that is to say that all columns are different permutations of any order of the combinations. If the assignment is commutative, the table is mirrored on the main diagonal (from top left to bottom right). If the assignment is involutive, ie if the decryption table and the encryption table are identical, the permutations of the various columns are symmetrical pennutations (these are those _: those carried out with themselves one after the other; the identical permutation results). Both requirements can be met at the same time. Uies trim, for example, in the case of the law of mixing according to Vernam's sign rules.

By di * German patent specification 9 59 020, which the French patents 10 50 064 and 62 545 corresponds, oleklronic devices have become known, which carry out the mentioned arbitrary mix assignments with great speed allow. The mixing devices change: 15 and 37 as well as the decoding devices 39 and 40; η F i g. 1 and 2 are significant and require additional storage facilities. The aul the PCM-related parts, however, remain the same and lead in FIG. 3 and 4 have the same reference numbers.

In Fig. 3, which represents the transmission system, follow the gate 13, which lets the code pulses of the code pulse generator s 14 through during its opening times, the pulse memory 58, which collects the code pulses of an r-digit combination that arrive one after the other until their respective r are together. At the r output terminals of this memory

the pulse combination is then available as a potential configuration. The memory 58 is controlled by the clock generator 22 with the clock frequency 2] and briefly releases the stored combination to the mixer 59 at the end of the clock period.

* 5 The key pulses delivered one after the other by the key pulse generator 16 are collected in the memory 60 until their respective r are together. The r-digit key pulse combination is then at the / ■ output terminals of this memory

as a potential configuration. The memory 60 is also controlled by the clock generator 22 with the clock frequency 2 j and, at the end of the clock period, briefly transfers the stored key combination to the second input of the mixer 59

free. But this no longer consists of one simple mixing gate as in FIG. 1 and 2, but is considerably more complicated in its structure. At the exit of the mixer 59, the encrypted pulse combinations are again available as potential configuration

functions that are still generated by a pulse generator in a timed sequence of occurring one after the other Impulses! 1 resolved and released as pulse combinations to the following devices will.

The final pulse generator 16 can experience certain modifications. So it is not necessary that the hole gaps of the Urlochstripe be a normal keyholestripe extension, ie without "cross-comparison query", from the r scanning devices

be sampled with a small time offset, but it. it is enough that this happens at the same time, whereby the pulse combinations as potential configurations at the output of the key pulse generator :; 16 appear. The memory 60

can then be omitted. If the "cross-comparison query" is used, it can do this be modified (see German patent specification 10 12 635), so that when a "hit" occurs, it is not from an alternating pulse train, but from

^{an arbitrary periodic fig of the 2 r} hole combinations offered with the same query speed, the hole combination currently present at the time of the hit is selected and simultaneously scanned and thus used as a potential

S ₅ configuration is available. In this case, too, the memory 60 can be used in its previous. No form. Instead of this, however, two exchangeable memories are required, but this time not for individual impulses,

but for r-digit pulse combinations, which are possible for safety reasons due to their time-irregular decoder 64 using a diode or toroidal core matrix , than pulse combinations - is better adapted. This she needs to graze. In this case, too, the 5 applies above all if the transmitter side has intermittently coded Goodall code by means of the key pulse combination generator, which is not operated according to the fact that only the first of several can be decoded using the Shannon method,
Pulse combinations delivered in one clock period - A diode or toroidal core matrix has used sources and the generator has continuously used an r-digit code for r inputs, to which and is switched off. ^{l0 the potential configurations to be} decoded ge-In F i g. 4 is the one with FIG. 3 corresponding give and 2 ^{r shown} after ascending amplitude receiving system. The gate 35, which has the quantum-ordered outputs, at each of which the code pulses of the code pulse generator 36 occurs to a potential change when it lets through its opening time, is followed by the memory 61, the relevant output associated potential configuration with the codes arriving one after the other at the input occurs. Each matrix output is impulses of an r-digit combination, until connected to the control input of a locking gate, their respective r are together. At the r output and the gate inputs are connected to the partial points of a clamp this memory is then the encrypted evenly subdivided voltage divider connected pulse combiner. sen as a potential configuration, at which one of the number of stages to be decoded before. The memory 61 is controlled by the clock generator 20 of the binary code proportional constant quantum 33 with the clock frequency 2 / and gives the voltage. If the stored combination shows a change in potential at a matrix output at the end of the pulse period, it opens freely to the backmixing device 62 for a short time. This time the associated gate, and at the gate output of the key pulse generator 38 appear successively in time the passed through it and that of the delivered key pulses are collected in the memory of the quantum voltage of the rather 63 assigned to their respective r together voltage divider as amplitude. All gate exits are connected to the r output terminals of this memory via decoupling diodes in parallel to an Aus-HeRi then the r-digit key pulse combination output line. Diode or Rmgkern as a potential configuration. The memory 63 is matrices as a decoder can also be advantageously used by the clock generator 33 with the clock 30 whenever any arbitrary frequency2 / is controlled and gives the stored key combination short to the pulse combinations at the end of the clock loan fixed initial assignment of the signal amplitude period of a binary code is prematurely on the second input of the backmixing device, which cannot be decoded in any other way. 62 free This is generally different from the transmitting and receiving system are in the so-called mixer 59 according to FIG. 3. It only resembles it in "On Line" mode, ie the entire case where the decryption key is the same as the apparatus for pulse code modulation and the encryption key, the mixed assignment coding is in the transmitter and receiver side tion is therefore involutive. At the output of the mixing connection line 20 between the image transmitter 1 and the device 62, the decrypted impulse combinations image receivers 49 are connected, without the two commercially available devices being present again as potential configurations at these functions Interventions to be carried out need a fair sequence of sequential occurrences. In the case of Abden on the transmitter and receiver side, pulses can be resolved by means of two circuits of this apparatus can be undecoded. encrypts with one another in the usual way, but it is also a known way of decoding.

For this purpose 3 sheets of drawings

Claims (11)

Patent claims: 1. Method for encrypting and decrypting halftone images in image telegraphy, under Use of pulse code modulation, with the image amplitudes of the images to be transmitted quantized and coded and transmitted as pulse combinations and these after transmission again to the original image amplitudes »are decoded, characterized in this that a certain initial assignment between the quantized image amplitudes and the Code-pulse combinations at the transmitter and receiver in a matching manner is continuously changed with a key 2. The method according to claim 1, characterized in that first the clear image amplitudes are mixed with key amplitudes and then the secret amplitudes are coded. ^ao 3. The method according to claim 1, characterized in that the quantized K arbildamplituden first encoded and then the clear pulse combinations with key pulse combinations be mixed. 4. The method according to claim 1 to 3, characterized characterized in that initially the clear image amplitudes mixed with key amplitudes according to a first law, then the secret amphetas encoded and then the secret pulse combinations with key pulse combinations be mixed again according to a second law. 5. The method according spoke 2, thereby modulo amplitude level (20 takes place. 6. The method according to claim 2, characterized in that that the mixture of the Klarbüdmit the key amplitudes by addition modulo amplitude level (20 and subsequent complementation to that reduced by one unit Amplitude level (2 ^ -1) takes place 7. The method according to claim 2, characterized in that that the mixture of the clear image with the key amplitudes after a lively, reversible unique mixing function 12. The method of claim 1 to 11, characterized in that the encrypted image transmission in the so-called "On Line" - is operated transmission method wherein the leases for the pulse code modulation and J £ ^nC _ildve 5 _circuiting i _un g sender and emp angswe _{the Verbindun} gsweg between Image suction _{image receivers} can be switched on without _any changes or interventions that E ^ _{Ger "th vorgenomme} n are, at this _Invention relates to methods for encrypting and decrypting halftone images in the BildteIe-E ™ ^USS J use of the pulse code ^{modules S ra} P ^h '^e; _b ^U _{ei the} image amplitudes of the tion to be transmitted w ° _{and coded and} as a pulse combination "" ° * _transferred and these nations after the transfer and β. _{Q Bildamplltud} en deco- again m α men. | _the encryption of semitones was »^^ _{approaches not yet} befnedi- * _{"D" solves} this may partly be explained by the quietly 8 * ° "· _{in a} ^S _usreic Hendes need this his OaBDi sn ^ ^^ ^ _{of creation wdt} . vorg, 1 g of military intelligence networks is the ""! gJS "of halftone images currently encrypted ™ ^επ \. South encryption method known, it w _dept . _and recording speed S of the pictures according to a key in fixed stages of buoe ^^ _{and receiver in} - ¹ it is known to provide additional locking '" _{h the} picture lines are randomly Störbildse ^ ung ™ d« nae ^ _Νβώ ^ _that ^ 4 «lines emzutug _wifd Transfer time (PP _{) known) various} JJ ^ _{image to be} transmitted by several ! ^ devices to be scanned at the same time and scanning device eg _{and E ß in} J «^ ™ ^ ⁰ ^ ,," After a key time has _{to be} mixed up. known, the mentioned procedures f The method according to claim 3, characterized in that the mixture of clear with the key impulse combinations impulswe.se takes place according to Vernam's sign rules. 9. The method according to claim 3, characterized in that the mixture of clear with the key combinations of impulses wise according to any, reversibly unique Mixing function takes place. 10. The method according to claim 1 to 3, characterized in that the mixing of the clear signals Lt key signals follows a ze.thch variable mixing law. 11. The method according to claim 5 and 6, characterized in that the mixture of the clear image the key amplitudes through reduced addition at the same time as the Q ^ ization and Coding in an electronic coding tube (cathode ray tube with quantization gatter and perforated code plate). fSF ^ ehKndg, where- images act and the band- ^ JL ^ aTbeStlich is larger, are in principle pulp ^ bed «· fb ™ g ^^ _{combined before} . the same 'verianre G «cWagen ^" ^ _{h is in common} that they ^ jSß _they require considerable additional ^^ JJ ^ _elek , ri _S chen effort in the telegraphy devices, so that commercial images ^le B "P ⁿ '^. _h can be used. Da ^blic ^ ^Ger ^ ⁿ g _n ricSgen mostly mecha _v ^E ; ⁿ _w ™ ^g _are ⁸ _whose periods verso that they are against an unauthorized person and _electronic means come to light work YES enngk ^ h ^, ch ^ ^ .. ^. ^ _{Degrees at} Jf'i ^ J ™ _{"can ziclen} ⁵ _as it is nowadays ge-go irnha J ⁿ 8 ™ _{inzelt okhe to recite} i also asked una vein